Australian and New Zealand Journal of Public Health 2… · Professor John Lynch School of Public...

ISSN 1326-0200 Vol. 43 No. 4, 2019Public Health Association

A U S T R A L I A

The Journal of the Public Health Association of Australia Inc.

Australian and New Zealand Journal of

Public Health

Sports Food and Beverage

General Public HealthIndigenous Health

Australian and New Zealand Journal of Public Health

Statement of policyThe Australian and New Zealand Journal of Public Health is the

journal of the Public Health Association of Australia. PHAA members have training in almost all of the human, natural and social sciences, at various levels of professional status. Some are employed to analyse the ideological, social or empirical features of the health service. Some begin from a basic, some from an applied, perspective; others come to research by reflecting on the work they do in health care – for example, organising industrial health services in a particular locality, trying to implement a patient-held record system or using lay helpers in a domiciliary care system. Others carry out formal epidemiological research into the correlates and causes of disease and of health-related behaviour.

The Australian and New Zealand Journal of Public Health is published six times a year, in February, April, June, August, October and December. Its contents are subject to normal refereeing processes. Finished discussions of research projects are the staple diet of the Journal, but there is space for reviews, views and historical pieces from time to time. The Journal is indexed by Australian Public Affairs Information Service, Cur rent Contents , Excer pta Medica, Index Medicus, the Cumulative Index to Nursing & Allied Health Literature and Social Sciences Citation Index and is available on microfiche from University Microfilms International.

Most of the disciplines embraced by PHAA publish journals that carry articles about facets of health, illness and health care. However, there is no other Australian journal that gives an overview of research across the broad range of PHAA interests, nor does any other journal aim to attract more than one or two of the many levels of workers in health care assessment and delivery. The Australian and New Zealand Journal of Public Health invites contributions which will add to knowledge in its fields of interest. It will give priority, after normal refereeing processes, to papers whose focus and content is specifically related to public health issues.

SubscriptionsPlease address all inquiries about subscriptions, membership, advertising and other PHAA matters to the Public Health Association of Australia Inc., PO Box 319, Curtin, ACT 2605. Phone (02) 6285 2373; Fax (02) 6282 5438; e-mail [email protected]; www.phaa.net.au

Editorial officePlease address all editorial correspondence to: The Editors, Australian and New Zealand Journal of Public Health, e-mail [email protected]

Editorial Board

Professor Ross BailieMenzies School of Health Research, Northern Territory

Dr Sandra CampbellCentre for Chronic Disease Prevention, James Cook University, Queensland

Professor Donna CrossTelethon Kids Institute, Western Australia

Professor Joan CunninghamMenzies School of Health Research, Northern Territory

Professor Chris Del MarFaculty of Health Sciences and Medicine, Bond University, Queensland

Professor Kevin DewSchool of Social and Cultural Studies, Victoria University of Wellington, New Zealand

Professor Annette DobsonSchool of Public Health, University of Queensland, Queensland

Dr Rhys JonesTe Kupenga Hauora Māori, University of Auckland, New Zealand

Professor John LynchSchool of Public Health, University of Adelaide, South Australia

Professor Robyn McDermottCentre for Chronic Disease Prevention, James Cook University, Queensland

Professor Robert McGeeDunedin School of Medicine, University of Otago, New Zealand

Professor Terry NolanSchool of Population and Global Health, The University of Melbourne, Victoria

Dr Yin ParadiesFaculty of Arts and Education, Deakin University, Victoria

Professor Andre RenzahoHumanitarian and Development Studies, Western Sydney University, New South Wales

Professor Peter SainsburyDirector, Population Health, South Western Sydney Local Health District, New South Wales

Professor Cindy ShannonPro-Vice Chancellor (Indigenous Education), The University of Queensland

Professor Alan ShiellSchool of Psychology and Public Health, La Trobe University, Victoria

Assoc. Prof. David ThomasTobacco Control Research, Menzies School of Public Health, Northern Territory

Professor Gavin TurrellSchool of Public Health and Social Work, Queensland University of Technology

Professor Alison VennMenzies Institute for Medical Research, University of Tasmania

Editor-in-Chief: Professor John Lowe

Adj. Assoc. Professor Priscilla Robinson School of Public Health, La Trobe University, Victoria

Dr Sandar Tin TinSchool of Population Health, The University of Auckland, New Zealand

Assoc. Professor Luke WolfendenSchool of Medicine and Public Health, The University of Newcastle, NSW

Editors:

Dr Melissa StonehamPublic Health Advocacy Institute WA, Curtin University, WA

Dr Hassan VallySchool of Psychology and Public Health, La Trobe University, Victoria

Dr Nikki PercivalAustralian Centre for Public and Population Health Research, University of Technology Sydney, NSW

Editor-in-Chief: Prof. John Lowe Book Review Editor: Jo-Anne Rayner Production: Journal Assist Pty Ltd.

ANZJPHThe Australian and New Zealand Journal of Public Health is the journal of the

Public Health Association of Australia Inc.

Volume 43, Number 4 August 2019

ContentsEditorial 305 The Public Health Association of Australia’s advocacy to prevent suicide

Samantha Battams, Fiona Robards

Commentary 307 Listen, understand, collaborate: developing innovative strategies to improve health service utilisation by Aboriginal and Torres Strait Islander men Kootsy Canuto, Stephen Harfield, Gary Wittert, Alex Brown

310 The important role of charity in the welfare system for those who are food insecure Fiona H. McKay, Rebecca Lindberg

Indigenous Health 313 Feasibility and acceptability of opportunistic screening to detect atrial fibrillation in Aboriginal adults Rona Macniven, Josephine Gwynn, Hiroko Fujimoto, Sandy Hamilton, Sandra C. Thompson, Kerry Taylor, Monica Lawrence, Heather Finlayson, Graham Bolton, Norman Dulvari, Daryl C. Wright, Boe Rambaldini, Ben Freedman, Kylie Gwynne

319 Anaemia in early childhood among Aboriginal and Torres Strait Islander children of Far North Queensland: a retrospective cohort study Dympna Leonard, Petra Buttner, Fintan Thompson, Maria Makrides, Robyn McDermott

328 Participant profile and impacts of an Aboriginal healthy lifestyle and weight loss challenge over four years 2012-2015 Anne C. Grunseit, Erika Bohn-Goldbaum, Melanie Crane, Andrew Milat, Aaron Cashmore, Rose Fonua, Angela Gow, Rachael Havrlant, Kate Reid, Kiel Hennessey, Willow Firth, Adrian Bauman

334 Breast screening attendance of Aboriginal and Torres Strait Islander women in the Northern Territory of Australia Kriscia A. Tapia, Gail Garvey, Mark F. McEntee, Mary Rickard, Lorraine Lydiard, Patrick C. Brennan

340 Limited progress in closing the mortality gap for Aboriginal and Torres Strait Islander Australians of the Northern Territory Tom Wilson, Yuejen Zhao, John Condon

Food and Beverage 346 The frequency and magnitude of price-promoted beverages available for sale in Australian supermarkets Christina Zorbas, Beth Gilham, Tara Boelsen-Robinson, Miranda R.C. Blake, Anna Peeters, Adrian J. Cameron, Jason H.Y. Wu, Kathryn Backholer

352 Development of Australia’s front-of-pack interpretative nutrition labelling Health Star Rating system: lessons for public health advocates Michael Moore, Alexandra Jones, Christina M. Pollard, Heather Yeatman

2019 vol. 43 no. 4 Australian and New Zealand Journal of Public Health 304

Food and Beverage cont. 355 The performance and potential of the Australasian Health Star Rating system: a four-year review using the RE-AIM framework Alexandra Jones, Anne Marie Thow, Cliona Ni Mhurchu, Gary Sacks, Bruce Neal

Sport 366 Unhealthy sport sponsorship at the 2017 AFL Grand Final: a case study of its frequency, duration and nature Tegan Nuss, Maree Scully, Melanie Wakefield, Helen Dixon

373 Challenges for sport organisations developing and delivering non-traditional social sport products for insufficiently active populations Kiera Staley, Alex Donaldson, Erica Randle, Matthew Nicholson, Paul O’Halloran, Rayoni Nelson, Matthew Cameron

General Public Health 382 Epidemiology of hospitalised traumatic brain injury in the state of New South Wales, Australia: a population-based study Ilaria Pozzato, Robyn L Tate, Ulrike Rosenkoetter, Ian D Cameron

389 New and old hotspots for rickettsial spotted fever acquired in Tasmania, 2012–2017 Gabriela Willis, Kerryn Lodo, Alistair McGregor, Faline Howes, Stephanie Williams, Mark Veitch

395 Public health and natural hazards: new policies and preparedness initiatives developed from an Australian bushfire case study Rachel Westcott, Kevin Ronan, Hilary Bambrick, Melanie Taylor

Letter 401 Observed vaping and smoking in outdoor public places: piloting objective data collection for policies on outdoor vaping George Thomson, Johanna Nee-Nee, Kirsty Sutherland, Rebecca Holland, Miriam Wilson, Nick Wilson

2019 vol. 43 no. 4 Australian and New Zealand Journal of Public Health 305© 2019 The Authors

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

In December 2018, the Public Health Association of Australia (PHAA) endorsed its first Suicide Prevention Policy. The

purpose of this article is to outline some concerning trends in suicide and suicide inequality, the opportunities to prevent suicide and the role of public health.

Suicide trends

The Australian suicide rate decreased in the period from 1994–1998 (17.8 per 100,000) to 2009–2013 (12.3 per 100,000).1 However, recent trends have showed a slight increase in the standardised suicide death rate between 2007 (10.6 per 100,000 people) and 2016 (11.7 per 100,000 people).2 The suicide death rate is now higher than the rate from motor vehicle accident deaths, and suicide is the leading cause of death of young Australians.3 Population groups most at risk include males (especially the middle-aged and older age group),4 Aboriginal and Torres Strait Islanders and those of lower socioeconomic status (SES). Suicide is a major cause of premature mortality for Aboriginal and Torres Strait Islander people, with a rate of suicide (23.8 per 100,000) more than twice the Australian national average. Indigenous young people (aged 15–24 years) are particularly vulnerable, with the suicide rate in 2016 almost four times that of non-Indigenous young people. Some groups from culturally and linguistically diverse backgrounds are also at higher risk of suicide, with suicide rates initially associated with country of birth5 and the experience of detention for asylum seekers.6 Suicidality has also been associated with the experience of disability in Australian men.7 Deaths from suicide have recently increased for Australian women; in 2017, the age-standardised suicide rate for females was higher than that of the previous ten years.14 In addition, deliberate self-harm is a significant issue in Australian society, particularly for young women: the rate of hospitalisation for females due to self-harm was 40% higher than for males from 1999–2000 to 2011–2012.15

Is there suicide inequality?

Suicide and suicidality disproportionally affect those who are poor and Aboriginal and Torres Strait Islanders. Suicide is related to unemployment and periods of economic crisis.16 Research from both Australia and Europe has indicated a recent increase in suicide inequality. In Europe, there were 1.82 more suicides in the lowest SES group than in the highest in the 1990s, and 2.12 more suicides from the lowest to highest group in the 2000s.8 In Australia, suicide inequality in older males (35–64 years) increased by 29% from 1999–2003 to 2004–2008, associated with an increase in suicide rates in low SES regions.1 The PHAA is committed to reducing health inequality and has recently updated its health inequity policy.

Framing suicide prevention

Risk factors for suicide are often framed in terms of individual psychological or life experience factors; for example, experience of a mental health condition or a sudden ‘crisis’ event, previous attempts at suicide, or having a friend or family member who has died by suicide. However, the risk factors for mental health conditions, suicide and suicidality are multifactorial, operate at many levels and may overlap. They may involve individual, relationship/family level, workplace, societal/community, political and economic levels. For example, gender and cultural factors linked to intimate partner violence may contribute to the problem of suicide, as the experience of intimate partner abuse has been linked to suicidality.9 Perceived racism has also been linked to suicidality, with a mediating role of depression and moderating role of religiosity.10 Male-dominated industries such as the construction industry have also been linked to higher suicide rates (especially for men),11 with research showing that those in the most unskilled occupations are most at risk within the construction industry.12 While economic crises are associated

with a spike in the suicide rate,16 research from Spain showed that this trend is more pronounced for those aged 35 to 54 years and unemployed males.13 In addition, the link between increasing inequalities and suicide and the trend in suicide inequity is not often part of the discourse on suicide.

What is the opportunity to prevent suicide?

Suicide prevention activity covers a broad range of policy and program activity that may include: limiting access to the means of suicide through legislation and policy; provision of education on mental health and suicide prevention, including in schools, workplaces and across community venues; providing support and transition for those affected by changing workplace conditions and retrenchment; training frontline workers on understanding suicide; provision of timely access to mental health information, support and services; timely community based support for those who have exhibited suicidality or have made a suicide attempt; and postvention support for those bereaved by suicide.

Additionally, suicide prevention strategies should consider the complex way in which individual, relationship/family level, workplace, societal/community, political and economic factors may overlap. Investment in tailored, multi-sectoral and community-level interventions and prevention for populations at high risk of suicide and self-harm is required.

Given the role of inequality in suicide, suicide prevention advocacy should also consider the social determinants of health and policies to reduce health inequities. More research is required on the social determinants of suicide.

Why the PHAA Policy Position Statement is important

The Policy Position Statement makes several important requests. It calls on the Australian Government to support and fully resource national and state and territory suicide prevention and mental health strategies, including those for Aboriginal and Torres Strait Islander people, and to develop specific strategies for high suicide risk groups, including middle-aged men.

doi: 10.1111/1753-6405.12909

The Public Health Association of Australia’s advocacy to prevent suicide Samantha Battams,1 Fiona Robards2 1. Southgate Institute of Health, Society and Equity, Flinders University, South Australia

2. The University of Sydney, Faculty of Medicine and Health, New South Wales

Editorial

306 Australian and New Zealand Journal of Public Health 2019 vol. 43 no. 4© 2019 The Authors

Editorial

Trends for the future include monitoring the increasing suicide inequality in Australia and understanding how socioeconomic inequalities have an impact upon suicide and interact with other issues, including individual, cultural and political/economic factors.

References1. Too L, Law P, Spittal M, Page A, Milner A. Widening

socioeconomic inequalities in Australian suicide, despite recent declines in suicide rates. Soc Psychiatry Psychiatr Epidemiol. 2018;53(9):969-76.

2. Australian Bureau of Statistics. 3303.0. - Causes of Death, Australia, 2016. (AUST): ABS; 2016.

3. Australian Bureau of Statistics. 3303.0 - Causes of Death, Australia, 2017. Underlying Causes of Death (Australia) Table 1.3. Canberra (AUST): ABS; 2018.

4. Burns RA. Sex and age trends in Australia’s suicide rate over the last decade: Something is still seriously wrong with men in middle and late life. Psychiatry Res. 2016;245:224-9.

5. Ide N, Kolves K, Cassaniti M, De Leo D. Suicide of first-generation immigrants in Australia, 1974-2006. Soc Psychiatry Psychiatr Epidemiol. 2012;47(12):1917-27.

6. Dudley M, Steel Z, Mares S, Newman L. Children and young people in immigration detention. Curr Opin Psychiatry. 2012;25(4):285-92.

7. Milner A, Bollier A-M, Emerson E, Kavanagh A. The relationship between disability and suicide: Prospective evidence from the Ten to Men cohort. J Public Health (Oxf). 2018. doi: 10.1093/pubmed/fdy19.

8. Lorant V, de Gelder R, Kapadia D, Borrell C, Kalediene R, Kovacs K, et al. Socioeconomic inequalities in suicide in Europe: The widening gap. Br J Psychiatry. 2018;212(6):356-61.

9. McLaughlin J, O’Carroll RE, O’Connor RC. Intimate partner abuse and suicidality: A systematic review. Clin Psychol Rev. 2012;32(8):677-89.

10. Walker RL, Salami TK, Carter SE, Flowers K. Perceived racism and suicide ideation: Mediating role of depression but moderating role of religiosity among African American adults. Suicide Life Threat Behav. 2012;44(5):548-59.

11. Milner A, King T. Men’s work, women’s work and suicide: A retrospective mortality study in Australia. Aust N Z J Public Health. 2019;43(1):27-32.

12. Milner A, Niven H, LaMontagne A. Suicide by occupational skill level in the Australian construction industry: Data from 2001 to 2010. Aust N Z J Public Health. 2014;38(3):281-5.

13. Córdoba-Doña JA, San Sebastián M, Escolar-Pujolar A, Martínez-Faure JE, Gustafsson PE. Economic crisis and suicidal behaviour: The role of unemployment, sex and age in Andalusia, southern Spain. Int J Equity Health. 2014;13:55.

14. Australian Bureau of Statistics. 3303.0 - Causes of Death, Australia, 2017. Canberra (AUST): ABS; 2018.

15. Harrison JE, Henley G. Suicide and Hospitalised Self-harm in Australia: Trends and Analysis. Injury Research and Statistics Series No.: 93. Canberra (AUST): Australian Institute of Health and Welfare; 2014.

16. Nordt C, Warnke I, Seifritz E, Kawohl, W. Modelling suicide and unemployment: A longitudinal analysis covering 63 countries, 2000–11. Lancet Psychiatry. 2015;2(3):239-45.

Correspondence to: Dr Samantha Battams, Southgate Institute of Health, Society and Equity, Flinders University, SA 5042; e-mail: [email protected]


There’s been enough talk, now is the time for action. Primary health care services (PHCSs) need to collaborate

and develop innovative strategies to increase the use of health services by Aboriginal and Torres Strait Islander men. Currently, Aboriginal and Torres Strait Islander men are viewed as being disinterested in their health, thus, the blame is being placed on the individuals themselves for their under-utilisation of PHCSs. In contrast to this misconception, studies have found that Aboriginal and Torres Strait Islander men are interested in their health but many face significant barriers that hinder access.1-7 In response, Aboriginal and Torres Strait Islander men have identified strategies for PHCSs to reduce barriers and increase their use, which fundamentally includes working with local men to develop innovative strategies.1,5-7

Aboriginal and Torres Strait Islander men are frequently described as having the worst health and social statistics in Australia. The life expectancy gap8 and burden of disease9

remains unacceptably high. The ill health of Aboriginal and Torres Strait Islander men is demonstrable across virtually all measures of mortality and morbidity;10 this group also experiences high rates of suicide, homelessness, unemployment and imprisonment, all of which contribute directly and indirectly to ill health and many other markers of wellbeing.11

Unfortunately, the commentary that often accompanies these statistics remains largely negative and either explicitly or implicitly places blame and personal responsibility for ill health and social disadvantage on

the lifestyle ‘choices’ of these men.12-14 Such blame is unhelpful, unwarranted and – in some cases – directly harmful. It is also often a result of an ideological position that seeks to place the onus of people’s own misfortunes on themselves, thus, ignoring the pervasive effects of disadvantage, inequality and structural racism on illness and its determinants. In addition, racism continues to shape Australian policies, laws and community perceptions, and plays an equally pivotal role in framing the social determinants of health for Aboriginal and Torres Strait Islander people.12

The causes of male health disadvantage are both complex and interwoven. Marmot15 suggested poverty and inequality are largely responsible for the significant life expectancy deficit faced by Aboriginal and Torres Strait Islander people; however, the social determinants, which play a significant part in the ill health of these same men, are but one facet in addition to a litany of other contributing factors that must urgently be addressed.

Health seeking

Generally, Australian men are considered reluctant to seek help for their own health issues. As Smith et al. explained, “it is commonly held that men delay help seeking because they are ignorant about and disinterested in their health”.16(p1) Such generalisations hide important contextual and more complete understandings of the reasons for poor healthcare use and rarely include the laymen’s perspectives relating to men’s help-seeking practices.16 Indeed, the

lack of men’s voices is also consistent within discussions of Aboriginal and Torres Strait Islander men and their under-utilisation of health services.

The available data detailing Aboriginal and Torres Strait Islander health service use is patchy17,18; however, most indicates that Aboriginal and Torres Strait Islander men use PHCSs at lower rates than their female counterparts, especially for preventative healthcare. Many authors suggest Aboriginal and Torres Strait Islander men tend to delay care, often presenting at a time of advanced or serious illness.1,10,11,19 Yet, access and utilisation are a function of multiple, complex and interacting factors that enable (or inhibit) Aboriginal and Torres Strait Islander men from accessing and using available care. These issues may include a lack of continuity of care, cultural factors pertaining to communication and understanding, counteracting social pressures, and both self-determination and control. Essentially, as Hayman et al. observed, part of the problem derives in the fact that “Aboriginal and Torres Strait Islander people are not sufficiently involved in planning, delivering and evaluating relevant healthcare services”.20(p485)

The perception that Aboriginal and Torres Strait Islander men are both disinterested in and reluctant to engage with their health is a common assumption, which, perhaps, stems from little being done to listen to and learn from their perspectives. Others, such as Brown et al. instead posit that Aboriginal and Torres Strait Islander men are very interested in their health and wish to engage with primary and other healthcare services, yet are rarely consulted on what they seek and how services can better meet their needs, and seldom informed about alternate approaches to healthcare access and use.21 Herein lies the enormous challenge facing services and policy makers alike.

Health service utilisation is critical, as access to and appropriate use of comprehensive and high-quality PHCSs can have a significant effect in the health and wellbeing of marginalised and disadvantaged populations.22-24 PHCSs and key stakeholders must first understand the reasons surrounding this phenomenon of under-utilisation, although identifying the barriers faced is simply not enough. Health services must be willing and able to make

doi: 10.1111/1753-6405.12922

Listen, understand, collaborate: developing innovative strategies to improve health service utilisation by Aboriginal and Torres Strait Islander menKootsy Canuto,1,2,3 Stephen Harfield,1,3 Gary Wittert,2 Alex Brown1,2,3,4

1. Wardliparingga Aboriginal Research Unit, South Australian Health and Medical Research Institute, South Australia

2. Freemasons Foundation Centre for Men’s Health, University of Adelaide, South Australia

3. Centre of Research Excellence in Aboriginal Chronic Disease Knowledge Translation and Exchange (CREATE), University of Adelaide, South Australia

4. Sansom Institute for Health Research, University of South Australia

Commentary



Commentary

the necessary changes to improve access, evaluate their strategies, share their findings and improve continuity of care. A fully committed, reliable and sustained approach is essential, as band-aid solutions will not and have not worked.

Barriers

The barriers to health service use for Aboriginal and Torres Strait Islander people were explored in the 2014–2015 Aboriginal and Torres Strait Islander Health Survey conducted by the Department of the Prime Minister and Cabinet.19 This review found that 35% of respondents believed they had been treated unfairly within the previous 12 months because they were an Aboriginal or Torres Strait Islander person. Of those, 13% reported they had avoided seeking healthcare due to experiencing unfair treatment in the past.19

The participants in three additional studies exploring the barriers and enablers for primary healthcare access faced by Indigenous men all felt health services and staff needed to be more culturally appropriate, while many also thought they lacked information regarding services available at primary health care centres.1,3,25 Additional barriers included distrust and fear of health services, as well as shame and stigma around sensitive health issues. This highlighted the importance of safe and supportive spaces for Aboriginal and Torres Strait Islander men especially when dealing with sensitive health, social and cultural concerns, an issue previously raised by community-based Indigenous researchers.26

Gender-specific services can certainly play another major role in establishing and sustaining accessible and culturally appropriate care.1,5,7,27 For example, the well-established Aboriginal community-controlled organisation Danila Dilba Health Service in Darwin demonstrates that gender-specific healthcare services are both a viable and highly accessed service.27 In addition, PHCSs can increase their cultural appropriateness by employing male health practitioners, offering choices to clients regarding the gender of their practitioner and holding men’s-only clinic days or times when men can visit these facilities and communicate with male staff for all their health needs.

Insufficient healthcare resourcing contributes towards the under-use of PHCSs for Aboriginal and Torres Strait Islander people.

In 2009, the National Health and Hospitals Reform Commission recommended an investment strategy for Aboriginal and Torres Strait Islander people’s health, stating this investment should be “proportionate to health need[s], the cost of service delivery, and the achievement of desired outcomes. This requires a substantial increase on current expenditure”.28(p20) Despite this, the 2014–2015 Australian Federal Budget saw aggressive budget cuts to Aboriginal and Torres Strait Islander affairs and health, particularly preventative healthcare, which has significantly affected the extent to which health services can provide them necessary amenities.29

Improving health services will not be the only change required to close the life expectancy gap, as systemic problems of social and economic disparity, discrimination and a lack of empowerment exist. To address this health crisis, changes in economic policy, improvements in education for Aboriginal and Torres Strait Islander males, access to sport and recreation facilities and programs, development of sustainable employment opportunities, a commitment to cultural maintenance, improved engagement with correctional services as well as increased health awareness are all needed. Essentially, addressing healthcare in isolation from sociocultural and economic factors will only ever have a limited effect. Notably, the 2016 Close the Gap Progress and Priorities report outlined many recommendations including: the introduction of ‘Closing the Gap Targets’ to reduce imprisonment; increasing focus on the needs of Aboriginal and Torres Strait Islander people with disabilities; a national inquiry into racism and institutional racism in healthcare; and a reform of the Indigenous Advancement Strategy.30

Despite the many barriers, Aboriginal and Torres Strait Islander men are putting up their hands in a collective show of need1,5,11 to encourage change and to be responsible for leading the way in the fight to turn around generations of disadvantage.

Looking forward

As Marmot suggested, “wider social policy will be crucial to reduction of inequalities in health”.15(p1103) The development of male health policy must rely on the strengths that already exist within Aboriginal and Torres Strait Islander men and communities, rather than the deficit approach that is currently favoured to frame Aboriginal and Torres

Strait Islander health and policy. Building on these strengths should be the cornerstone of future health and development, and an essential investment in the future generations of Aboriginal and Torres Strait Islander people.

A recent systematic review of primary healthcare interventions for Indigenous people with chronic disease highlighted five key enablers and inhibiting factors for program development to affect “upon intervention implementation and/or sustainability within a [primary health care] setting”.31(p9) These included design attributes, workforces, the importance of patient-provider partnerships, the adequate development of clinical pathways and mechanisms to improve access to services. Essentially, these findings should be considered when attempting to implement strategies specific to the needs of Aboriginal and Torres Strait Islander men.

The time has come to collaborate and share knowledge and experiences, to put aside individual egos and to be honest – even about our collective failures to adequately and purposefully engage men. Findings need to be published, including unsuccessful programs, to help others learn from past experiences. We need to stop describing problems and blaming individuals, and start acknowledging Aboriginal and Torres Strait Islander men as the dynamic, essential elements of families, communities and societies they have always been.32 The inherent personal and cultural strengths and attributes of Aboriginal and Torres Strait Islander men must be unshackled, and positive energy directed towards the development of new ways forward by men and their communities, who are empowered and supported to do so.

Funding alone will not close the life expectancy gap. PHCSs can have the latest technology in purpose-built centres, employ some of the best staff available, and provide a plethora of programs, but all of this remains ineffective if the men themselves choose not to use them. In the Torres Strait Islands, there is an expression derived from traditional dance called ‘mark time’, which refers to a dancer stepping in beat with the music while remaining on the spot. Although you are moving, you are also going nowhere. Likewise, PHCSs and key stakeholders need to urgently rethink the future direction of engaging Aboriginal and Torres Strait Islander men and must no longer simply ‘mark time’.


Commentary

Short-term funding is also problematic. It is common for programs or interventions implemented by PHCSs to cease due to funding cuts, despite their outcomes. In fact, as O’Dea explained, “the challenge is to sustain these interventions over the long term in the frequently under-resourced primary health care clinics”.33(p5) Services also have to manage the fallout from defunded programs which includes the loss of engagement, rapport and trust with local Aboriginal and Torres Strait Islander men, not to mention the subsequent turnover of staff that affects continuity of care. Despite the issues of funding, which is often outside of the control of PHCSs, prioritising engagement with local Aboriginal and Torres Strait Islander men (and the broader community) is essential. Engagement is a low-cost exercise for most PHCSs, with the exception of some remote services or those currently severely under-resourced, but it does require a change of attitude.

Aboriginal and Torres Strait Islander men hold the key to their future, as they know what they need and what will get them through the doors. These men need – and want – to take their health in their own hands; however, it is unrealistic to expect them to improve their current situation alone. Proper engagement with and commitment to Aboriginal and Torres Strait Islander men’s health is a logical first step for PHCSs. Ultimately, a collaborative effort from researchers, PHCSs, peak health bodies and government is required to empower Aboriginal and Torres Strait Islander men and their communities to develop and implement new engagement strategies. Sadly, if this is not the case, closing the life expectancy gap will remain nothing more than an advertising slogan.

References1. Canuto K, Wittert G, Harfield S, Brown A. “I feel more

comfortable speaking to a male”: Aboriginal and Torres Strait Islander men’s discourse on utilizing primary health care services. Int J Equity Health. 2018;17:185.

2. Isaacs AN, Maybery D, Gruis H. Mental health services for aboriginal men: Mismatches and solutions. Int J Ment Health Nurs. 2012;21(5):400-8.

3. Isaacs AN, Maybery D, Gruis H. Help seeking by Aboriginal men who are mentally unwell: A pilot study. Early Interv Psychiatry. 2013;7(4):407-13.

4. Adams M, Collins VR, Dunne MP, De Kretser DM, Holden CA. Male reproductive health disorders among Aboriginal and Torres Strait Islander men: A hidden problem? Med J Aust. 2013;198(1):33-8.

5. Wenitong M, Adams M, Holden CA. Engaging Aboriginal and Torres Strait Islander men in primary care settings. Med J Aust. 2014;200:632-3.

6. Hayman N. Medical and Clinical Issues for Indigenous Men. Aborig Isl Health Work J. 2000;24:4-6.

7. Brown A, Blashki G. Indigenous male health disadvantage–linking the heart and mind. Aust Fam Physician. 2005;34(10):813-19.

8. Phillips B, Morrell S, Taylor R, Daniels J. A review of life expectancy and infant mortality estimations for Australian Aboriginal people. BMC Public Health. 2014;14(1):1.

9. Australian Institute of Health and Welfare. Australian Burden of Disease Study: Impact and Causes of Illness and Death in Australia 2011. Australian Burden of Disease Study Series No.: 3. Canberra (AUST): AIHW; 2016.

10. Brown A, Walsh W, Lea T, Tonkin A. What becomes of the broken hearted? Coronary heart disease as a paradigm of cardiovascular disease and poor health among indigenous Australians. Heart Lung Circ. 2005;14(3):158-62.

11. Adams M, Danks B. A Positive approach to addressing indigenous male suicide in Australia. Aborig Isl Health Work J. 2007;31(4):28-31.

12. Eckermann A, Dowd T, Chong E, Nixon L, Gray R, Johnson S. Binan Goonj: Bridging Cultures in Aboriginal Health. 3rd ed. London (UK): Elsevier Health Sciences APAC; 2010.

13. Rix E, Barclay L, Wilson S. Can a white nurse get it? ‘Reflexive practice’ and the non-Indigenous clinician/researcher working with Aboriginal people. Rural Remote Health [Internet]. 2014 [cited 2017 Feb 6];14:2679. Available from: http://www.rrh.org.au/articles/subviewnew.asp?ArticleID=2679

14. Australian Health Ministers’ Advisory Council. Aboriginal and Torres Strait Islander Health Performance Framework 2017. Canberra (AUST): Government of Australia; 2017.

15. Marmot M. Social determinants of health inequalities. Lancet. 2005;365(9464):1099-104.

16. Smith JA, Braunack-Mayer A, Wittert G, Warin M. “It’s sort of like being a detective”: Understanding how Australian men self-monitor their health prior to seeking help. BMC Health Serv Res. 2008;8:1-10.

17. Australian Department of Health and Ageing. National Aboriginal and Torres Strait Islander Health Plan 2013-2023. Canberra (AUST): Government of Australia; 2013.

18. Deeble J. Assessing the Health Service Use of Aboriginal and Torres Strait Islander Peoples. Canberra (AUST): National Health and Hospitals Reform Commission; 2009.

19. Department of the Prime Minister and Cabinet. Access to services compared with need Aboriginal and Torres Strait Islander. In: Health Performance Framework 2014 Report. Canberra (AUST): Government of Australia; 2014.

20. Hayman NE, Wenitong M, Zangger JA, Hall EM. Strengthening cardiac rehabilitation and secondary prevention for Aboriginal and Torres Strait Islander peoples. Med J Aust. 2006;184(8):485.

21. Brown A, Scales U, Beever W, Rickards B, Rowley K, O’Dea K. Exploring the expression of depression and distress in aboriginal men in central Australia: A qualitative study. BMC Psychiatry. 2012;12(1):1-13.

22. Briscoe A. Indigenous men’s health: Access strategy. Aborig Isl Health Work J. 2000;24(1):7-11.

23. Davy C, Harfield S, McArthur A, Munn Z, Brown A. Access to primary health care services for Indigenous peoples: A framework synthesis. Int J Equity Health. 2016;15:163.

24. Ware V. Improving the Accessibility of Health Services in Urban and Regional Settings for Indigenous People. Canberra (AUST): Australian Institute of Health and Welfare; 2013.

25. Hughes CK. Factors associated with health-seeking behaviors of Native Hawaiian men. Pac Health Dialog. 2004;11(2):176-82.

26. Bulman J, Hayes R. Mibbinbah and spirit healing: Fostering safe, friendly spaces for indigenous males in Australia. Int J Men Health. 2011;10(1):6-25.

27. Danila Dilba Health Service. Engaging ATSI Males to our Clinic [PowerPoint Slides]. Canberra (AUST): National Aboriginal Community Controlled Health Organisation; 2017.

28. National Health and Hospitals Reform Commission. A Healthier Future for all Australians: Final Report June 2009. Canberra (AUST): Government of Australia; 2009.

29. Russell L. Impact of the 2014-15 Federal Budget on Indigenous Programs and Services. Sydney (AUST): University of Sydney Menzies Centre for Health Policy; 2014.

30. Close the Gap Campaign Steering Committee. Close the Gap: Progress and Priorities Report 2016. Canberra (AUST): Australian Human Rights Commission; 2016.

31. Gibson O, Lisy K, Davy C, Aromataris E, Kite E, Lockwood C, et al. Enablers and barriers to the implementation of primary health care interventions for Indigenous people with chronic diseases: A systematic review. Implement Sci. 2015;10(1):1-11.

32. Hammond C. Making positive resources to engage Aboriginal men/fathers. Aborig Isl Health Work J. 2010;34(5):23-5.

33. O’Dea K. (2005). Preventable chronic diseases among Indigenous Australians: The need for a comprehensive national approach. Heart Lung Circ. 2005;14(3):167–71.

Correspondence to: Mr Kootsy Canuto, Wardliparingga Aboriginal Research Unit, South Australian Health and Medical Research Institute, Adelaide, South Australia; e-mail: [email protected]



A number of changes have been made to the Australian welfare system over the past two decades that have

significant ramifications for household food security. Of greatest concern are changes made by the Australian Government that have resulted in payments such as Youth Allowance, Newstart and the Parenting Payment to fall below the poverty line (see Table 1). To compensate for low welfare payments, many Australians now rely on the private sector and on charities for food aid and other essentials to mitigate the impacts of austerity. In this commentary, we discuss the hypothesis that food charity is an inexpensive policy alternative to ensuring an adequate standard of living for welfare-reliant households. We conclude by predicting the significant public health ramifications of this approach.

Twenty years in the making

While Australia has a comprehensive welfare system, providing a range of payments and services to individuals and families across the life span, a change instigated by Prime Minister John Howard in the late 1990s has seen the culture of welfare in Australia move from a rights-based entitlement system to a system based on ‘mutual obligation’. Mutual obligation is underpinned by the goal of increasing the economic and social participation of citizens, with the expectation that any person who can should be employed or completing further education.1,2 The effect of this expectation is that individuals on different support categories receive different payment types with stricter entitlement rules based on their theoretical ability to enter the workforce.1 This is problematic; while recipients complete compulsory employment skills and related activities to access income support, the number of job seekers outstrips the number of jobs available.3,4

The Australian welfare system has been further reformed over the past two decades.

One of the more substantive changes has been to the eligibility criteria for a number of welfare payments,5 for example, the Parenting Payment Single. Before 2006, single parents on this payment were able to remain on the benefit until their youngest child turned 16.2 Changes in 2006 saw the child’s age lowered to eight years, with subsequent changes lowering it to six years. If, at this time, parents were still in need of financial assistance, they were encouraged to apply for the Newstart Allowance, a benefit otherwise provided to employment seekers at a fortnightly rate several hundred dollars less than the Parenting Payment, and with a range of mutual obligation requirements.5 There is a body of research that shows that this welfare payment change has had a detrimental effect on families, with some forced to move into insecure housing, often foregoing food.6

Mutual obligation requirements, and differences in indexation, mean that not all welfare recipients and their benefits are treated equally. Like many other countries, welfare payments in Australia are indexed in order to maintain their real value over time; without indexation, the value of benefits

would be eroded by inflation. Until recently, the welfare payment awarded to those on a range of pension payments (including the aged, disability, and carer pension) was indexed to 27.7% of male total average weekly wages.7 This indexation resulted in a larger annual increase in benefit for those receiving these payment types as a way to ensure that pensioners maintained a certain standard of living, relative to the rest of the population. Other benefits, particularly those provided to job seekers and students, have been indexed to the Consumer Price Index (CPI) rather than wages, meaning that these payments have not had a rate increase in real terms for more than 20 years.8

Further compounding these inequalities in payment rates are years of different benchmarking, resulting in a large gap between pension rates and the payments made to potential workforce participants. For example, recipients who are no longer actively encouraged into employment, including those aged over 65 and disability pensioners, receive $926.20 per fortnight for a single, or $1,396.20 per fortnight for a couple.9 These payments put recipients above the Australian poverty line of $866 per fortnight for a single person living alone after housing cost have been paid.10 However, the situation is different for those recipients who are of working age but who are un- or under- employed, with payments falling below the poverty line, putting recipients at an elevated risk of poverty.10 Under these payment types, a single recipient of the Newstart allowance with no children receives $555.70 and a couple receives $1,003.40 per fortnight, while a student on a Youth Allowance payment is eligible for a maximum of $455.20 per

doi: 10.1111/1753-6405.12916

The important role of charity in the welfare system for those who are food insecure Fiona H. McKay,1 Rebecca Lindberg2

1. School of Health and Social Development, Deakin University, Victoria

2. The Institute for Physical Activity and Nutrition (IPAN) and School of Exercise and Nutrition Sciences, Deakin University, Victoria

Table 1: Most common welfare payments in Australia.a

Payment type Recipient Indicative maximum payment for a single

(per fortnight)Aged Pension Aged 65.5 years $926.20b

Disability Pension Have a permanent and diagnosed disability or medical condition, or get a Department of Veterans’ Affairs special rate disability pension due to total and permanent incapacity.

$926.20b

Newstart Between 22 and 65.5 years of age, be looking for employment. $692.90c

AusStudy At least 25 years of age, in full-time education. $592.40c

Youth Allowance Less than 24 years of age, in full-time study or apprenticeship or 16–21 years of age and looking for full-time work.

$592.40c

Parenting Payment Primary carer of child under eight years of age if single, or six years of age if partnered. $990.30d

Notes:a: Source: Australian Department of Human Services. Payment Rates [Internet]. Canberra (AUST): Government of Australia; 2019 [cited 2019 Jan 11].

Available from: https://www.humanservices.gov.au/individuals/services/centrelink. Totals correct at 10 June 2019.b: Maximum rate, may include Energy supplement, Pension supplementc: May include Rent assistanced: May include Rent assistance and Family Tax benefit

Commentary


Commentary

fortnight.9 While these recipients may also be eligible for other benefits including rent assistance or a family tax benefit, even with these additional benefits, all payments fall below the poverty line. The Australian Government justifies these low rates on the grounds that they are intended to provide only short-term relief, and that a more generous rate would discourage the unemployed from seeking employment.11

Coping strategies for a $135 shortfall

According to Australian Council of Social Services (ACOSS), after taking account of housing costs, more than 10% of people live below the poverty line in Australia, closer to 20% for children.10 When looking at the depth of poverty for those below the poverty line, ACCOS suggests that one-in-eight people are, on average, about $135 short per week, with many of those who live below the poverty line relying on some form of welfare benefit as their main source of income.10 As a way to cope with the increasing cost of living and the low welfare entitlement, many welfare recipients, especially Newstart recipients, are turning to charity for assistance. This assistance may come in the form of cash payments to cover the costs of utilities, legal and medical advice, or material and food aid.

Households on low incomes are more likely to be food insecure (see Box 1), as essentials such as housing and utility bills are prioritised over food. Few welfare recipients have the discretionary income to purchase the foods they require from the two main retailers, with research suggesting that the cost of a week’s groceries for a single person is $122, going up to $336 for a couple with children.13 These households employ an array of strategies to feed themselves and their families, with many calling on family and friends for assistance,

skipping meals, or bulking out meals with simple carbohydrates, as well as increasingly accessing a range of food aid charities for all or most of their weekly groceries.6 This is consistent with recent research showing that food insecurity is more prevalent in households where members receive a payment type with a low financial value, such as the Newstart or Parenting Payment as their main source of income.14

The rise of food charity in Australia

Food aid has grown across Australia since the 1990s, with rapid increase over the past decade (see Figure 1). While there are thousands of emergency relief programs providing subsidised and free meals across the country, including those that offer pre-packaged food parcels, school breakfasts, and prepared meals, there are four major national food banking and rescue programs that support this front-line sector. These main organisations – FareShare, OzHarvest, SecondBite and Foodbank Australia – together rescued approximately 19 tonnes of food in 2008–2009 for re-distribution; by 2016–2017, our estimates suggest this sector has almost tripled in size (51,126 tonnes). While publicly available documents are unclear on the precise types of food distributed, or the extent to which these foods meet Australian Dietary Guidelines or are able to meet the cultural and dietary preferences of food-insecure Australians, these tonnes of charitable food are a mix of perishable surplus food from farms, quick service restaurants, bakeries and retailers, as well as shelf-stable products from food manufacturers including pasta, breakfast cereal, tinned foods and sugar-sweetened beverages.

Over the past few years, we have interviewed hundreds of Australians who were using or had used emergency food relief.15-17 What we found is that the term ‘emergency food relief’ is a misnomer. Many of the participants we spoke with had been accessing emergency food relief for years, with some inter-generational users obtaining half or more of their dietary requirements from church groups, community centres and/or outreach services. A range of reasons have led people to rely on these frontline services, including poor health (both chronic and acute), long-term and short-term unemployment, high costs of living, domestic violence and family breakdowns. We have spoken to people of all age groups, and across rural, regional, and metropolitan areas in Victoria,

to try to understand their experiences of the welfare system and how these fit into the charitable food sector. We found that many people perceive the support they receive as a privilege and a gift and assume that their own issues – rather than systemic issues – are the cause of their hunger and inability to meet the cost of living. Our conclusion is that food aid is masking the impacts of a severely ineffective social welfare system.

The Australia Government currently spends $157 billion on welfare, or around $6,500 per person. This includes cash payments, welfare services and unemployment benefits.18 There have been calls in recent years to increase the Newstart allowance, both by a cash amount and also by changing the indexation so that Newstart is more closely aligned with the cost of living in Australia; such a change is projected to lift more than 300,000 people out of poverty, costing the government $3.2 billion dollars in the first year.19 Such an increase would also reduce the burden on charities, even those government funded; currently, the Federal Government spends more than $200 million across 180 relief agencies to help Australians in need, with $4.5 million in funding to three organisations, Foodbank, SecondBite and OzHarvest.20 This federal funding of food relief is, however, modest in comparison to the billions that are projected to be needed to ensure all citizens of Australia have a minimum standard of living. This short-term saving though, will have long-term costs.

Crystal ball gazing

In Australia, thanks largely to entrenched inequality, it can take up to four generations to move out of poverty, that is from the bottom 10% of income to the mean income in society.21 Escaping or avoiding poverty is easier if welfare payments sit close to the poverty line, as is often the case for aged pensioners who may own their home, have no dependent children and have some superannuation. However, for a majority of those receiving Youth Allowance, Newstart, or a Parenting Payment with entitlements that fall well below the poverty line, it is almost impossible to break out of poverty – further embedding and fostering intergenerational poverty.

If there is a not a substantial change in the level of poverty in Australia and the trends to date continue, what are the public health impacts for up to four generations of Australian households who will likely need food charity to cope? Some of the most

Box 1: Definition and Dimensions of Food Insecurity (adapted from reference 12).

Food insecurity exists when individuals, households or even whole communities, have inadequate access to healthy affordable culturally appropriate food.

This experience can have:

Quantitative dimensions. Where individuals run low on food, go without eating, and/or restrict portions.

Qualitative dimensions. Where individuals have limited dietary variety and increase dietary monotony, as well as not being able to meet preferences for healthy and/or cultural foods.

Psychological dimensions. Where individuals feel uncertain/worried about food supplies, have feelings of deprivation and lack of agency/choice/dignity.

Social dimensions. Where individuals are unable to maintain socially prescribed ways of acquiring and/or eating food.


Commentary

concerning include the potential negative effect of the overreliance on charity for health in terms of suboptimal nutrition. Previous research has found that food charities are not in a position to provide sufficient quality or quantity of food, instead providing people with food for approximately three days.22 Our crude snapshot of the growth of the food banking and rescue organisations (Figure 1) suggests that, for the one million estimated food insecure households, there was 51 million kilograms of food available in 2016/17. That is not even one kilogram per food insecure household per week.

Also of concern is the erosion of dignity for many of those who are forced to rely on charity for their essential needs. Research highlights the inherent shame associated with the use of emergency food relief. While the food might be ‘free’, there are hidden ‘costs’ of social stigma and shame, with these feelings having the potential to worsen existing health problems and create further humiliation.23 Finally, of great concern is the large number of children living in poverty in Australia – up to one in five. The research shows that those growing up in poverty are more prone to behavioural problems than other children,24 are more likely to be absent from school,25 and have poorer overall health and wellbeing.26 A reliance on food aid and other charities is unlikely to diminish without some significant changes to the Australian welfare system.

References1. Cook K. Neoliberalism, welfare policy and health: A

qualitative meta-synthesis of single parents’ experience of the transition from welfare to work. Health. 2012;16:507-30.

2. McKenzie HJ, McHugh C, McKay FH. Life on newstart allowance: A new reality for low-income single mothers. J Fam Stud. 2019;25(1):18-33.

3. Australian Bureau of Statistic. 6354.0 - Job Vacancies, Australia, Nov 2018. Canberra (AUST): ABS; 2018.

4. Department of Jobs and Small Business. Vacancy Report. Canberra (AUST): Government of Australia; 2018.

5. Jovanovski N, Cook K. How Australian welfare reforms shape low-income single mothers’ food provisioning practices and their children’s nutritional health. Crit Public Health. 2019. doi.org/10.1080/09581596.2019.1577951

6. McKenzie HJ, McKay FH. Food as a discretionary item: The impact of welfare payment changes on low-income single mother’s food choices and strategies. J Poverty Soc Justice. 2017;25:35-48.

7. Department of Social Services. Social Security Guide. Canberra (AUST): Government of Australia; 2019.

8. Deloitte. Analysis of the Impact of Raising Benefit Rates. Melbourne (AUST): Deloitte; 2018

9. Australian Department of Human Services. Payment Rates [Internet]. Canberra (AUST): Government of Australia; 2019 [cited 2019 Jan 11]. Available from: https://www.humanservices.gov.au/individuals/services/centrelink/age-pension/eligibility/payment-rates

10. Davidson P, Saunders P, Bradbury B, Wong M. 2018 Poverty in Australia. Sydney (AUST): Australian Council of Social Service; 2018.

11. Mendes P. The changing nature of the Australian welfare state: A critical analysis of the ACOSS campaign to increase the Newstart Allowance. Aust J Polit Sci. 2015;50:427-41.

12. Loopstra R. Interventions to address household food insecurity in high-income countries. Proc Nutr Soc. 2018;77:270-81.

13. Australian Securities and Investments Commission. Australian Spending Habits 2016. Gippsland (AUST): Government of Australia; 2018.

14. Temple J, Booth S, Pollard C. Social Assistance Payments and Food Insecurity in Australia: Evidence from the Household Expenditure Survey. Int J Environ Res Public Health. 2019;16(3):455.

15. McKay FH, Dunn M. Food security among asylum seekers in Melbourne. Aust N Z J Public Health. 2015;39:344-9.

16. Lindberg R, Lawrence M, Caraher M. Kitchens and pantries—helping or hindering? the perspectives of emergency food users in Victoria, Australia. J Hunger Environ Nutr. 2017;12:26.

17. Haines BC, McKay FH, Dunn M, Lippi K. The role of social enterprise in food insecurity among asylum seekers. Health Soc Care Community. 2018;26:829-38.

18. Australian Institute of Health and Welfare. Australia’s Welfare 2017. Canberra (AUST): AIHW; 2017

19. Australian Council of Social Service. Raise the Rate 2018. Strawberry Hills (AUST): ACOSS; 2018.

20. Fletcher P, MP. Media Release: Liberal-National Government Invests Over $200 Million to Support Australians Experiencing Hardship. Canberra (AUST): Ministers for the Department of Social Services; 2018

21. Organisation for Economic Co-operation and Development. A Broken Social Elevator? How to Promote Social Mobility. Paris (FRA): OECD; 2018.

22. Bazerghi C, McKay FH, Dunn M. The Role of Food Banks in Addressing Food Insecurity: A Systematic Review. J Community Health. 2016;41:732-40.

23. Garthwaite K. Stigma, shame and’people like us’: An ethnographic study of foodbank use in the UK. J Poverty Soc Justice. 2016;24:277-89.

24. Costello EJ, Compton SN, Keeler G, Angold A. Relationships between poverty and psychopathology: A natural experiment. JAMA. 2003;290:2023-9.

25. Zhang M. Links between school absenteeism and child poverty. Pastor Care Educ. 2003;21:10-17.

26. Brooks-Gunn J, Duncan GJ. The effects of poverty on children. Future Child. 1997;7(2):55-71.

Supporting InformationAdditional supporting information may be found in the online version of this article:

Supplementary Table 1: The rise of food charity in Australia.

Correspondence to: Dr Fiona H. McKay, School of Health and Social Development, Faculty of Health, Deakin University, Locked Bag 20000, Geelong, VIC 3220; e-mail: [email protected]

The data in this figure was extracted principally from annual reports. Where organisations reported on meals, a 500g meal was assumed in order to calculate kilograms and tonnes. Typically financial years were reported on, but occasionally calendar years were used to prepare the figure. For these reasons the data is indicative approximates only.

422 1,250

240

17,573

19,485

571 1,650

702

19,000

21,923

222 2,

399

839

21,000

24,460 540 1,600

1,312

15,626

19,078

665

1,180 3,900

19,600

25,345

560 2,538 5,200

29,000

37,298

545 3,669

4,300

33,000

41,514

661

4,525

10,000

31,390

46,576

846

5,780

11,000

33,500

51,126 CHARITABLE FOOD SECTOR

TONNES PER ANNUM

FareShare OzHarvest SecondBite

Foodbank Australia Sector total (FS+OH+SB+FBA)

The data in this figure was extracted principally from annual reports. Where organisations reported on meals, a 500g meal was assumed in order to calculate kilograms and tonnes. Typically financial years were reported on, but occasionally calendar years were used to prepare the figure. For these reasons the data is indicative approximates only.

422 1,250

240

17,573

19,485

571 1,650

702

19,000

21,923

222 2,

399

839

21,000

24,460

540 1,600

1,312

15,626

19,078

665

1,180 3,900

19,600

25,345

560 2,538 5,200

29,000

37,298

545 3,669

4,300

33,000

41,514

661

4,525

10,000

31,390

46,576

846

5,780

11,000

33,500

51,126 CHARITABLE FOOD SECTOR

TONNES PER ANNUM

FareShare OzHarvest SecondBite

Foodbank Australia Sector total (FS+OH+SB+FBA)

Note:The data in this figure was extracted principally from annual reports. Where organisations reported on meals, a 500g meal was assumed in order to calculate kilograms and tonnes. Typically financial years were reported on, but occasionally

calendar years were used to prepare the figure. For these reasons the data is indicative approximates only.

Figure 1: Charitable food sector tonnes per annum.


Atrial fibrillation (AF) is an established antecedent for stroke and other forms of cardiovascular disease.1

Cardiovascular disease is the main cause of mortality among Aboriginal and Torres Strait Islander people in Australia (hereafter Aboriginal people).2

Handheld electrocardiogram (ECG) devices have been recommended internationally as preferred screening tools for the diagnosis of atrial fibrillation.3 These devices can attach to mobile phones, are typically referred to as iECG and have demonstrated effectiveness and acceptability in clinical and community settings such as dental vans, pharmacies4 and in general practice.5,6 Both systematic and opportunistic screening of adults for AF increased the detection rate of new cases compared with routine practice and opportunistic screening has greater cost-effectiveness than systematic screening.7 Competence and confidence of nurses facilitated iECG screening whereas a lack of staff availability and technical issues obstructed screening.6 A recent study found the majority of iECG-screened participants were satisfied with the device, finding it easy to use without restricting activities or causing anxiety.8

A scoping review on AF in Indigenous populations internationally found higher AF hospitalisation rates relative to other populations and occurrence at younger ages and with more comorbidity.9 National data reports the AF rate, as either a principal or additional diagnosis, was 1.4 times as

high for Aboriginal Australians as for other Australians.10 One study in Western Australia reporting on AF as the primary outcome,11 conducted in a hospital inpatient setting, found higher rates in comparison with non-Aboriginal counterparts. A further study of hospital admissions found AF to occur

Feasibility and acceptability of opportunistic screening to detect atrial fibrillation in Aboriginal adultsRona Macniven,1,2 Josephine Gwynn,1,2 Hiroko Fujimoto,1 Sandy Hamilton,3 Sandra C. Thompson,3 Kerry Taylor,4 Monica Lawrence,5 Heather Finlayson,6 Graham Bolton,6 Norman Dulvari,7 Daryl C. Wright,8 Boe Rambaldini,1 Ben Freedman,1,2 Kylie Gwynne1

1. Faculty of Medicine and Health, Sydney Medical School, Poche Centre for Indigenous Health, The University of Sydney, New South Wales2. Charles Perkins Centre D17, The University of Sydney, New South Wales3. Poche Centre for Indigenous Health, School of Indigenous Studies, The University of Western Australia, Crawley, Western Australia4. Poche Centre for Indigenous Health, Alice Springs, Northern Territory 5. Poche Centre for Indigenous Health, Flinders University of South Australia, Adelaide, South Australia 6. Brewarrina Multipurpose Service, Brewarrina, New South Wales7. Albury Wodonga Aboriginal Health Service, Glenroy, New South Wales8. Tharawal Aboriginal Corporation, Airds, New South WalesCorrespondence to: Dr Rona Macniven, Faculty of Medicine and Health, Sydney Medical School, Poche Centre for Indigenous Health, Rm 224, Edward Ford Building A27,

The University of Sydney, NSW 2006; e-mail: [email protected]: August 2018; Revision requested: December 2018; Accepted: March 2019The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,

provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Aust NZ J Public Health. 2019; 43:313-18; doi: 10.1111/1753-6405.12905

Abstract

Objective: Examine the feasibility and acceptability of an electrocardiogram (ECG) attached to a mobile phone (iECG) screening device for atrial fibrillation (AF) in Aboriginal Controlled Community Health Services (ACCHS) and other community settings.

Methods: Semi-structured interviews were conducted with ACCHS staff in urban, rural and remote communities in three Australian states/territories. Quantitative and qualitative questions identified the enabling factors and barriers for staff and Aboriginal patients’ receptiveness to the device. Mean quantitative scores and their standard deviation were calculated in Microsoft Excel and qualitative questions were thematically analysed. Results: Eighteen interviews were conducted with 23 staff across 11 ACCHS. Quantitative data found staff were confident in providing iECG screening and managing the referral pathway, and thought the process was beneficial for patients. Qualitative data highlighted the usefulness of the device to undertake opportunistic screening and acceptability in routine practice, and provided opportunities to engage patients in education around AF. Conclusion: The iECG device was well accepted within ACCHSs and was feasible to use to screen for AF among Aboriginal patients.

Implications for public health: The device can be used in clinical and community settings to screen Aboriginal people for atrial fibrillation to help reduce rates of stroke and other cardiovascular diseases.

Key words: indigenous health, rural and remote health, primary health care, screening

INDIGENOUS HEALTH


Macniven et al. Article

in Aboriginal people 15 years earlier than in non-Aboriginal people, as well as higher overall rates in comparison to non-Aboriginal people and higher long-term mortality rates among the Aboriginal patients.12 However, little data on the prevalence of AF in Australian Aboriginal communities is available.

Given the limited available and inconsistent data on AF, and acknowledged importance as a risk factor for cardiovascular disease, early screening and detection holds promise in improving clinical and population outcomes. Early screening and detection usually occurs in primary health care settings and for Aboriginal people the Aboriginal Community Controlled Health Services (ACCHS) play a pivotal role in delivering health care in urban, regional and remote community settings. A key service performed by ACCHS is adult health checks, which are recommended every two years for Aboriginal adults aged 15-54 years and comprise screening of multiple health variables and risk factors including cardiovascular disease and health promotion.13 The checks are typically conducted by Aboriginal Health Workers (AHWs) and pathways exist for subsequent referral to relevant health professionals and specialists, although these can be variable depending on location, remoteness and system factors.14 The feasibility and acceptability of the iECG device to health staff working within the Aboriginal Community Controlled Health sector as a tool for determining AF prevalence among Aboriginal adults is unknown. AF is not only often asymptomatic,15 but Aboriginal Australians are more likely to delay accessing the healthcare system until later in a disease process or may not seek timely help in emergencies due to issues such as fear, racism and service access.16,17 However, ACCHS can help overcome these barriers and has achieved better health outcomes for Aboriginal people than mainstream services.17 This study aims to determine the feasibility and acceptability to health staff of opportunistic screening through the use of an iECG device to detect AF among Aboriginal adults within community controlled health settings.

Methods

Study design The study adopted a mixed methods design and took place during 2017. It was

co-designed and implemented with the participating communities and community services. The full protocol of a study to conduct opportunistic screening for AF among Aboriginal adults that included the examination of the feasibility and acceptability of the device has been described elsewhere.18 The study was approved by Aboriginal Health and Medical Research Council (AHMRC) of NSW (1135/15), the Western Australian Aboriginal Health Ethics Committee (WAAHEC) (HREC706) and the Central Australian Human Research Ethics Committee in the NT.

ParticipantsInterviewees were 18 ACCHS staff (Aboriginal and non-Aboriginal), including AHWs and registered nurses (RNs). They were purposefully sampled from the 11 ACCHS involved in the study due to their specific involvement in the full study as the ACCHS contact personnel and/or had responsibility through their professional role in conducting iECG screening with patients. Participating ACCHS were located in urban (major cities, N=2), regional (N=7) and remote (N=2) areas 19 within New South Wales (NSW; N=7), Western Australia (WA; N=3) and the Northern Territory (N=1). At the commencement of the study, staff received face-to-face training at their ACCHS in the use of the device, consent processes for patients, cardiovascular health promotion and treatment, data collection and the clinical pathway for patients with a non-normal result. The screening process involved using a dedicated smartphone with an iECG device to screen a patient and using the device software to transmit the ECG result to the study database via the telephone data network using an activated sim card. Internet connectivity was required to transmit the results to a secure website for data storage but was not required for the screening itself. Screening occurred from June 2016 to February 2018 and the iECG device

was retained by ACCHS at the end of the full study for ongoing use in routine practice. All interviewees provided informed verbal and written consent to take part.

MeasuresThe study measures were developed in partnership with Aboriginal investigators and communities in each of the three states/territories that the study was conducted in to ensure their contextual integrity. Face validity of the measures were also determined through this co-design process. Quantitative and qualitative questions identified the enabling factors and barriers for AHWs and other ACCHS staff using the iECG in their roles and Aboriginal patients’ receptiveness to the iECG as perceived by the iECG screeners. Semi-structured interviews included seven quantitative five-point Likert question items (strongly agree, agree, neither agree nor disagree, disagree, strongly disagree). Staff were asked whether they felt they were provided with sufficient training for the study and their confidence in providing iECG screening, and in managing the pathway and treatment plan for patients who required follow-up. Interviewees were also asked whether they believed patients (who were subsequently diagnosed with AF) followed their treatment plan for the condition, whether the process was beneficial for participants (who were screened) and the time commitment required for the study. Five qualitative open-ended questions asked interviewees what they thought was useful for patients, how patients responded to screening, what interviewees liked and found challenging about the process and any suggestions for improvement (Box 1). Potential participants were able to choose to take part through a group interview if preferred.

Procedure Data collection occurred during the second half of 2017. Purposefully sampled interviewees in each ACCHS were approached by the researcher coordinating the study by telephone, face-to-face or by email, as appropriate, and invited to participate and/or suggest other suitable interviewees within the ACCHS. Interviews were conducted at a time and location convenient to the interviewees by one of five researchers (RM, HF, SH, KT, HF) with the exception of two interviews, which were conducted by two researchers following the community’s

Box 1: Qualitative open-ended questions.1. What aspects of the screening process do you think were

useful for patients?2. How did you feel the patients responded to the screening

process?3. Were there any things you liked about the screening

process? Did you find it worked well?4. Were there any aspects of the screening process that you

did not like or were difficult for you? If so, can you describe these in more detail?

5. Can you suggest any ways of improving the screening process?


Indigenous Health Opportunistic screening to detect atrial fibrillation in Aboriginal adults

request. Each of the researchers had tertiary training in qualitative research methods and prior experience of conducting interviews. Face-to-face interviews were conducted where feasible and typically took place in private in the staff member’s closed office, or interviews occurred by telephone where distance, travel or time constraints existed. Interviews were audio-recorded and transcribed verbatim, other than in the Northern Territory where no audio recording occurred for cultural reasons in the relevant remote community. In this interview, the participants and interviewer talked as a small group in an informal setting with respect to the participant cultural preferences, responses were agreed on among the group and data were recorded through the detailed field notes of the interviewer. Recruitment continued until no new information emerged from interviews. Interview duration was typically 30 minutes (minimum 15 minutes; maximum 60 minutes). Quantitative responses were recorded in a Microsoft Excel spreadsheet and qualitative responses were entered into QSR NVivo software version 10 (QSR International) for data management. Community members from each participating site were screened for AF using the iECG device, and details of this study component are provided in a separate forthcoming publication. In summary: participants were recruited from two states (NSW N=419; WA N=161) and the NT (N=39) and across three geographical location types (remote N=41, regional N=459; and urban N=119); 619 iECG screens were collected; and results were recorded as either Unclassified, Normal or Possible AF and referred locally for confirmatory 12 lead ECG where an Unclassified or Possible AF result was reported.

Data analyses For the quantitative questions, the five-level Likert question items were scored from 5=strongly agree to 1=strongly disagree. Mean scores and their standard deviation were calculated in Microsoft Excel. For the qualitative questions, thematic analysis20 was used to capture key themes around feasibility and acceptability of the device within ACCHS, including perceived barriers and enablers to its use. An inductive approach was used to code the data transcripts and identify frequently occurring themes that emerged from the coded content. Data analyses were conducted and cross-checked by two

researchers who had conducted the majority of the interviews. A member-checking21 process was undertaken through consultation with the Aboriginal investigators and community members to finalise the results and include their correct interpretation.

Results

Information on the geographical location and occupational characteristics of the interviewees as well as the type of interview is presented in Table 1.

Quantitative dataMean score and standard deviation of five-item Likert scale questions are presented in Table 2. Overall, interviewees reported they were provided with sufficient training for their role in the screening study and did not believe that the study took too much of their time. They expressed confidence in providing an iECG screening and in managing the referral follow-up pathway for patients where required. Most agreed that the process of screening to detect and manage AF in Aboriginal patients, regardless of the diagnosis, was beneficial for patients although less were convinced that most patients diagnosed with AF would follow their treatment plan.

Qualitative data Qualitative data outlining four main themes of feasibility; acceptability; use as an educational tool; barriers to use are described with exemplary quotes in Figures 1-4.

Theme 1: Feasibility of the iECG for ACCHS screening

The majority of interviewees spoke about the usefulness of the device to undertake opportunistic screening in their roles. Specifically, they found the iECG simple, quick and easy to use and liked its portable nature.

Table 1: Interviewee characteristics (N=18).Characteristic NumberState/ Territory NSW

WA

NT

12

4

2Geographical classification Urban

Regional

Remote

2

13

3Professional role RN

AHW

AHW/RN

AHW/Manager

Manager

9

4

2

2

1Interview type Face to face

Telephone

12

6

Table 2: Mean score and standard deviation of five-item Likert scale questions (5=strongly agree; 1=strongly disagree).Question Percent ‘agree’ or

strongly agree’1. ‘I was provided with sufficient training for my role in this study’ 88.9%2. ‘I was confident in providing an iECG screening’ 100.0%3. ‘I was confident in managing the referral pathway for patients who required a confirmation ECG by the GP’ 94.4%4. ‘After a patient was diagnosed with Atrial Fibrillation, I was confident helping support their treatment plan’ 94.4%5. ‘I believe most patients followed their treatment plan’ 61.1%6. ‘I believe this process of screening to detect and manage AF in Aboriginal patients, regardless of the diagnoses, was beneficial for patients’

83.3%

7. ‘I believe that this study took too much of my time’ 16.7%

Several interviewees also thought they could use the device in the future as part of standard adult health checks for Aboriginal patients.

Theme 2: Acceptability of the iECG among staff and patients

Most interviewees described how the device was acceptable for use with patients in their routine practice and enhanced diagnostic and other screening processes for early intervention and chronic disease management. They spoke of how the device was particularly useful beyond clinical settings in the community where it seemed to have greater acceptability among patients and provided more flexible options. This greater acceptability related to comfort and anonymity of community-based screening, with participants indicating how some patients could feel uncomfortable with perceived implications of screening results in the clinical setting. While all interviewees described how the device was generally well received by patients, some interviewees


communities, the importance of AHWs in providing culturally-competent care in partnership with other health professionals and their expertise was described, taking account of complexities around community factors. While patients appeared to be most comfortable with AHWs, conflicting cultural issues relating to traditional views regarding the heart, AHW professional authority and caution around heart health and cultural boundaries relating to their professional and personal roles in the community were described by remote AHWs. These issues related to the management of patients and relationships with community members and the interactions of these factors with the roles of non-Aboriginal health professionals whose expertise and lack of cultural conflict could provide clarity for patients and support the initial screening by AHWs.

Theme 3: Use as an educational tool

Several interviewees described how the iECG device provided unique opportunities to engage patients in education around AF and their heart, and to empower patients to find out more about their heart health. Some staff also spoke of how using the device for screening led them to want to learn more about AF and cardiovascular disease themselves in their professional role.

Theme 4: Barriers and enablers to using the device existed

Some challenges or barriers were also described. These were a mixture of cultural and logistical issues, but many of the latter related to the requirements of the study protocol. A number of logistical issues were described across urban, regional and remote communities relating to the internet connectivity required to upload data collected by the device. Interviewees outlined some challenges in ensuring patients could receive appropriate follow-up within the required time period and that patients returned for follow-up of unclassified or positive readings, particularly when the device was used in the community setting. The consent process and paperwork requirements of the study were also described by interviewees as a barrier to both their and their patients’ participation. Interviewees spoke about how training procedures delivered by the researchers and onsite assistance helped to overcome some of the barriers described. They also highlighted how educational resources would also help overcome barriers around knowledge of AF and the screening process.

Discussion

Our findings demonstrate the feasibility and acceptability of a portable, handheld iECG device for the screening and detection of AF within ACCHS and related community settings across a range of urban, regional and remote areas in Australia.

Several studies have previously established the feasibility and acceptability of a range of screening tools among Aboriginal populations22,23 as well as iECG devices in mainstream populations.5,7 This study is the first to examine these factors in relation to the iECG screening tool in Aboriginal populations. Our quantitative data found that both Aboriginal and non-Aboriginal ACCHS staff were confident in providing iECG screening and managing the referral follow-up pathway, and felt the process was beneficial for patients. The feasibility and acceptability of screening tools within the ACCHS setting has been described previously23 and our findings confirm the usefulness of screening in this setting across urban, rural and remote settings. Interviews confirmed the feasibility and usefulness of iECG screening in the community setting despite several barriers. These findings are consistent with the results

Figure 1: Qualitative interviewee quotes: Feasibility.

“It worked very well, it was quick, results immediately, just so simplified and when you’re in outreach and there’s only 2 nurses, this sort of technology is good for us, especially when time management is difficult.” [Remote AHW/RN]

“It’s portable, provides opportunistic testing away from the main clinic, gave the patients a visual that they found interesting.” [Regional RN]

“We really liked this because we can make this part of their adult health check.” [Remote AHW]

“I can include the iECG in the regular routine with Blood Pressure etc.” [Remote AHW/RN]

Figure 2: Qualitative interviewee quotes: Acceptability.

“A very good tool for me as it was a fast way of diagnosis. It sped up the diagnostic procedure and sped up the treatment pathway.” [Urban RN]

“If we were out in the field, on outreach so could have family do it where they felt more comfortable in their own home. Wherever we did it they were happy and it didn’t take up too much time.” [Regional RN]

“There was excitement through to the other end of the scale – absolute fear about what the iECG would actually disclose.” [Remote AHW]

“I liked how it was simple, 30 seconds, much easier than a regular 12 lead ECG so that was really good.” [Regional RN]

“I liked the fact that it a good tool for generating yarning about your heart and what was normal.” [Urban RN]

“It’s better to involve the nurse and doctor to talk about the treatment plan, because AHWs don’t really know enough about this AF. Also some people don’t like health workers involved because they worry about privacy because we are part of the community. Even though we know we have to keep things confidential, some people don’t want to see us.” [Remote AHW]

Figure 3: Qualitative interviewee quotes: Use as an educational tool.

“Opening that dialogue around heart health, whether it was having that conversation if they had a preexisting condition, asking about what medication they are on. It was a good engagement tool for people who would otherwise not be engaged in that kind of conversation.” [Regional AHW/RN]

“I would like more training in AF to be up to date about that heart problem and having cardiac resources around AF that are culturally appropriate and respectful and in language.” [Remote AHW]

Figure 4: Qualitative interviewee quotes: Barriers and enablers to use.

“Internet coverage is very slow for us, staff don’t always have email, we shouldn’t have to be faced with this stuff. I could personally see that if the ECG was ok but if I couldn’t upload it to Dr in town I would have to have a backup plan.” [Remote AHW/RN]

“The difficulty is that the 24-hour requirement to be followed up by the Dr is hard here as the Dr only comes here once a week.” [Remote AHW/RN]

“The issue was getting them back to clinic for any follow up required.” [Regional RN]

“Sometimes it [the device] couldn’t decide the diagnosis when it was clearly sinus rhythm and that was a bit frustrating.” [Urban RN]

“I think from screening tool without needing to have all the data entered into the phone… I think that would be really useful tool, just like taking blood pressure, taking pulse. I think it’s just because this is a study that you have to collect data on the person, and getting consent all that sort of issues.” [Regional RN]

“I think the best time was when you guys were here being a part of it, that was really valuable. It’s difficult to overcome the barriers we did face, I can’t think of a way other than having someone at the service that could solely do it.” [Regional RN]

“Having cardiac resources around AF that are culturally appropriate and respectful and in language.” [Remote AHW]

described worry or anxiety expressed by patients in anticipation of what the device might diagnose, or in response to a positive or unclassified result for AF.

There were also high levels of acceptability of the device and screening process among ACCHS staff, despite some challenges in establishing the project into their work patterns. Specifically, interviewees spoke of the time benefits of the iECG compared to a full ECG test and how it provided an opportunity to discuss broader aspects that promoted health with patients. In remote



of the screening process that achieved recruitment of 619 Aboriginal patients aged 45 years and over across the 11 ACCHS, reported in detail separately.18

Community settings, such as events and home visits, gave opportunities to engage with patients’ family members and the wider local Aboriginal community and to undertake broader preventive screening beyond the traditional clinical service. Given known barriers experienced by Aboriginal people in engaging with the health sector for health promotion,24 iECG screening in community settings enhances opportunities for health promotion and engagement. This is particularly valuable in rural and remote settings where disparities in health service provision exist25 as well as among Aboriginal populations broadly.26 Our qualitative data also highlighted the usefulness of the device to undertake opportunistic screening and its acceptability in routine practice, specifically in providing an easy-to-use resource to assist cardiac diagnosis that was favoured by patients. These are positive enablers to achieving greater access to, and update of, preventive and treatment services within primary care, known to be linked with better health outcomes.26 A couple of interviewees suggested incorporating the iECG screening tool into Aboriginal adult health checks, which provide comprehensive health assessments and enable opportunities to provide health advice and risk factor modification.27 This would align with the interviewee recommendations of the potential for the iECG device to be used as an educational tool for wider health promotion and a component of empowerment through improving health literacy28; we recommend the process of formal addition to adult health checks13 be investigated further.

Several barriers to screening were described by ACCHS staff. These included a lack of time to complete screening, logistical and technical issues with the device, implementation of the protocol such as a lack of opportunity for timely follow-up with a doctor, and internet connectivity issues. Some of these barriers related only to the study protocol, not the use of the device itself, such as the requirement to upload data to the study and so would not be present in ongoing use or might be overcome with an upgraded smartphone. The 24-hour time period required for follow-up with a doctor was reported as challenging to

achieve in remote areas where a doctor may only provide weekly community visits, yet the screening process was considered to have overall benefits in identifying patients potentially at risk of AF to ACCHSs staff that may not otherwise have occurred. Few interviewees reported actual barriers to patient engagement with the device generally positively received. Only a small number of participants, mainly AHWs in remote areas, expressed patient concern, fear and confusion related to the device and its results. Strategies described to overcome these reported barriers included supporting the positive interpersonal relationships that provide trust between AHWs and patients.29 The barrier related to obtaining participant consent and related research paperwork would not exist outside of the research processes. The protocol for this study18 stated that the local Aboriginal healthcare workforce would collect the data, however, this was not the case in all participating sites (Table 1). This was due to a variety of reasons: AHW availability and role in the clinical setting; hesitancy from a few AHWs to collect the data, requiring more support and time to become confident in the use of the iECG than was feasible for the research team to deliver at the site; cultural issues around privacy; and concern about giving ‘bad news’ to participants.However, the final study sample gave a broader perspective of the ACCHS screening process including managers as well as AHWs and RNs. Strengths of this study include the examination of a novel device in both a clinic and community setting, the multi-site recruitment of interviewees across three states/territories, urban, regional and remote areas and across a range of professional roles within ACCHS. The study was designed and implemented in close collaboration with communities and community services, an approach vital to achieving culturally relevant acceptance and engagement in cardiac care.30 While the study participants gave many positive perspectives of the screening, some difficulties were also described such as the difficulty in obtaining consent and ensuring follow-up took place, as well as fear of the

device and its results among some people. Overall, we therefore consider the study to give accurate data about the feasibility and acceptability of the device for screening. Our findings summary and suggestions for future use are outlined in Table 3.

Conclusion

Overall, the iECG device was well accepted within ACCHS and was feasible to use to screen for AF among Aboriginal patients in both clinical and community settings. A number of barriers to screening within these settings were identified, but solutions to overcome these barriers emerged and the use of the device created interest in relevant training and educational resources. Screening through the iECG device is a feasible and acceptable way to look for untreated AF in a community or clinic setting. It has the potential, if widely utilised and followed up, to reduce the unfortunate outcome of stroke that is experienced at a younger age and higher rates by Aboriginal people, and to contribute to improving the health literacy and health of Aboriginal people.

Acknowledgements

In memory of Norman Dulvari whose contribution to this research was invaluable and without whom would not have been the success it has.

The authors acknowledge the Aboriginal Health and Medical Research Council of New South Wales and the support and participation of the communities involved in the design and conduct of the study and the participants for contributing their time and perspectives.

References1. Allan V, Honarbakhsh S, Casas J-P, Wallace J, Hunter

R, Schilling R, et al. Are cardiovascular risk factors also associated with the incidence of atrial fibrillation?: A systematic review and field synopsis of 23 factors in 32 population-based cohorts of 20 million participants. Thromb Haemost. 2017;117(5):837-50.

2. Australian Institute of Health and Welfare. Cardiovascular Disease, Diabetes and Chronic Kidney Disease—Australian Facts: Aboriginal and Torres Strait Islander People. Cardiovascular, Diabetes and Chronic Kidney Disease Series No.: 5. Canberra (AUST): AIHW; 2015.

Table 3: Future suggestions for iECG device use.Finding Future suggestionTheme 1: Feasibility of the iECG for ACCHS screening Provide up to date devices in ACCHSsTheme 2: Acceptability of the iECG among staff and patients Include in Indigenous adult health checksTheme 3: Use as an educational tool Provide further training to ACCHSs staff on the deviceTheme 4: Barriers and enablers to using the device existed Streamline screening & follow-up processes

Indigenous Health Opportunistic screening to detect atrial fibrillation in Aboriginal adults


3. Freedman B, Camm J, Calkins H, Healey JS, Rosenqvist M, Wang J, et al. Screening for atrial fibrillation. A report of the AF-SCREEN International Collaboration. Circulation. 2017;135(19):1851-1867.

4. Lowres N, Neubeck L, Salkeld G, Krass I, McLachlan AJ, Redfern J, et al. Feasibility and cost-effectiveness of stroke prevention through community screening for atrial fibrillation using iPhone ECG in pharmacies. The SEARCH-AF Study. Thromb Haemost. 2014;111(6): 1167-76.

5. Orchard J, Freedman S, Lowres N, Peiris D, Neubeck L. iPhone ECG screening by practice nurses and receptionists for atrial fibrillation in general practice: the GP-SEARCH qualitative pilot study. Aust Fam Physician. 2014;43:315-19.

6. Orchard J, Lowres N, Freedman SB, Ladak L, Lee W, Zwar N, et al. Screening for atrial fibrillation during influenza vaccinations by primary care nurses using a smartphone electrocardiograph (iECG): A feasibility study. Eur J Prev Cardiol. 2016;23 Suppl 2:13-20.

7. Moran PS, Teljeur C, Ryan M, Smith SM. Systematic screening for the detection of atrial fibrillation. Cochrane Database Syst Rev. 2016;(6):CD009586.

8. Halcox JPJ, Wareham K, Cardew A, Gilmore M, Barry JP, Phillips C, et al. Assessment of remote heart rhythm sampling using the alivecor heart monitor to screen for atrial fibrillation: The REHEARSE-AF Study. Circulation. 2017;136(19):1784-94.

9. Katzenellenbogen JM, Woods JA, Teng T-HK, Thompson SC. Atrial fibrillation in the Indigenous populations of Australia, Canada, New Zealand, and the United States: A systematic scoping review. BMC Cardiovasc Disord. 2015;15:87.

10. The Australian Commission on Safety and Quality in Health Care and the Australian Institute of Health and Welfare. The Second Australian Atlas of Healthcare Variation. Sydney (AUST): ACSQHC; 2017.

11. Wong CX, Brooks AG, Cheng Y-H, Lau DH, Rangnekar G, Roberts-Thomson KC, et al. Atrial fibrillation in Indigenous and non-Indigenous Australians: A cross-sectional study. BMJ Open. 2014;4(10):e006242.

12. Katzenellenbogen JM, Teng THK, Lopez D, Hung J, Knuiman MW, Sanfilippo FM, et al. Initial hospitalisation for atrial fibrillation in Aboriginal and non-Aboriginal populations in Western Australia. Heart. 2015;101(9):712-19.

13. Spurling GK, Hayman NE, Cooney AL. Adult health checks for Indigenous Australians: The first year’s experience from the Inala Indigenous Health Service. Med J Aust. 2009;190(10):562-4.

14. Macniven R, Hunter K, O’Brien C, Jeffries TL Jr, Shein G, Saxby A, et al. Primary, specialist and allied health services delivered to rural and remote communities and their access by Aboriginal people: Protocol for a mixed methods study. JMIR Res Protoc. 2019;8(2):e11471.

15. European Heart Rhythm Association, European Association for Cardio-Thoracic Surgery, Camm AJ, Kirchhof P, Lip GY, Schotten U, et al. . Guidelines for the management of atrial fibrillation: The Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC). Eur Heart J. 2010;31(19):2369-429.

16. Australian Institute of Health and Welfare. The Health and Welfare of Australia’s Aboriginal and Torres Strait Islander Peoples: 2015. Canberra (AUST): AIHW; 2015.

17. Panaretto KS Wenitong M, Button S, Ring IT. Aboriginal community controlled health services: Leading the way in primary care. Med J Aust. 2014;200(11):649-52.

18. Gwynne K, Flaskas Y, O’Brien C, Jeffries TL, McCowen D, Finlayson H, et al. Opportunistic screening to detect atrial fibrillation in Aboriginal adults in Australia. BMJ Open. 2016;6(11):e013576.

19. Australian Bureau of Statistics. 1270.0.55.005 - Australian Statistical Geography Standard (ASGS): Volume 5 - Remoteness Structure, July 2011. Canberra (AUST): ABS; 2013.

20. Braun V, Clarke V. Successful Qualitative Research: A Practical Guide for Beginners. London (UK): Sage Publications; 2013.

21. Lincoln YG, Guba EG. Naturalistic Inquiry. Newbury Park (CA): Sage Publications; 1985.

22. Garvey G, Thewes B, He VFY, Davis E, Girgis A, Valery PC, et al. Indigenous cancer patient and staff attitudes towards unmet needs screening using the SCNAT-IP. Support Care Cancer. 2016;24(1):215-23.

23. Noble NE, Paul CL, Carey ML, Sanson-Fisher RW, Blunden SV, Stewart JM, et al. A cross-sectional survey assessing the acceptability and feasibility of self-report electronic data collection about health risks from patients attending an Aboriginal Community Controlled Health Service. BMC Med Inform Decis Mak. 2014;14(1):1-8.

24. Aspin C, Brown N, Jowsey T, Yen L, Leeder S. Strategic approaches to enhanced health service delivery for Aboriginal and Torres Strait Islander people with chronic illness: A qualitative study. BMC Health Serv Res. 2012;12:143.

25. Gruen RL, Bailie RS, Wang Z, Heard S, O’Rourke IC. Specialist outreach to isolated and disadvantaged communities: A population-based study. Lancet. 2006;368(9530):130-8.

26. Gwynne K, Jeffries T, Lincoln M. Improving the efficacy of healthcare services for Aboriginal Australians. Aust Health Rev. 2018. doi: 10.1071/AH17142.

27. DiGiacomo M, Abbott P, Davison J, Moore L, Davidson PM. Facilitating uptake of aboriginal adult health checks through community engagement and health promotion. Qual Prim Care. 2010;18(1):57-64.

28. Vass A, Mitchell A, Dhurrkay Y. Health literacy and Australian Indigenous peoples: An analysis of the role of language and worldview. Health Promot J Austr. 22(1):33-7.

29. Abbott P, Gordon E, Davison J. Expanding roles of Aboriginal health workers in the primary care setting: Seeking recognition. Contemp Nurse. 2007;26(1):66-73.

30. Tibby D, Corpus R, Walters DL. Establishment of an innovative specialist cardiac indigenous outreach service in rural and remote Queensland. Heart Lung Circ. 2010;19(5):361-6.



Anaemia is a global health issue that particularly affects women and young children.1 Causes of anaemia include

nutrient deficiencies – lack of folate, vitamin B12 and/or iron-infections, inflammation and genetic conditions.1

Anaemia in the first thousand days, from conception to age two years, can compromise the health of mothers and their pregnancy outcomes as well as the health and early childhood development of their children.2 The most common cause of anaemia in early life is iron deficiency, as iron requirements increase due to expanding blood supply and other tissue growth.1,3 Prevention of iron deficiency and/or anaemia is necessary for optimal child health and development.2,4

During the first months of life, the main source of iron is not breast milk or infant formula but the iron provided to the baby by its mother during the last ten weeks of pregnancy.5 Iron status of an infant at birth reflects the iron status of the mother during pregnancy.6 In low-income settings, anaemia of a mother in pregnancy is a strong predictor of anaemia in the early life of her child.7 Birthweight matters, as smaller babies have smaller iron endowment.6 Cord clamping practices at birth are also important, as delayed clamping can increase body iron of the newborn by about 30% compared to early clamping.8,9

A baby of healthy birthweight, born at full term to a well-nourished mother, typically has sufficient iron for the first six months of

life.6,8 After this, nutrient-dense solid foods rich in iron are required.10,11 Traditionally, Aboriginal and Torres Strait Islander Australians consumed many iron-rich foods such as insects, shellfish, animal blood and organs.12-14 Today, however, Aboriginal and Torres Strait Islander people consume diets that are less nutritious than the diets of other Australians.15 These nutrient-poor diets, often commencing from early life, are associated

with high rates of food insecurity compared to other Australians, especially among those living in remote locations.15-17 Food insecurity increases the risk of anaemia among women and their children.18

Anaemia among Aboriginal and Torres Strait Islander children is a long-standing concern in remote communities in the Northern Territory and Western Australia.19-22 A recent Northern Territory report showed 29.0% of children

Anaemia in early childhood among Aboriginal and Torres Strait Islander children of Far North Queensland: a retrospective cohort study Dympna Leonard,1 Petra Buttner,1 Fintan Thompson,1 Maria Makrides,2,3 Robyn McDermott1

1. Australian Institute of Tropical Health and Medicine, College of Public Health, Medical and Veterinary Sciences, James Cook University, Queensland 2. Healthy Mothers, Babies and Children, South Australian Health and Medical Research Institute3. Discipline of Paediatrics, School of Medicine, The University of Adelaide, South AustraliaCorrespondence to: Ms Dympna Leonard, Australian Institute of Tropical Health and Medicine, College of Public Health, Medical and Veterinary Sciences, James Cook

University, PO Box 6811, Cairns, QLD 4870; e-mail: [email protected]: December 2018; Revision requested: March 2019; Accepted: April 2019The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,



Abstract

Objective: Early childhood anaemia affects health and neurodevelopment. This study describes anaemia among Aboriginal and Torres Strait Islander children of Far North Queensland.

Methods: This retrospective cohort study used health information for children born between 2006 and 2010 and their mothers. We describe the incidence of early childhood anaemia and compare characteristics of children and mothers where the child had anaemia with characteristics of children and mothers where the child did not have anaemia using bivariate and multivariable analysis, by complete case (CC) and with multiple imputed (MI) data.

Results: Among these (n=708) Aboriginal and Torres Strait Islander children of Far North Queensland, 61.3% (95%CI 57.7%, 64.9%) became anaemic between the ages of six and 23 months. Multivariable analysis showed a lower incidence of anaemia among girls (CC/MI p<0.001) and among children of Torres Strait Islander mothers or both Aboriginal and Torres Strait Islander mothers (CC/MI p<0.001) compared to children of Aboriginal mothers. A higher incidence of anaemia was seen among children of mothers with parity three or more (CC/MI p<0.001); children born by caesarean section (CC/MI p<0.001); and children with rapid early growth (CC/MI p<0.001).

Conclusion: Early childhood anaemia is common among Aboriginal and Torres Strait Islander children of Far North Queensland. Poor nutrition, particularly iron deficiency, and frequent infections are likely causes.

Implications for public health: Prevention of early childhood anaemia in ‘Close the Gap’ initiatives would benefit the Aboriginal and Torres Strait Islander children of Far North Queensland – and elsewhere in northern Australia.

Key words: anaemia, Aboriginal, Torres, child, mother, Queensland

INDIGENOUS HEALTH


Leonard et al. Article

aged six to 23 months (n=956) were anaemic in 2016/17.23

Comparable information on anaemia among other Australian children is limited. Localised surveys (2010) reported 1–6% of toddlers had iron deficiency anaemia increasing to 14% among those of Asian background, while one national survey from 1995 reported that 2% of 1–4-year-old children were anaemic.24-26

In remote Far North Queensland (Map 1), 71.5% of the population are Aboriginal and/or Torres Strait Islander people (n=14,107).27 A recent audit from eight Cape York Aboriginal communities reported that 32.3% of children aged six to 23 months were anaemic.28 However, published information is lacking for the wider Far North Queensland region.

The current study was undertaken to investigate anaemia among Aboriginal and Torres Strait Islander mothers and their children in Far North Queensland. Here we describe early childhood anaemia, defined as anaemia at age six to 23 months, and the characteristics associated with early childhood anaemia among Aboriginal and Torres Strait Islander children.

Methods

This retrospective cohort study used information from three existing health service data collections, extracted, linked

and de-identified by the Queensland Health Statistical Services Branch. The process of securing this information has been previously described.29 Briefly, data recorded between 2000 and 2015 were extracted from the Queensland Perinatal Data Collection (PDC)30; the Queensland Health Pathology Services Data Collection (Auslab)31; and the community health services electronic record system, Ferret,32 used mainly in remote Far North Queensland (Map 1 and Supplementary Table 1).

Study data were provided for two cohorts of Aboriginal and Torres Strait Islander children and their mothers: the Cape York cohort and the 2009–2010 cohort. The Cape York cohort includes children of the remote Cape York communities only, born between 2006 and 2008. The 2009–2010 cohort includes children born to Aboriginal and/or Torres Strait Islander mothers with a Queensland Perinatal Data Collection (PDC) record of birth in 2009 or 2010 in Far North Queensland, which includes the Torres region, Cape York and Cairns and Hinterland. Children included in this analysis are those with a Ferret record in addition to a PDC record. The Ferret system was implemented mainly in discrete Aboriginal and Torres Strait Islander communities across Far North Queensland (Map 1). Twelve of these localities were in Cape York, 21 in the Torres region and five in the Cairns and Hinterland region.

Longitudinal information on child growth and haemoglobin levels was recorded on the Ferret system. To ensure independence of events for statistical analysis, only the first child born to each mother between 2006 and 2010 was included.

Ethics approval was granted by Queensland Health Far North Queensland Human Research Ethics Committee (HREC/15/QCH/50-980) in June 2015. Approval under the Queensland Public Health Act 2005 was granted by the Director-General of Queensland Health in February 2016. The complete linked de-identified data was provided to the research group in May 2017.

Study variables and definitions Anaemia was defined as per Queensland Health clinical guidelines (haemoglobin <105 g/L from six to 11 months; haemoglobin <110g/L for children from 12 to 23 months).33 Children aged six to 23 months, with at least one haemoglobin level recorded below the respective criteria for age at the date of measurement, were considered to have anaemia. The haemoglobin levels reported here for children were measured on capillary blood using a HemoCue®.

Some characteristics are as reported on the Perinatal Data Collection30 (mother’s usual residence, ethnicity, parity, smoking in pregnancy, pregnancy induced

Map 1: Caption Far North Queensland – Hospital and Health Service boundaries and localities of Ferret electronic health records system.


Indigenous Health Early childhood anaemia

hypertension; baby’s sex, gestational age at birth, birthweight, method of birth). Birthweight z-scores adjusted for sex and gestational age, for babies with gestational age of 33 weeks or more were calculated using the INTERGROWTH-21ST Neonatal Size Calculator.34,35 The INTERGROWTH-21st standards for newborns are designed to complement the WHO Child Growth Standards.34 Weight for age z-scores for the first weight measure at age four to six months recorded on the Ferret system were calculated using the STATA ‘zscore06’ module, which is based on the 2006 World Health Organization sex-specific child growth standards.36,37

Where measurements were available, z-score-change was calculated (z-score for the first weight measurement at age four to six months minus birthweight z-score for weight for gestational age). Z-score-change is a measure of change in weight for age z-scores in the first months of life. A positive value indicates an increase in weight for age, a negative value indicates a decline in weight for age, while a zero value indicates no change in weight for age (Supplementary Table 2).

Other characteristics (maternal body mass index [BMI] and age; baby’s prematurity and/or low birthweight) were derived from Perinatal Data Collection information using criteria specified by the Australian Institute of Health and Welfare and the National Health and Medical Research Council, unless otherwise referenced.38,39 Information on maternal glucose tolerance, haemoglobin, ferritin, red cell folate (RCF) and vitamin B12 levels are as recorded on the Queensland Pathology Auslab system. Maternal anaemia in the third trimester of pregnancy was defined as an Auslab record of mother’s haemoglobin level <110 g/L as per Queensland Health clinical guidelines, measured on a date between estimated day 186 of pregnancy and the date of birth of the child.24,40 Supplementary Table 2 provides further details on definitions including implausibility criteria. Implausible values were considered missing.

The Socio-Economic Index for Areas (SEIFA 2011) ranks Australian Bureau of Statistics Statistical Local Areas (SLA) by deciles of relative socio-economic advantage and disadvantage.41 A ranking of ‘1’ indicates greatest relative disadvantage while a ranking of ‘10’ indicates greatest relative advantage.

The appropriate SEIFA decile ranking was allocated to each mother based on her usual place of residence.

Statistical analysisCategorical variables were described using absolute and relative frequencies. The distributions of numerical variables were assessed; symmetrically distributed numerical characteristics were described using mean values, 95% confidence intervals (95%CI), and ranges; numerical values with a skewed distribution were described using median, inter-quartile ranges (IQR) and ranges. The cumulative incidence of anaemia between age six to 23 months was presented with 95% confidence interval (95%CI). Mean haemoglobin levels using the first haemoglobin reading for each child, and incidence of anaemia were presented by six-month age groups (six–11 months, 12–17 months, 18–23 months). Children were included in one or more of the six-month age intervals if the appropriate measurements were available at that age but once only in analysis for the six to 23-month age group. Characteristics of the children and their mothers were compared between those children who had early childhood anaemia and those who did not, using bivariate logistic regression analyses adjusted for cohort.

The following characteristics were considered during multivariable analyses (Cohort 1 “2009-2010 cohort” n=407; Cohort 2 “Cape York cohort” n=301): sex of the baby; birthing method (non-instrumental vaginal, instrumental vaginal, caesarean section); gestational age of baby; whether baby was premature or not; birthweight of baby; z-score-change (z-score for weight for age at first weight at age four to six months less z-score for birthweight); feeding method to age four months (only breast milk, only infant formula, both breast milk and formula); ethnicity of mother (Aboriginal, Torres Strait Islander, both); region of residence of mother; SEIFA category for residence of mother; age of mother when baby was born; BMI category of mother (underweight, normal weight, overweight, obese); categories of parity (0-2, >=3); smoking during pregnancy; five or more antenatal care visits; pregnancy induced hypertension; mother had pre-existing diabetes; mother had gestational diabetes); low RCF level before or during pregnancy; low B12 level before or during pregnancy; mother anaemic in the third trimester of pregnancy.

Multivariable logistic regression analyses were conducted to identify independent risk factors for early childhood anaemia for the complete case analysis. Backward and forward stepwise modelling procedures were initially conducted to establish basic multivariable models for the combined cohorts. Characteristics that were not part of the basic models were assessed for potential confounding effects. A confounder was assumed to be a variable that changed estimates of characteristics in the basic model by 10% or more.42

Multivariable multiple imputation was conducted using Stata’s MI commands for sequential imputation using chained equations. Missing values were imputed for BMI of mother; parity; smoking during pregnancy; mother anaemic in the third trimester of pregnancy; mother with pre-existing diabetes; mother with gestational diabetes; number of antenatal visits five or more; feeding method to age four months; and z-score-change from birth to age four to six months. Low RCF and B12 levels before or during pregnancy were not imputed because these characteristics were missing in close to 80% of cases. Examination of patterns of missing data was conducted using Pearson’s chi-square and Fisher’s exact tests to compare the occurrence of missing values in characteristics (Supplementary Table 3). Patterns of missing values were assessed and judged to be “missing at random”.43 Linear regression was used to impute missing values of continuous characteristics; logistic regression was used to impute missing values of dichotomous characteristics; ordinal logistic regression was used to impute missing values of the categories of BMI. Imputation models were based on the following variables with nil missing data: early childhood anaemia; sex of baby; gestational age of baby; baby premature; birthing method; birthweight of baby; pregnancy induced hypertension, ethnicity of mother, age of mother, SEIFA index; antenatal care received; and cohort. Forty imputed data sets were created. Multivariable logistic regression analyses were conducted to identify independent risk factors for early childhood anaemia for imputed data.

Results of multivariable models for complete case and imputed data analyses are presented as odds ratios (OR) and 95% confidence intervals. P values of less than 0.05 were considered statistically significant.


Exclusions: Non‐Indigenous mother n = 15 Mother not resident in FNQ n = 10 Second birth (n = 40) to same

mother Second born twin n = 3

957 unique mother and child pairs

After exclusions: Children n = 645 Mothers n = 645

Exclusions: Mother not resident in FNQ n = 5 Second birth (n = 43) to same mother Third birth (n = 1) to same mother Second born twin n = 3 Mothers already included in Cape York

cohort (n = 78)

After exclusions: Children n = 312 Mothers n = 312

Figure 1: Flow diagram – early childhood anaemia among two cohorts of Aboriginal and Torres Strait Islander children and their mothers in Far North Queensland; data available and exclusions

Child seen by health services aged 6 to 23 months n = 904

Child seen by health services and haemoglobin measured aged 6 to 23 months

n = 708

Babies (n = 2167) with a Perinatal Data Collection (PDC) record of birth to an Aboriginal and/or Torres Strait Islander mother (n = 1993) in Far

North Queensland (FNQ) in 2009 or 2010

Children n = 775 of n = 735 mothers had Ferret longitudinal health record

Children n = 380 of n = 339 mothers had Ferret longitudinal health record

PDC records available for n = 2548 babies and their mothers n = 2332

Cape York Aboriginal and/or Torres Strait Islander babies (n = 434) born between 2006 and 2008, included in

previous study. Perinatal Data Collection (PDC) records located for 381

babies and their mothers (n = 339)

Figure 1: Flow diagram – early childhood anaemia among two cohorts of Aboriginal and Torres Strait Islander children and their mothers in Far North Queensland; data available and exclusions.

Figure 2. Incidence of Anaemia among Aboriginal and Torres Strait Islander children (n = 708) of Far North Queensland from age 6 to 23 months, and by 6 month age groups (%, 95% confidence interval)

0.0 20.0 40.0 60.0 80.0

6 ‐ 11 months (n=492)

12 ‐ 17 months (n=515)

18 ‐ 23 months (n=462)

6 ‐ 23 months (n=708)

Incidence of anaemia (%; 95%CI)

Age grou

p

Figure 2: Incidence of anaemia among Aboriginal and Torres Strait Islander children (n = 708) of Far North Queensland from age six to 23 months, and by six-month age groups (%, 95% confidence interval).

Analysis was conducted using Stata version 13 (StataCorp, Lakeway Drive, College Station, Texas).

Results

Linked de-identified data was provided in May 2017 for 2,548 Aboriginal and Torres Strait Islander children born to 2,332 mothers in Far North Queensland between 2006 and 2010. Ferret records were available for 1,155 children of 1,074 mothers. The number of children for whom this information was available is close to the estimated 1,147 child residents based on census population figures for those localities (Figure 1, Map 1, Supplementary Table 4). Information was excluded where the mother was non-Indigenous (n=15), not resident in Far North Queensland (n=15) and where the child was not the first child born to his/her mother in the cohort years (n=90). Seventy-eight mothers were excluded from the 2009–2010 cohort because they were already included in the Cape York cohort. After exclusions, the number of unique mother and child pairs was 957 (Figure 1).

Ferret records showed at least one visit to health services between age six and 23 months for 904 (94.5%) of these 957 children, of whom 708 (74.0%) had a haemoglobin level recorded at least once between the ages of six and 23 months (Supplementary Tables 5 and 6). No significant differences were seen between children for whom haemoglobin measurements were available and those without haemoglobin measurements, except that children from Cape York were more likely to have been seen by health services and have had a measurement of haemoglobin made, compared to children from elsewhere (p<0.001), see Supplementary Tables 5 and 6.

Of these 708 children, 61.3% (95%CI 57.7%, 64.9%) had at least one haemoglobin measure showing anaemia; the incidence of anaemia by six-month age groups was highest at 12–17 months (Figure 2). Mean haemoglobin was above the level indicating anaemia (105 g/L) at six to 11 months; 109.8 g/L (95%CI 108.7,110.9) but below the level indicating anaemia (110 g/L) at 12–17 months; 109.3 g/L 95%CI 108.3, 110.3) and close to that level (110 g/L) at 18–23 months; 111.8 (95%CI 110.8, 112.8), see Supplementary Figure 1. Among children anaemic at six to 11 months who had subsequent haemoglobin measurements, 102 out of 150 (68%) were also anaemic at



12–17 months and 69 out of 138 (50%) at 18–23 months.

Haemoglobin measurements were not available for 249 children between the age of six and 23 months. If it is assumed that all these children did not have anaemia, the number of children without anaemia (n=249 + 274) would be n=523. Under this hypothetical assumption, the incidence of early childhood anaemia would be 45.4% (95%CI 42.2%, 48.5%).

Bivariate analysis (n=708) showed the incidence of early childhood anaemia was higher among boys (65.8%) compared to girls (56.5%, p<0.001) and among children born by caesarean section (69.5%) compared to those born by vaginal birth (57.7%, p<0.001). Children who had early childhood anaemia had lower mean birthweight (3,159g vs. 3,217g, p=0.01) and lower mean birthweight z-score (+0.082 vs. +0.274, p=0.001), and higher mean gains in z-score (+0.254 v-0.013, p<0.001) for weight for age in early life, see Tables 1a and 1b).

Children of Aboriginal mothers (71.4%) had higher incidence of early childhood anaemia than children of mothers who were Torres Strait Islander (46.9%) or both Aboriginal and Torres Strait Islander (43.9%, p<0.001), see Table 2).

Multi-variable analysis (Table 3) showed higher incidence of early childhood anaemia among children born by caesarean section compared to children born by vaginal birth (p<0.001), children with higher gains in weight for age (p<0.001) and children of mothers with a parity of three children or more (p<0.001), and lower incidence of early childhood anaemia among girls compared to boys (p<0.001). Children whose mothers were Torres Strait Islander (p<0.001) or both Aboriginal and Torres Strait Islander (p<0.001) had lower incidence of early childhood anaemia compared to children whose mothers were Aboriginal (Table 3). The analysis was repeated using mother’s region of residence instead of mother’s ethnicity. Children of mothers who were resident in the Torres Strait and Northern Peninsula Area (MI p=0.003) and children of mothers resident in Cairns and Hinterland (MI p=0.005) had lower incidence of early childhood anaemia compared to children of mothers resident in Cape York (Supplementary Table 7).

Multi-variable analysis showed disparate results in respect of smoking in pregnancy. Multiple imputation analysis, but not complete case analysis, showed significantly

higher incidence of early childhood anaemia among children of mothers who smoked in pregnancy compared to children of mothers who did not smoke in pregnancy (MI p=0.023), see Table 3. Birthweight, age of mothers and anaemia of mothers in the third trimester of pregnancy were found to be confounding factors in multivariable analyses.

Discussion

This study shows that early childhood anaemia was common among these (n=708)

Aboriginal and Torres Strait Islander children of Far North Queensland with the incidence of anaemia being 61.3% between the age of six and 23 months. Many children with anaemia before the age of 12 months were still anaemic in the second year of life. Mean haemoglobin levels were low; below the diagnostic level for anaemia from 12 to 17 months. Our findings are consistent with reports of high rates of early childhood anaemia among Aboriginal and Torres Strait Islander infants and young children elsewhere in northern Australia.24,44,45

Table 1a: Incidence of Early Childhood Anaemia (anaemia between age six and 23 months) among Aboriginal and Torres Strait Islander children (n=708) of Far North Queensland by characteristics of the children.Characteristics of children n Child ever anaemic

age 6–23 months n (%) [95%CI]

P value (logistic regression adjusted for

cohort – unless stated otherwise) Cohorts: Both combined 708 434 (61.3%) [57.7%, 64.9%] n/a 2009–2010 births cohort 407 199 (48.9%) [44.1%, 53.8%] chi2 <0.001 Cape York cohort 301 235 (78.1%) [73.4%, 82.8%]Gender: Male 363 239 (65.8%) [60.9%, 70.7%) <0.001 Female 345 195 (56.5%) [51.3%, 61.8%]Birth method: Vaginal 473 273 (57.7%) [53.2%, 62.2%] base Vaginal/Instrumental 35 22 (62.9%) [46.0%, 79.7%] <0.001 Caesarean 200 139 (69.5%) [63.1%, 75.9%] <0.001Birth Weight category: Low birth weight (<2,500g) 81 52 (64.2%) [53.5%, 74.9%] 0.008 Normal (2,500–4,000g) 580 355 (61.2%) [57.2%, 65.2%] base Marcosomic (>=4,000g) 47 27 (57.4%) [42.8%, 72.1%] 0.782Gestational age category: Preterm (<37 weeks) 82 50 (61.0%) [50.2%. 71.8%] 0.893 Full-term (>=37 weeks) 626 384 (61.3%) [57.5%, 65.2%]Feeding method birth to 4-6 months; n=544 (164 missing) Only breast milk 228 157 (68.9%) [62.8%, 74.9%] base Only infant formula 56 30 (53.6%) [40.1%, 67.0%] 0.249 Breast milk and formula 260 159 (61.2%) [55.2%, 67.1%] 0.018

Table 1b: Incidence of Early Childhood Anaemia (anaemia between age six and 23 months) among Aboriginal and Torres Strait Islander children (n = 708) of Far North Queensland by characteristics of the children.Characteristics of children All

n=708 Ever anaemic age

6–23 months n=434

Not anaemic age 6–23 months

n=274

P value (logistic regression

adjusted for cohort)Gestational age at birth weeks – median (IQR) [range]

39 (38–40) [26–42]

39 (38–40) [27–42]

39 (38–40) [26–42]

0.036

Birth weight - grams mean (95% CI) [range]

3,181 (3,136, 3,225)

[800–5,320]

3,159 (3,102, 3,215)

[800–5,320]

3,217 (3,145, 3,288)

[960–4,780]

0.010

Z-score for birth-weight for gestational age mean (95%CI) [range] n=692 (missing n=16)

+0.16 (+0.07, +0.24)

[-2.9–+4.3] n=692

+0.082 (-0.021, +0.18)

[-2.9–+4.3] n=426

+0.274 (+0.142, +0.405)

[-2.4–+3.2] n=266

0.001

Z-score-change birth to first weight at age 4 -6 months mean (95% CI) [range] n=527(missing n=181)

0.16 (+0.057, +0.255)

[-3.7–+3.7] n=527

+0.254 (+0.125, +0.382)

[-3.7–+3.7] n=334

-0.013 (-0.167, +0.140)

[-3.5–+3.0] n=193

<0.001



Our findings show that children of Aboriginal mothers had higher incidence of early childhood anaemia compared to children of mothers who were Torres Strait Islander or both Aboriginal and Torres Strait Islander. Further analysis by mother’s region of residence showed the same pattern. These results reflect the different history of Aboriginal people of Cape York compared to people of the Torres Strait. Government policies forcibly relocated Queensland Aboriginal people from their traditional lands to mission settlements, some of which are now the remote communities of Cape York.48 This “large scale relocation did not occur in the Torres Strait”.48 Despite Government restrictions and impositions, Torres Strait Islander peoples largely remained on their traditional lands, a key factor in preserving cultural continuity, including traditional food systems.48

The high cost of nutritious food has been widely reported, while household food insecurity is exacerbated by smoking.49,50 The implications of poor health of mothers on the future health of their children have been raised previously.51 The intergenerational association reported here, of high parity and maternal smoking with early childhood anaemia, reflects the shared experiences of food insecurity of these mothers and their children in a context of poverty and social disadvantage that is particularly challenging in Cape York.48

The limitations of this study are those associated with the use of routine health service data, including missing information.29 The multiple imputation methodology was used to adjust for missing values and results are presented for both complete case and multiple imputation analyses. However, some information was not recorded on the electronic data collections accessed for this study. For example, information about treatment of anaemia was not available for mothers or children. It may be that treatment of maternal anaemia protects the unborn child from subsequent anaemia, but this hypothesis could not be tested. Similarly, the lack of information on treatment of children meant that the effect of treatment at first diagnosis of early childhood anaemia on subsequent haemoglobin levels could not be assessed.

In addition, many (26.0%) of the 957 children with a Ferret record did not have a measure of haemoglobin recorded between the age of six and 23 months; most (n=196, 20.5%) were

The finding of more early childhood anaemia among children born by caesarean or vaginal/instrumental births may reflect the urgency of such births, with early cord clamping reducing transfer of placental blood to the

newborn.8,9 Caesarean births are increasing among Indigenous mothers; this finding may be particularly relevant for the Torres Strait where diabetes in pregnancy and births by caesarean section are common.46,47

Table 2: Incidence of Early Childhood Anaemia (anaemia between age six and 23 months) among Aboriginal and Torres Strait Islander children (n = 708) of remote Far North Queensland by characteristics of their mothersCharacteristics of mothers n Child ever anaemic

age 6–23 months n (%) [95%CI]

P value (logistic regression

adjusted for cohort)Ethnicity Aboriginal 423 302 (71.4%) [67.1%, 75.7%] base Torres Strait Islander 228 107 (46.9%) [40.4%, 53.5%] <0.001 Both Aboriginal and Torres Strait Islander 57 25 (43.9%) [30.6%, 57.1%] <0.001Region of residence Cairns and Hinterland 56 29 (51.8%) [38.3%, 65.3%] 0.025 Cape York 442 318 (72.0%) [67.7%, 76.2%] base Torres Strait and Northern Peninsula Area 210 87 (41.4%) [34.7%, 48.1%] 0.023SEIFA – usual residence Mother resident in SEIFA 1 620 377 (60.8%) [57.0%, 64.7%] 0.509 Mother resident in SEIFA 2 - 10 88 57 (64.8%) [54.6%, 75.0%]Body Mass Index of mothers (n=481, missing n=227) Underweight (9.8%) 47 34 (72.3%)[59.1%, 85.6%] 0.037 Healthy weight (35.8%) 172 93 (54.1%) [46.5%, 61.6%] base Overweight (25.2%) 121 63 (52.1%) [43.0%, 61.1%] 0.459 Obese (29.3%) 141 68 (48.2%) [39.9%, 56.6%] 0.032Teenage mothers Teenage mother 163 95 (58.3%) [50.6%, 65.9%] 0.162 Mother age 20 years or older 545 339 (62.2%) [58.1%, 66.3%] Antenatal visits (missing n=1) Less than 5 visits 83 43 (51.8%) [40.8%,62.8%] <0.001 5 visits or more 642 390 (62.5%) [58.7%, 66.3%]Parity (missing n=233) nil to 2 277 161 (58.1%) [52.3%, 64.0%] 0.103 3 or more 198 127 (64.1%) [57.4%, 70.9%]Smoked in pregnancy (n=703, missing n=5) Yes 439 275 (62.6%) [58.1%, 67.2%] 0.020 No 264 155 (58.7%) [52.7%, 64.7%]Gestational Diabetes (n=421, missing n=287) Yes 75 42 (56.0%) [44.5%, 67.5%] 0.329 No 346 198 (57.2%) [52.0%, 62.5%]Pre-existing Diabetes (n=587, missing n=121) Yes 33 23 (69.7%) [53.1%, 86.2%] <0.001 No 554 330 (59.6%) [55.5%, 63.7%]Pregnancy Induced Hypertension (PIH) Yes 45 31 (68.9%) [54.8%, 83.0%] 0.026 No 663 403 (60.8%) [57.1%, 64.5%]Anaemia in third trimester (n=657, missing n=51) Yes 336 202 (60.1%) [54.9%, 65.4%] 0.014 No 321 199 (62.0%) [56.7%, 67.3%]Iron deficiency in pregnancy (n=385, missing n=323) Yes 185 110 (59.5%) [52.3%, 66.6%] 0.775 No 200 123 (61.5%) [54.7%, 68.3%]Low Red Cell Folate (RCF) before/during pregnancy (n=158, missing n=550) Yes 20 16 (80.0%) [60.8%, 99.2%] <0.001 No 138 71 (51.5%) [43.0%, 59.9%]Low B12 before/during pregnancy (n=131, missing n=577) Yes 22 7 (31.8%) [10.7%, 53.0%] 0.151 No 109 62 (56.9%) [47.4%, 66.3%]



seen by health services but haemoglobin levels were not recorded. Children in Cape York were more likely to be seen and to have a haemoglobin level recorded than children from elsewhere. No other differences were identified. However, the reason for this missing data is not known and there may be an unidentified bias in the availability of relevant information.

Another limitation of our study is that haemoglobin levels were measured on capillary blood using HemoCues®. The use of capillary blood and these devices may underestimate haemoglobin levels, which would result in overestimation of the incidence and prevalence of anaemia.52 Validation studies suggest that the HemoCue® is suitable for screening purposes but, where anaemia is suspected, other methods should also be used.53 However, health service protocols for diagnosis and treatment of anaemia in remote Far North Queensland are based on HemoCue® measurements.33

The information presented here is for children born between 2006 and 2010. It is possible that the situation in respect of early childhood anaemia has changed. However, in Cape York in 2014 and 2015, about one-in-three children aged six to 23 months were anaemic (n=155, 32.3% anaemic, 95%CI 24.8%, 39.7%) indicating that early childhood anaemia continues to be a problem in Cape York.28 Comparable information is not available from elsewhere in Far North Queensland.

There is no information to identify the cause(s) of the anaemia reported here. Nutrient deficiencies such as iron deficiency cause anaemia, as do chronic infections.1,54 Iron deficiency is the ‘usual suspect’ as a nutrition-related cause of anaemia in early life because of the high requirements for iron due to rapid growth.10 Iron requirements per kilogram of body weight are higher at six to 12 months than at other stages of the life cycle.5 Australian estimates show that the daily iron requirements of a child aged seven to 12 months are higher than those of an

adult man: Estimated Average Requirement (EAR) child seven to 12 months – 7mg; EAR male aged 19 years or more – 6mg.55 Milk is not a rich source of iron; during the first months of exclusive breastfeeding, a baby draws on iron stores acquired before birth from the mother.5 Subsequently the small quantity of solid food consumed by young children must provide most of these high iron requirements.10 Consequently, iron-rich and/or iron-fortified first foods are recommended in Australia.11

Two studies in the Northern Territory showed the association of iron deficiency with childhood anaemia in similar settings. In one study, among young Aboriginal children (n=74) with anaemia, most (n=62, 84%) had iron deficiency anaemia, with folate deficiency and chronic infections identified as causes of anaemia in the other children.21 Another Northern Territory study among school-age children with anaemia (n=201) found that iron therapy was effective in resolving anaemia among 83% of the 66 children for whom follow-up measurements were available.22

Table 3: Risk factors for Early Childhood Anaemia (n=708); multi-variable analysis – complete case analysis and analysis with imputed data.Complete case analysis

n=329

Imputed data analysis n=708

Characteristic Anaemia

n=203

(61.7%)

No Anaemia

n=126

(38.3%)

Odds-ratio

(95% CI)a

p-value Number of missing values (%)

Anaemia

n=434

(61.3%)

No Anaemia

n=274

(38.7%)

Odds-ratio

(95% CI)a

p-value

Sex of child

Male

Female

115 (56.7%)

88 (43.4%)

59 (46.8%)

67 (53.2%)

1

0.63 (0.59, 0.67) P<0.001

0

239 (55.1%)

195 (44.9%)

124 (45.3%)

150 (54.7%)

1

0.62 (0.55, 0.71) P<0.001Z-score-change from birth to 4 to 6 months / / 1.3 (1.2, 1.4) p<0.001 181 (25.6%) / / 1.2 (1.1, 1.3) P<0.001

Age of motherb / / 0.97 (0.92, 1.02) P=0.224 0 / / 0.99 (0.98, 1.00) P=0.096Ethnicity of mother

Aboriginal

Torres Strait Islander

Both

144 (70.9%)

46 (22.7%)

13 (6.4%)

55 (43.7%)

54 (42.9%)

17 (13.5%)

1

0.34 (0.21, 0.53)

0.26 (0.17, 0.39)

P<0.001

P<0.001

0

302 (69.6%)

107 (24.7%)

25 (5.8%)

121 (44.2%)

121 (44.2%)

32 (11.7%)

1

0.35 (0.22, 0.56)

0.28 (0.19, 0.42)

P<0.001

P<0.001Parity

Up to 2 children

3 or more children

124 (61.1%)

79 (38.9%)

79 (62.7%)

47 (37.3%)

1

2.1 (1.7, 2.5) P<0.001

233 (32.9%)

284 (65.4%)

150 (34.6%)

193 (70.4%)

81 (29.6%)

1

1.8 (1.4, 2.5) P<0.001Birth method

Vaginal

Vaginal instrumental

Caesarian

125 (61.6%)

8 (3.9%)

70 (34.5%)

100 (79.4%)

3 (2.4%)

23 (18.3%)

1

3.1 (1.9, 5.2)

3.0 (2.9, 3.1)

P<0.001

P<0.001

0

273 (62.9%)

22 (5.1%)

139 (32.0%)

200 (73.0%)

13 (4.7%)

61 (22.3%)

1

1.4 (1.1, 1.9)

1.7 (1.4, 2.1)

P=0.013

P<0.001Mother anaemic in third trimesterb

No

Yes

103 (50.7%)

100 (49.3%)

63 (50.0%)

63 (50.0%)

1

1.0 (0.5, 2.2) P=0.934

51 (7.2%)

216 (49.8%)

218 (50.2%)

131 (47.8%)

143 (52.2%)

1

0.89 (0.77, 1.03) P=0.122Mother smoked during pregnancy

No

Yes

75 (37.0%)

128 (63.1%)

50 (39.7%)

76 (60.3%)

1

1.0 (0.7, 1.5) P=0.964

5 (0.7%)

156 (35.9%)

278 (64.1%)

109 (39.8%)

165 (60.2%)

1

1.2 (1.02, 1.3) P=0.023Notes:Both models were adjusted for the confounding effect of birth weight (no missing values imputed). a: 95% CI = 95% confidence interval. b: Mother anaemic in third trimester and mothers’ age were identified as confounding variables in complete case data analysis. Imputed data are averages of 40 imputations.



Several of the risk factors identified here are consistent with iron deficiency as a cause of the anaemia: birth by caesarean section; rapid early growth; and boys compared to girls, as boys typically have higher early weight gains than girls.56,5

Infections are another probable cause of anaemia in these children, with high rates of infectious diseases reported for Aboriginal and Torres Strait Islander children of remote Far North Queensland.57 There is a bidirectional relationship between infectious disease and nutrition status; frequent illness impairs nutrition status and poor nutrition status increases susceptibility to infection.58 The immune response to infections restricts iron availability to infectious organisms; when prolonged, this immune response can lead to anaemia.59

However, the diagnosis of anaemia based on haemoglobin only, as in this study, cannot identify the cause(s) of anaemia. The lack of information on the causes of early childhood anaemia is not only a limitation of this study but also a limitation of methods currently available to identify causes of anaemia in early childhood; in particular, the assessment of iron status is complex in the presence of infection.60,54

Prevention of early childhood anaemia is important as successful treatment of anaemia may not reverse the associated neurological deficits.2 Where the prevalence of early childhood anaemia is high (20% or more), WHO recommends interventions that combine promotion of breastfeeding and healthy food with home fortification of solid foods using multi-micronutrient preparations for babies/children aged six to 23 months.1,61-63 Such interventions have been demonstrated to be acceptable, safe and effective in the prevention and treatment of early childhood anaemia in low-income settings where the infectious disease burden is high.61-63 One such intervention, the Fred Hollows Foundation Early Childhood Nutrition and Anaemia Prevention Project (ECNAPP), was successfully piloted in six remote communities across northern Australia in 2010–2012.64 Nutrition-focused interventions will be strengthened by complementary interventions to improve food security and reduce infections in early life.58

Improved nutrition in the first one thousand days of life – through pregnancy up to the age of two years – provides “a golden

opportunity to impact neurodevelopment and brain function through the lifespan”.2 Prevention of early childhood anaemia, included as a key strategy to ‘Close the Gap’ between Aboriginal and Torres Strait Islander Australians and other Australians,65 would benefit the Aboriginal and Torres Strait Islander children of Far North Queensland – and elsewhere in northern Australia.

Acknowledgements and funding

Dympna Leonard was supported by a National Health and Medical Research Council post-graduate scholarship APP1092732.

Other agencies provided non-financial support to this research. We acknowledge and thank the Aboriginal and Torres Strait Islander leaders of the key community-controlled health service organisations in far north Queensland who considered and endorsed the proposed research, providing the support that made this research possible. In addition, we acknowledge and thank the Queensland Health Data Custodians and their research and data management staff for their assistance and support for this work.

References 1. World Health Organization. Nutritional Anaemias: Tools

for Effective Prevention and Control. Geneva (CHE): WHO; 2017

2. Georgieff MK, Ramel SE, Cusick SE. Nutritional influences on brain development. Acta Paediatr. 2018;107(8):1310-21.

3. Zimmermann MB, Hurrell RF. Nutritional iron deficiency. Lancet. 2007;370:511-20.

4. Prado EL, Dewey KG. Nutrition and brain development in early life. Nutr Rev. 2014;72(4):267-84.

5. Domellöf M. Iron requirements in infancy. Ann Nutr Metab. 2011;59(1):59-63.

6. Scholl TO. Maternal iron status: Relation to fetal growth, length of gestation, and iron endowment of the neonate. Nutr Rev. 2011;69 Suppl 1:23-9.

7. Balarajan Y, Ramakrishnan U, Özaltin E, Shankar AH, Subramanian SV. Anaemia in low-income and middle-income countries. Lancet. 2011;378(9809):2123-35.

8. Dewey KG, Chaparro CM. Iron status of breast-fed infants. Proc Nutr Soc. 2007;66(3):412-22.

9. McDonald SJ, Middleton P, Dowswell T, Morris PS. Effect of timing of umbilical cord clamping of term infants on maternal and neonatal outcomes. Cochrane Database Syst Rev. 2013 Jul 11;(7):CD004074.

10. Dewey KG. The challenge of meeting nutrient needs of infants and young children during the period of complementary feeding: An evolutionary perspective. J Nutr. 2013;143(12):2050-4.

11. National Health and Medical Research Council. Infant Feeding Guidelines. Canberra (AUST): NHMRC; 2013.

12. O’Dea K, Jewell PA, Whiten A, Altmann SS, Strickland ST, Oftland OT. Traditional diet and food preferences of Australian aboriginal hunter-gathers. Philos Trans R Soc Lond B Biol Sci. 1991;334(1270):233-40; discussion 240-1.

13. Lee AJ. The transition of Australian aboriginal diet and nutritional health. World Rev Nutr Diet. 1996;79:1-52.

14. Food Standards Australia New Zealand. Indigenous Food 2010 Composition of Indigenous Foods of Australia-updated [Internet]. Canberra (AUST): FSANZ; 2015 [cited 2018 Jul].

15. Australian Bureau of Statistics. 4727.0.55.005- Australian Aboriginal and Torres Strait Islander Health Survey: Nutrition Results-Food and Nutrients 2012-13. Canberra (AUST): ABS; 2016.

16. Eades SJ, Read AW, McAullay D, McNamara B, O’Dea K, Stanley FJ. Modern and traditional diets for Noongar infants. J Paediatr Child Health. 2010;46(7-8):398-403.

17. Leonard D, Aquino D, Hadgraft N, Thompson F, Marley J. Poor nutrition from first foods: A cross sectional study of complementary feeding of infants and young children in six remote Aboriginal communities across northern Australia. Nutr Diet. 2017;74(5):436-45.

18. Moradi S, Arghavani H, Issah A, Mohammadi H, Mirzaei K. Food insecurity and anaemia risk: A systematic review and meta-analysis. Public Health Nutr. 2018;21(16):1-13.

19. Brewster DR. Iron deficiency in minority groups in Australia. J Paediatr Child Health. 2004;40:422-3.

20. Edmond K. Anaemia in mothers and infants living in disadvantaged communities. J Trop Pediatr. 2014;60(6):407-8.

21. Kruske SG, Ruben AR, Brewster DR. An iron treatment trial in an Aboriginal community: Imporving non-adherence. J Paediatr Child Health. 1999;35:153-8.

22. Udovicich C, Perera K, Leahy C. Anaemia in school-aged children in an Australian Indigenous community. Aust J Rural Health. 2017;25(5):285-9.

23. Northern Territory Health. Healthy Under Five Kids 2017-18. Darwin (AUST): Government of the Northern Territory; 2018.

24. Pasricha S-R, Flecknoe-Brown SC, Allen KJ, et al. Diagnosis and management of iron deficiency anaemia: A clinical update. Med J Aust. 2010;193(9):525-32.

25. Zhou SJ, Gibson RA, Gibson RS, Makrides M. Nutrient intakes and status of preschool children in Adelaide, South Australia. Med J Aust. 2012;196(11):696-700.

26. Mackerras D, Hutton SI, Anderson PR. Haematocrit levels and anaemia in Australian children aged 1-4 years old. Asia Pac J Clin Nutr. 2004;13(4):330-5.

27. Queensland Statistician’s Office. Resident Population Profile 2017 [Internet]. Brisbane (AUST): State Government of Queensland; 2017 [cited 2017 Jul 11].

28. Apunipima Cape York Health Council. Anaemia in Young Children of Cape York: Results of a Chart Audit. Cairns (AUST): Apunipima Cape York Health Council; 2016.

29. Leonard D, Buettner P, Thompson F, Makrides M, McDermott R. Linking ‘data silos’ to investiate anaemia among Aboriginal and Torres Strait Islander mothers and children in Far North Queenland. Aust N Z J Public Health. 2018;42(5):256-62.

30. Queensland Health. Queensland Perinatal Data Collection Manual for the Completion of Perinatal Data. Brisbane (AUST): State Government of Queensland; 2016.

31. Queensland Health. Queensland Health Pathology Services. Cairns (AUST): State Government of Queensland; 2016.

32. Queensland Health. Ferret-Better Health Outcomes Project. Cairns (AUST ): State Government of Queensland; 2005.

33. Queensland Health. Primary Clinical Care Manual: Paediatrics. 9th ed. Carins (AUST): State Government of Queensland; 2017.

34. Villar J, Ismail LC, Victora CG et al. International standards for newborn weight, length and head cicumference by gestational age and sex: The Newborn Cross-Sectional Study of the INTERGROWTH-21st Project. Lancet. 2014;384(9946):857-68.

35. Intergrowth-21st Network. Newborn Size Pakage [Internet]. Version 1.3.5. Oxford (UK): University of Oxford Oxford Maternal & Perinatal Health Institute; 2014 [cited 2018 Apr]. Availabe from: https://intergrowth21.tghn.org/newborn-size-birth/#c4

36. de Onis M, Onyango A, Borghi E, et al. Worldwide implementation of the WHO Child Growth Standards. Public Health Nutr. 2012;15(9):1603-10.



37. World Health Organization. WHO Child Growth Standards: Methods and Development: Length/height-for-age, Weight-for age, Weight-for-length, Weight-for-height and Body Mass Index-for-age. Geneva (CHE): WHO; 2006.

38. National Health and Medical Research Council. Australian Dietary Guidelines. Canberra (AUST): NHMRC; 2013.

39. Australian Institute of Health and Welfare. Australia’s Mothers and Babies 2015-in brief. Canberra (AUST): AIHW; 2017.

40. Queensland Health. Primary Clinical Care Manual: Sexual and Reproductive Health. 9th ed. Cairns (AUST): State Government of Queensland; 2016.

41. Australian Bureau of Statistics. Australian Census of Population and Housing 2011: Socio-Economic Indexes for Areas (SEIFA). Canberra (AUST): ABS; 2013.

42. Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiologic Research. Principles and Quantitative Methods. New York (NY): Van Nostrand Reinhold; 1982.

43. Sterne JAC, White IR, Carlin JB, et al. Multiple imputation for missing data in epidemiological and clinical research: Potential and pitfalls. BMJ. 2009;338:b2393.

44. Aquino D, Leonard D, Hadgraft N, Marley J. High prevalence of early onset anaemia amongst Aboriginal and Torres Strait Islander infants in remote northern Australia. Aust J Rural Health. 2018;26(4):245-50.

45. Bar-Zeev SJ, Kruske SG, Barclay LM, Bar-Zeev N, Kildea SV. Adherence to management guidelines for growth faltering and anaemia in remote dwelling Australian Aboriginal infants and barriers to health service delivery. BMC Health Serv Res. 2013;13:250.

46. Thompson F, Dempsey K, Mishra G. Trends in Indigenous and non-Indigenous caesarean section births in the Northern Territory of Australia, 1986-2012: A total population-based study. BJOG. 2016;123(11):1814-23.

47. Falhammar H, Davis B, Bond D, Sinha AK. Maternal and neonatal outcomes in the Torres Strait Islands with a sixfold increase in type 2 diabetes in pregnancy over six years. Aust N Z J Obstet Gynaecol. 2010;50(2):120-6.

48. Australian Law Reform Commission. Pathways to Justice-An Inquiry into the Incarceration Rate of Aboriginal and Torres Strait Islander Peoples (ALRC Report 133). Canberra (AUST): Government of Australia; 2017.

49. Queensland Health. Healthy Food Access Basket 2014 [Internet]. Cairns (AUST): Queensland Health; 2014 [cited 2018 Nov 29]. Available from: https://www.health.qld.gov.au/research-reports/reports/public-health/food-nutrition/access/guidelines

50. Markwick A, Ansari Z, Sullivan M, McNeil J. Social determinants and lifestyle risk factors only partially explain the higher prevalence of food insecurity among Aboriginal and Torres Strait Islanders in the Australian state of Victoria: A cross-sectional study. BMC Public Health. 2014;14(598):1-10.

51. McDermott R, Campbell S, Li M, McCulloch B. The health and nutrition of young indigenous women in north Queensland-intergenerational implications of poor food quality, obesity, diabetes, tobacco smoking and alcohol use. Public Health Nutr. 2009;12(11):2143-9.

52. Gwetu TP, Chhagan MK, Craib M, Kauchali S. Hemocue validation for the diagnosis of anaemia in children: A semi-systematic review. Pediat Therapeut. 2013;4:187.

53. Sanchis-Gomar F, Cortell-Ballester J, Pareja-Galeano H, Banfi G, Lippi G. Hemoglobin point-of-care testing: The HemoCue system. Lab Autom. 2013;18(3):198-205.

54. Stoltzfus RJ, Klemm R. Research, policy, and programmatic considerations from the Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) Project. Am J Clin Nutr. 2017;106 Suppl 1:428-34.

55. National Health and Medical Research Council. Nutrient Reference Values for Australia and New Zealand. Canberra (AUST): NHMRC; 2006.

56. World Health Organization. MultiCentre Growth Study Velocity Report. Geneva (CHE): WHO; 2009.

57. Rothstein J, Heazlewood R, Fraser M, Paediatric Outreach Service. Health of Aboriginal and Torres Strait Islander children in remote Far North Queensland: findings of the Paediatric Outreach Service. Med J Aust. 2007;186(10):519-21.

58. Dewey KG, Mayers DR. Early child growth: How do nutrition and infection interact? Matern Child Nutr. 2011;7 Suppl 3:129-42.

59. Ganz T. Iron in innate immunity: Starve the invaders. Curr Opin Immunol. 2009;21(1):63-7.

60. Lynch S, Pfeiffer CM, Georgieff MK, et al. Biomarkers of nutrition for development (bond)-iron review. J Nutr. 2018;148 Suppl 1:1001-67.

61. Dewey KG, Yang ZY, Boy E. Systematic review and meta-analysis of home fortification of complementary foods. Matern Child Nutr. 2009;5(4):283-321.

62. Nyhus Dhillon C, Sarkar D, Klemm RD, et al. Executive summary for the micronutrient powders consultation: Lessons learned for operational guidance. Matern Child Nutr. 2017;13 Suppl 1. doi: 10.1111/mcn.12493.

63. Bhutta ZA, Das JK, Rizvi A, et al. Evidence-based interventions for improvement of maternal and child nutrition: What can be done and at what cost? Lancet. 2013;382(9890):452-77.

64. Aquino D, Marley J, Senior K, et al. The Early Childhood Nutrition and Anaemia Prevention Project-Summary Report. Darwin (AUST): The Fred Hollows Foundation; 2013.

65. Department of the Prime Minister and Cabinet. Closing the Gap-Prime Ministers Report 2018. Canberra (AUST): Government of Australia; 2018.

Supporting Information

Additional supporting information may be found in the online version of this article:

Supplementary Figure 1: Mean haemoglobin (g/L) (with 95% confidence interval) for Aboriginal and Torres Strait islander children of Far North Queensland by six month age groups.

Supplementary Table 1: Data collections used to source information to investigate Early Childhood Anaemia (anaemia at age six to 23 months) among Aboriginal and Torres Strait Islander children in Far North Queensland.

Supplementary Table 2: Definitions of variables used to describe characteristics of Aboriginal and Torres Strait Islander children (born between 2006 and 2010) and their mothers in Far North Queensland.

Supplementary Table 3: “Missingness” of characteristics in analysis of risk factors for early childhood anaemia in 708 Aboriginal children from North Queensland.

Supplementary Table 4: Early childhood anaemia among two cohorts of Aboriginal and Torres Strait Islander children and their mothers in Far North Queensland; comparisons of study participant numbers with census information.

Supplementary Table 5: Far North Queensland Aboriginal and/or Torres Strait Islander children with a Ferret record (n = 957): comparing those seen by health services at age six to 23 months (n = 904) with those not seen (n = 53).

Supplementary Table 6: Far North Queensland Aboriginal and/or Torres Strait Islander children with a Ferret record seen by health services at age six to 23 months (n = 904) - comparing those with at least one haemoglobin measure (n = 708) with those with no haemoglobin measure (n = 196) at age six to 23 months.

Supplementary Table 7: Risk factors for Early Childhood Anaemia (anaemia between six and 23 months) among Aboriginal and Torres Strait Islander children of Far North Queensland (n = 708): Results of multi variable analysis using region of residence of mother instead of ethnicity in imputed data model (all other variables as in Table 3).



Positive health impacts from small-scale or local lifestyle behaviour change interventions for Aboriginal Australians

have been demonstrated. For example, Canuto et al.1 observed decreased weight, body mass index (BMI) and blood pressure among urban Aboriginal women after 12 weeks of a structured exercise and diet program. A small trial of a 12-week supervised exercise program2 demonstrated improved BMI, waist and glucose metabolism measures in Aboriginal men. Although this and other short-term clinical interventions3-5 have shown positive effects, and cultural aspects were incorporated, they were more focused on treatment of disease and individuals and therefore may not achieve population-level change for Aboriginal Australians. Activities of daily living or sport-focused interventions may confer greater benefits because of access, affordability and their ability to build social capital among Aboriginal Australians.6-8 Team sports provide physical health and personal development opportunities and may provide a platform for community development benefits (social connection, cultural identity and life skills) through community sporting clubs.8-10 Unfortunately, these seldom achieve measurable and sustained large-scale community reach, particularly for Aboriginal peoples, with barriers such as costs of transport, membership and a lack of sport infrastructure

(facilities and sponsors) often affecting participation in Aboriginal communities.8

Community-based lifestyle behaviour change interventionsComparatively few community-based physical activity (PA) and nutrition interventions targeting Aboriginal

Australian adults have been evaluated.11,12 Three early interventions had variations in approach (including health education,13-15 environmental and policy changes14) and duration (six weeks13 to a few years14,15) but demonstrated waist13,15 and weight reductions,13 reduced sedentary behaviour,14 and increased vigorous PA.14 In a review of

Participant profile and impacts of an Aboriginal healthy lifestyle and weight loss challenge over four years 2012-2015Anne C. Grunseit,1,2 Erika Bohn-Goldbaum,1,2 Melanie Crane,1,2 Andrew Milat,1,3 Aaron Cashmore,3,4 Rose Fonua,5 Angela Gow,5 Rachael Havrlant,6 Kate Reid,5 Kiel Hennessey,6 Willow Firth,7 Adrian Bauman1,2

1. The Australian Prevention Partnership Centre, New South Wales 2. Sydney School of Public Health, University of Sydney, New South Wales3. Centre for Epidemiology and Evidence, NSW Ministry of Health, New South Wales4. School of Public Health and Community Medicine, University of NSW, New South Wales 5. NSW Office of Preventive Health, New South Wales6. Agency for Clinical Innovation, New South Wales7. South Coast Women’s Health and Welfare Aboriginal Corporation, New South WalesCorrespondence to: Dr Anne C. Grunseit, The Australian Prevention Partnership Centre, Sydney School of Public Health, Level 6, Charles Perkins Centre, University of Sydney,

Camperdown, NSW 2006; e-mail: [email protected]: January 2019; Revision requested: April 2019; Accepted: May 2019The authors have stated the following conflict of interest: Authors Fonua, Gow and Reid have a non-financial competing interest in that they work for the organisation that

operates the program. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.


Abstract Objective: To explore participation, consistency of demographic and health profiles, and short-term impacts across six Aboriginal Knockout Health Challenge (KHC) team-based weight loss competitions, 2012 to 2015.

Methods: Data comprised one competition each from 2012 and 2013 and two per year in 2014 and 2015. We compared baseline and change (pre- to post-competition) in weight, fruit and vegetable consumption, physical activity and waist circumference (baseline only) across competitions using mixed models.

Results: Numbers of teams and participants increased from 2012 to 2015 from 13 and 324 to 33 and 830, respectively. A total of 3,625 participants registered, representing 2,645 unique people (25.4% repeat participation). Participants were mainly female and >90% were classified obese at baseline. Baseline weight and weight lost (between 1.9% and 2.5%) were significantly lower in subsequent competitions compared with the first. Improvements in fruit and vegetable consumption and physical activity were comparable across competitions.

Conclusion: The KHC has increasing and sustained appeal among Aboriginal communities, attracting those at risk from lifestyle-associated chronic disease and effectively reducing weight and promoting healthy lifestyles in the short term.

Implications for public health: Community-led programs generated by, and responsive to, Aboriginal Australians’ needs can demonstrate consistent community reach and sustained program-level lifestyle improvements.

Key words: physical activity, obesity, intervention, Aboriginal and Torres Strait Islander, weight loss

INDIGENOUS HEALTH


PA or sport programs targeting Aboriginal and Torres Strait Islander people, MacNiven found 39 programs addressing PA or sport in adults operating in or since 2012, primarily in the grey literature. Only 25 reported any evaluation data, few are publicly available and not all include health outcomes.12 Of the three including weight-related outcomes, only one was evaluated and showed significant reductions in waist circumference, weight and BMI, and significant increases in intake of vegetables and of fruit.16 Together, these results suggest that effects of community-based PA interventions with Aboriginal Australians can result in modest but significant changes in risk profiles, but more evaluations are needed to build the evidence base.

Impact of lifestyle-related disease on Aboriginal populationsIt is clear that non-communicable chronic diseases (NCDs) continue to be leading contributors to disease burden of Australians.17 They are particularly prevalent among Aboriginal and Torres Strait Islander Australians and are responsible for much of the health gap between Indigenous and non-Indigenous Australian populations.18 In 2013, 29% of Aboriginal and Torres Strait Islander Australian adults were overweight and 37% were obese, these rates being 1.2 and 1.6 times higher, respectively, than in the non-Indigenous population.19 Other contributors to NCD development include dietary and PA habits: 42% of Aboriginal and Torres Strait Islander adults residing in urban and regional areas consume adequate fruit, but only 5% consume the recommended vegetable intake; just over half do not meet the recommended amount of PA, comparable to non-Indigenous populations.19 Smoking rates among Aboriginal and Torres Strait Islander people aged 15 and over have been estimated at 44%, 2.6 times higher than among non-Indigenous people.19 For this population, 37% of the burden of disease is preventable through addressing these and other modifiable risk factors such as alcohol consumption.18,19 Therefore, scaled-up community-wide program evaluations are required to build a comprehensive evidence base for NCD prevention with this population. Evaluation of scaled-up interventions is particularly important as it provides evidence that promising interventions are feasible, appropriate and effective when implemented across diverse communities or, as in the case

of the Knockout Health Challenge (KHC), across the state of New South Wales (NSW). The lag between evidence generation and implementation at scale is a major impediment to health improvement for Aboriginal people, as it denies or delays community access to effective programs.

The NSW Aboriginal Knockout Health ChallengeThe community-based interventions described above were not sustained over time and were relatively small-scale, and the studies evaluating them were published more than 10 years ago. More recently, the NSW Aboriginal KHC has been running team-based competitions in weight loss through PA and healthy eating in NSW, Australia, with up to 830 participants per competition (www.nswknockouthealthchallenge.com.au). Community teams with 20 to 30 Aboriginal adults compete in the KHC for prize money that funds community initiatives in health, sport, nutrition or fitness. Teams are formed through existing social networks and local promotional activities and self-determine the frequency and type of activities they do, but all include one group PA training and one healthy eating activity. In 2012 and 2013, there was one 17-week competition, with two 10-week competitions in 2014 and two 12-week competitions in 2015. The Challenge has links to the Koori Knockout competition (https://www.facebook.com/nswkko/) and the Rugby League more generally, which promote the event. Community engagement is further strengthened through Challenge Town Committees20 comprising volunteers across local government, health services and land councils who support and promote the Challenge in their local area. Further details are provided in the Supplementary Materials and elsewhere.16 An initial evaluation in 2013 showed significant reductions in waist circumference, weight and BMI and significant increases in intake of vegetables and of fruit at the end of the KHC.16

Given the lack of large-scale sustained interventions, in combination with a paucity of peer-reviewed studies on community-based interventions in this population– especially sport-based interventions8 – we compared participation and intervention-level impacts of the KHC over four years. Specifically, we examined whether the KHC continued to attract a relevant demographic and at-risk population, the extent of complete and repeat participation, and whether short-

term impacts on participants’ risk profiles were consistent and sustained across six competitions, 2012 to 2015.

Methods

Study design and data collectionChange in participant outcomes from baseline to completion for each competition was monitored for prize allocation and ongoing delivery of the program using single group pre-post design. Written consent allowed the use of these data for prize calculation and research purposes. This analysis used information from one KHC competition from 2012 and 2013 and two per year in 2014 and 2015. A summary of the data collected over the years 2012–2015 is shown in Table S1. Prior to each competition, participants joined a team and recorded their name, date of birth, and gender on a registration form. From 2013, self-reported current smoking status and fruit and vegetable intake (servings of each on a typical day) was recorded and, from 2014, PA (frequency in last 7 days of 20 minutes or more vigorous PA, 30 min or more of walking, and 30 min or more of moderate PA) was included using validated questions.21 Starting weight (to nearest 0.1kg), height (cm), and waist measurement (cm) were measured objectively by a doctor or registered nurse and also documented on the registration form. Registration and consent forms were collated by team managers and forwarded to the event organisers. At the conclusion of the competition, participants’ weights were recorded by a health professional and self-reported lifestyle risk factors via questionnaire using the same questions as at registration.

ParticipantsParticipants from teams with 20 members or more at registration were included in the analysis. A total of 3,625 participants were registered for the six competitions from 2012 to 2015, representing 2,645 unique people, as 671 (25.4%) people took part in more than one competition.

Data treatment and measures Data for the six competitions were merged, with probabilistic record matching by participant name, sex and date of birth through an independent data linkage agency (The Centre for Health Record Linkage – http://www.cherel.org.au). Where a range was given for health behaviours,

Indigenous Health Aboriginal Australian weight loss challenge


the lowest number was entered (e.g. 2–3 serves of vegetables was recoded as two serves) because people tend to over-estimate healthy behaviours.22-24 Adequate PA was defined as three or more vigorous sessions/week; or five or more walking or moderate sessions/week; or 1–2 vigorous sessions/week and 3–4 walking or moderate sessions/week, according to previous procedures.25 Fruit and vegetable intakes were also categorised for meeting current dietary recommended levels of two serves of fruit and five serves of vegetables per day.26

AnalysisAcross the six competitions, the number of teams, registrants and mean age of registrants were calculated for each competition, along with the proportion in each competition who were male/female, repeat registrants, and completers of the competition (had both start and end weight). The intraclass correlation (ICC) for the primary variable of interest (percentage weight change) within teams ranged between 0.064 (competition 6, 2015) to 0.295 (competition 1, 2012) across the six competitions, demonstrating a clear clustering effect for team, which was stronger for the early KHCs. Over all competitions, the ICC for person was 0.170.

Baseline health characteristics (weight, waist circumference, meeting minimum fruit and vegetable consumption, and sufficient PA) were compared across competitions with 2012 as the reference category for weight and waist circumference, 2013 for smoking, fruit and vegetable intake and 2014 for PA. Crossed random effects rather than nested models were used to account for the clustering of observations within person and within team because people did not uniquely nest within teams but changed teams across competitions.27 Linear models were used for continuous variables, and logit models with QR decomposition were used for dichotomous outcome variables.28

To examine the consistency of short-term impacts, difference scores were calculated by subtracting baseline from post-intervention scores for each competition. Crossed random effects models on the difference scores using the same reference competitions described above were generated; differences were normally distributed so linear models were used for all outcomes. A threshold of 0.05 was used for statistical significance and all analyses were conducted using Stata 14.2 (College Station, TX, USA) and included gender and age (continuous) as fixed effects.

Ethics approval for the secondary analysis was provided by the Aboriginal Health and Medical Research Council (Project #1125/15).

Results

Participation over competitionsParticipation and competition completion rates for the 3,625 participants (including repeat participants) are shown in Table 1.

The number of teams registering sufficient members to start the competition increased across four competitions: 2012, 2013 and the first competitions in 2014 and 2015. Second competitions in 2014 and 2015 attracted many repeat participants (almost half of those registering in 2014) with the majority of repeaters carrying over from the first to the second competitions (207/227 and 178/246, respectively). The number of unique participants for 2014 (both competitions combined) was 1,105 and for 2015 was 1,226. The completion rate dropped from almost three-quarters in the first competition in 2012 to around half to two-thirds of registrants thereafter. The majority of participants were female, and the average age of participants was 39.1 years (SD=12.5).

Registrant profile across competitionsParticipants’ baseline health characteristics for the six competitions are shown in Table 2; beta coefficients with 95%CI comparing across competitions are in Supplementary Table S2.

Baseline weight, adjusted for gender and age, was significantly lower for all competitions compared with 2012 except for competition 3 (beta=-1.09 [95%CI: -2.39, 0.21], p=0.100) and marginally lower in competition 5 (beta=-1.27 [95%CI: -2.61, 0.08], p=0.065). However, all competitions attracted registrants with an average of Class 2 obesity (https://www.cdc.gov/obesity/adult/defining.html). The smallest significant weight loss (adjusted for age and sex) was 1.7kg (competition 3) and largest 2.8kg (competition 4), with BMI following a similar pattern of results. Starting waist circumference only differed significantly from 2012 in competition 4 (2014), by an average of 3.76cm (95%CI: -7.51, -0.01; p=0.049).

In terms of change in other risk factors, there were no significant differences between the first competition where this was measured and subsequent competitions in the proportion of registrants meeting PA or vegetable consumption

Table 1: Participation in Aboriginal Knockout Health Challenges 2012 to 2015.Challenge

Year 2012 2013 2014 2015competition 1 2 3 4 5 6Teamsa 13 22 30 18 33 22Registrants (start) 324 585 828 484 830 574Repeat participants NA 112 (19.2%) 167 (20.2%) 227 (46.9%) 231 (27.8%) 246 (42.3%)Completersb 239 (73.8%) 379 (64.9%) 544 (65.7%) 259 (53.5%) 531 (64.0%) 323 (56.3%)Males 89 (27.5%) 159 (27.7%) 229 (27.7%) 132 (27.3%) 237 (28.6%) 123 (21.6%)Females 235 (72.5%) 416 (72.4%) 599 (72.3%) 352 (72.7%) 592 (71.4%) 447 (78.4%)Age in years (SD) 38.5 (11.7) 40.1 (13.6) 39.4 (12.5) 38.5 (11.9) 38.8 (12.4) 39.0 (12.2)Notes:a: To be eligible to start in a competition a team must have at least 20 registered membersb: Participant had both start and end weight.

Table 2: Participant health characteristics at start of each competition 2012-2015.Competition (year)

At start of competition 1 (2012) N=324

2 (2013) N=585

3 (2014) N=828

4 (2014) N=484

5 (2015) N=830

6 (2015) N=574

Mean weight in kg (SD) 102.2 (21.6) 98.5 (22.4)** 100.9 (22.9) 97.0 (22.9)** 97.1 (24.2) 97.2 (23.0)**

Mean BMI (SD) 37.1 (7.5) 35.7 (7.8)* 36.6 (7.7) 35.1 (7.7)** 35.4 (8.0) 35.6 (8.1)*

Mean waist in cm (SD) 115.8 (15.8) 113.3 (16.8) 115.9 (18.3) 111.0 (18.1)* 111.4 (19.3) 112.6 (18.3)Meet fruit rec n (%) NA 258 (47.4) 367 (57.4)* 239 (61.0)** 367 (44.3) 259 (46.8)Meet veg rec n (%) NA 46 (8.4) 65 (8.7) 63 (14.5) 67 (8.1) 43 (7.8)Sufficient PA n (%) NA NA 416 (51.4) 231 (50.0) 460 (51.9) 264 (47.7)Daily smoker n (%)a NA 153 (28.7) 248 (30.0) 142 (29.3) 254 (31.3) 127 (23.4)Notes:* Significantly different compared with competition 1 (compared with competition 2 for fruit, vegetables and smoking, and competition 3 for PA) at p<0.05,

** at p<0.01a: The model did not converge for current smoker; no formal comparison available.

Grunseit et al. Article



recommendations at baseline; the latter was consistently very low with less than 10% meeting vegetable guidelines for all except one competition (Table 2). However, registrants had higher odds of meeting fruit recommendations in competitions 3 (adjusted odds ratio (AOR)=1.58; 95%CI: 1.09, 2.27) and 4 (AOR=1.92; 95%CI: 1.27, 2.92) than competition 2. Smoking rates appeared similar for the first four competitions and dropped in the last competition.

Change compared across competitionsTable 3 shows health characteristics at the beginning and end for each competition. Beta coefficients and 95%CI for comparisons of the magnitude of change pre- to post-competition compared with competition 1 (competition 2 for fruit and vegetable consumption and competition 3 for the PA measures) are given in Table S3.

Despite the varying competition duration across the years (see methods above), there were few discernible systematic differences in competition outcomes. The analysis showed that although short-term impacts were modest, there were consistent changes across the years in the direction of improvements in health indicators. In detail, the amount of weight lost was greatest in competition 1 whether measured by per cent weight lost, kilograms or BMI (Tables 3 and S3). Participants who completed competition 1 lost on average about 5% of their starting body weight, but the average percentage weight lost for subsequent competitions was between 1.9% and 2.5%, adjusted for sex and age (Table S3). A supplementary analysis including only new (as opposed to

repeat) registrants showed the same pattern of effects (Table S4). By contrast, the change in fruit and vegetable consumption and sessions of walking, moderate and vigorous PA were comparable across the competitions. Only in competition 6 was there a marginally higher increase in vegetable consumption compared with competition 2, when it was first measured (beta= 0.37, 95%CI: -0.03, 0.77; p=0.072).

Discussion

Community engagement in program design and implementation is considered a critical determinant of program effectiveness of healthy lifestyle programs for Aboriginal and Torres Strait Islander people,14,29 yet few of these types of programs are evaluated and reported in the peer review literature. The KHC is an example of a community-led (and government-supported) intervention run over successive years and reaching many hundreds of Aboriginal Australians at risk of chronic disease due to lifestyle risk factors. An initial evaluation of the 2013 KHC indicated participants felt they benefitted not only physically, but also from feeling more socially connected, with improved self-esteem, reduced stress and better linkages with their local Aboriginal Medical Service.20 Our study of six successive competitions during 2012–2015 extends these findings by examining temporal changes in participation and impact effect size, making this one of the few studies to examine sustainability of healthy lifestyle programs for Aboriginal Australians.

It is clear the KHC attracts and retains the interest and active participation of the target population. Increasing participation

may in part be due to efforts made by KHC organisers to incentivise, promote and adapt the intervention over time. For example, there are start-up funds for teams and free t-shirts, caps and water bottles for individual participants, a strategy found to be effective in engaging Aboriginal people in the 1 Deadly Step program.30 Moreover, the KHC is led by local Aboriginal communities and participants are local Aboriginal people. The program is promoted by the NSW Aboriginal Rugby League Knockout, which engenders team spirit according to previous research,20 and prize funds can be used to support local teams joining that competition, leveraging one sport intervention with another. KHC program staff visit teams, informally gathering information about what program components work well and what needs to change. Therefore, the sustainability and growth of KHC may be due to the flexibility and adaptability required for successful intervention on complex issues in populations coping with multiple disadvantage.13,14

In years with two competitions (2014 and 2015), the second competition attracted a smaller number of participants than the first. The second competition in a year runs in the coldest months which likely contributes to the reduced participation, as previous research shows PA participation drops in colder months31; the majority of participating teams are based in regional areas, which are subject to colder temperatures and may have less access to all-weather facilities. Organisers could consider the feasibility of partnering with organisations with indoor facilities to enhance the attractiveness of the second competition in each year.

Completion rates in the KHC were highest for the first competition (74%) and subsequently were between 55% and 65%. Attrition rates among weight loss interventions in the literature vary between 23% and 90% and depend on intervention characteristics, setting, population and length of the intervention.32 The vast majority of research examining attrition in weight loss interventions targets mainstream populations in intensive clinic-based programs. One community-based six-month weight loss program reported a 47% completion rate,33 lower than that found here and perhaps demonstrating the strength of the group-based approach of the KHC.34 Team-based weight loss has been effective in other populations,9 but may be particularly effective for retention in the KHC because

Table 3: Pre and post participant health characteristics from pre to post competition 2012-2015 and percent weight change.

Competition (year)Outcome 1 (2012)

N=2392 (2013) N=379

3 (2014) N=544

4 (2014) N=259

5 (2015) N=531

6 (2015) N=324

% weight change -4.7 -2.3** -2.2** -2.8 -2.7** -2.6**

pre post pre post pre post pre post pre post pre postWeight kg 103.0 98.1 97.2 95.0** 101.0 98.7** 95.2 92.6** 98.4 95.7** 97.0 94.5**

BMI 37.5 35.8 35.1 34.4** 36.0 35.2** 33.9 33.2** 35.3 34.4** 35.0 34.0**

Fruit serves/day NA 1.6 2.0 1.8 2.2 1.9 2.1 1.3 1.7 1.5 1.9Veg serves/day NA 2.3 2.8 2.2 2.8 2.7 3.0 2.0 2.6 2.1 3.0Walking NA NA 2.2 3.0 1.9 2.6 2.1 2.7 1.9 2.8Moderate PA NA NA 1.6 2.5 1.7 2.4 1.6 2.2 1.5 2.2Vigorous PA NA NA 1.8 2.5 1.6 2.2 1.9 2.5 1.7 2.4Notes:* Significantly different compared with competition 1 (compared with competition 2 for fruit and vegetables, and competition 3 for walking, moderate and

vigorous PA) at p<0.05, ** at p<0.01


the format capitalises on local pride and social connections.6,35 The completion rate observed for the KHC therefore fares well and may even be slightly underestimated, because completion was defined as providing both start weight and final weight, and participants may take part in KHC activities until the end but not provide a final weight.

The health profile for eating, smoking and PA behaviours of registrants attracted to subsequent competitions showed no difference compared to the first competition where that behaviour was measured, but registrants weighed significantly less in competitions 2–6 compared with the first. It is unclear why; we did not observe a sustained downward trend. One possible explanation is that those who were most motivated and most in need participated in the first KHC. Despite this, around three-quarters of KHC registrants met the definition of obesity. Further, prevalence of meeting guidelines, especially for vegetable consumption, was low, demonstrating the KHC consistently attracts those at high risk of chronic disease. KHC registrants also reflect the broader Aboriginal population in terms of PA and fruit intake, although they fare somewhat better on smoking rates and recommendations for vegetable consumption.19 The KHC is therefore well-targeted.

The predominance of females (up to 78%) is disproportionate to their representation in the ‘at-risk’ population,36 but reflects other volunteer weight loss programs.37 Aboriginal and Torres Strait Islander men have poor health-seeking behaviours compared to their female counterparts, for reasons ranging across cultural (e.g. traditional gender-related law) and societal (sex-specific difference in health) factors.38 Introduction of measures to increase male participation in future Challenges are warranted and could include informal consultations to better understand the program needs of Aboriginal men; actively engaging more men in team management and program coordination and support roles; or enlisting male Aboriginal health workers to promote the KHC as used in other settings.39

Losing at least 5% of body weight is considered to be clinically meaningful40 and while this threshold was not achieved, on average, the percentage lost is comparable with other lifestyle behaviour change programs.41,42 For four competitions it exceeded 2.5%, where benefits for glycaemic measures start to improve,43 despite an increasing proportion of repeat participation.

Research is mixed as to whether repeat weight loss attempts are associated with greater likelihood of weight loss.44,45 Worick et al. (1993) reported that although repeat participants in their worksite team weight loss annual competition lost weight, they also risked ‘weight cycling’, whereby weight lost during the competition was regained in the inter-competition period.46 The 2013 KHC evaluation showed almost one-third of those followed up regained weight nine months post-competition.20 Our analysis showed that the weight lost and change in diet and PA behaviours were similar whether the analysis was on all participants or confined to only those who were new to the KHC. Future research could examine within-individual patterns of weight from the end of one competition to the start of the next to further investigate inter-competition regain and maintenance and associated correlates.

KHC, despite being a ‘weight loss challenge’, targets other healthy lifestyle factors and showed consistent improvements across six competitions, including increased fruit and vegetable intake and increased proportion of people achieving recommended levels of PA, which benefit health. Previous epidemiological studies have shown that those who meet recommendations for PA have better health outcomes than those who do not within the same weight class.47 Smoking behaviour was not a focus of the intervention until 2018, when referrals to a smoking cessation program were formally included. Future analyses may examine changes in smoking outcomes among participants.

Strengths and limitationsDespite growing evidence for effective community-based lifestyle interventions for Aboriginal people, previous studies have been conducted in a single Aboriginal community, with one notable exception – the evaluation of the 2013 KHC. Our evaluation across six different competitions offers unique insights into program implementation at scale and under real-world conditions, addressing a gap in the intervention research evidence base.48 Limitations include, first, the absence of a control group, meaning the effects observed may be unrelated to KHC participation but are unlikely in the absence of any intervention. Second, attrition rates, although comparable to other weight loss interventions, were high and those lost to follow-up may have lost less weight or not made behavioural changes, risking an

overestimation of impact. However, given the consistency of the findings over the six competitions, it is more likely that the effects were associated with participation. Thirdly, there was no information retained on teams that did not make the minimum 20-registrant cut-off point at the start of each competition; these data have been collected since 2016. Finally, because KHC was a real-world competition and relied on non-research staff to collate data, baseline estimates of behaviours in some cases may have taken place once competition activities commenced, thereby possibly biasing impact estimates downwards. Further, self-report measures may also introduce biases towards more healthy behaviours; however, the primary outcome of weight was objectively measured at both pre- and post-intervention and showed patterns consistent with the other self-reported health behaviours.

Conclusion

The KHC has shown to promote (at least) short-term reductions in weight and improvements in lifestyle-related risk factors promoting healthy lifestyles among Aboriginal communities in NSW. Addressing the key chronic disease risk factors, the KHC has potential to make an important contribution to closing the gap in health outcomes between Aboriginal and non-Indigenous people. Future research should explore characteristics of non-completers and qualitatively explore non-completion, reduced participation in the second competition in the year, and factors that hinder or encourage male participation in the KHC. Finally, future analyses should focus on repeating participants and their patterns of weight maintenance, regain or further loss between finishing and starting a new competition.

Acknowledgements

The authors would like to acknowledge Dr Erin Passmore’s contribution to initiating the project and facilitating access to the data.

Data availabilityThe datasets generated and/or analysed during the current study are not publicly available due to the conditions of ethics approval. Data are available from the authors upon reasonable request and with permission of the NSW Ministry of Health.

Grunseit et al. Article


FundingThe work was funded by the National Health and Medical Research Council of Australia (NHMRC) through its Partnership Centre grant scheme [grant number GNT9100001]. NSW Health, ACT Health, the Australian Government Department of Health, the Hospitals Contribution Fund of Australia and the HCF Research Foundation have contributed funds to support this work as part of the NHMRC Partnership Centre grant scheme. The contents of this paper are solely the responsibility of the individual authors and do not reflect the views of the NHMRC or funding partners.

References1. Canuto K, Cargo M, Li M, D’Onise K, Esterman A,

McDermott R. Pragmatic randomised trial of a 12-week exercise and nutrition program for Aboriginal and Torres Strait Islander women: Clinical results immediate post and 3 months follow-up. BMC Public Health. 2012;12(1):933.

2. Mendham AE, Duffield R, Marino F, Coutts AJ. A 12-week sports-based exercise programme for inactive Indigenous Australian men improved clinical risk factors associated with type 2 diabetes mellitus. J Sci Med Sport. 2015;18(4):438-43.

3. Davey M, Moore W, Walters J. Tasmanian Aborigines step up to health: Evaluation of a cardiopulmonary rehabilitation and secondary prevention program. BMC Health Serv Res. 2014;14:349-58.

4. Dimer L, Dowling T, Jones J, Cheetham C, Thomas T, Smith J, et al. Build it and they will come: Outcomes from a successful cardiac rehabilitation program at an Aboriginal Medical Service. Aust Health Rev. 2013;37:79-82.

5. O’Dea K. Marked improvement in carbohydrate and lipid metabolism in diabetic Australian Aborigines after temporary reversion to traditional lifestyle. Diabetes. 1984;33(6):596-603.

6. Browne-Yung K, Ziersch A, Baum F, Gallaher G. Aboriginal Australians’ experience of social capital and its relevance to health and wellbeing in urban settings. Soc Sci Med. 2013;97:20-8.

7. Stronach M, Maxwell H, Taylor T. ‘Sistas’ and Aunties: sport, physical activity, and Indigenous Australian women. Ann Leis Res. 2016;19(1):7-26.

8. Standing Committee on Aboriginal and Torres Strait Islander Affairs. Sport - More Than Just a Game. Canberra (AUST): Parliament of Australia House of Representatives; 2013. p. 1-123.

9. Leahey TM, Kumar R, Weinberg BM, Wing RR. Teammates and social influence affect weight loss outcomes in a team-based weight loss competition. Obesity. 2012;20(7):1413-18.

10. McMahon NE, Visram S, Connell LA. Mechanisms of change of a novel weight loss programme provided by a third sector organisation: A qualitative interview study. BMC Public Health. 2016;16(1):1-11.

11. Pressick EL, Gray MA, Cole RL, Burkett BJ. A systematic review on research into the effectiveness of group-based sport and exercise programs designed for Indigenous adults. J Sci Med Sport. 2016;19(9):726-32.

12. Macniven R, Elwell M, Ride K, Bauman A, Richards J. A snapshot of physical activity programs targeting Aboriginal and Torres Strait Islander people in Australia. Health Promot J Austr. 2017;28(3):185-206.

13. Egger G, Fisher G, Piers S, Bedford K, Morseau G, Sabasio S, et al. Abdominal obesity reduction in Indigenous men. Int J Obes. 1999;23:564-9.

14. Rowley KG, Daniel M, Skinner K, Skinner M, White GA, O’Dea K. Effectiveness of a community‐directed ‘healthy lifestyle’program in a remote Australian Aboriginal community. Aust N Z J Public Health. 2000;24(2):136-44.

15. Chan L, Ware R, Kesting J, Marczak M, Good D, Shaw J. Short term efficacy of a lifestyle intervention programme on cardiovascular health outcome in overweight Indigenous Australians with and without type 2 diabetes mellitus: The healthy lifestyle programme (HELP). Diabetes Res Clin Pract. 2007;75:65-71.

16. Passmore E, Shepherd B, Milat A, Maher L, Hennessey K, Havrlant R, et al. The impact of a community-led program promoting weight loss and healthy living in Aboriginal communities: The New South Wales Knockout Health Challenge. BMC Public Health. 2017;17(1):951-60.

17. Australian Institute of Health and Welfare. Australian Burden of Disease Study: Impact and Causes of Illness and Death in Australia 2011. Canberra (AUST): AIHW; 2016.

18. Australian Institute of Health and Welfare. Australian Burden of Disease Study: Impact and Causes of Illness and Death in Aboriginal and Torres Strait Islander People 2011. Canberra (AUST): AIHW; 2016.

19. Australian Institute of Health and Welfare. The Health and Welfare of Australia’s Aboriginal and Torres Strait Islander Peoples. Canberra (AUST): AIHW; 2015.

20. ARTD Consultants. Evaluation of the NSW Knockout Health Challenge 2013. Sydney (AUST): ARTD; 2013.

21. Dal Grande E, Fullerton S, Taylor AW. Reliability of self-reported health risk factors and chronic conditions questions collected using the telephone in South Australia, Australia. BMC Med Res Methodol. 2012;12(1):108-17.

22. Miller TM, Abdel-Maksoud MF, Crane LA, Marcus AC, Byers TE. Effects of social approval bias on self-reported fruit and vegetable consumption: A randomized controlled trial. Nutr J. 2008;7(1):1-7.

23. McDonald SP, Maguire GP, Hoy WE. Validation of self-reported cigarette smoking in a remote Australian Aboriginal community. Aust N Z J Public Health. 2003;27(1):57-60.

24. Warner ET, Wolin KY, Duncan DT, Heil DP, Askew S, Bennett GG. Differential accuracy of physical activity self-report by weight status. Am J Health Behav. 2012;36(2):168-78.

25. Smith BJ, Marshall AL, Huang N. Screening for physical activity in family practice: Evaluation of two brief assessment tools. Am J Prev Med. 2005;29(4):256-64.

26. National Health and Medical Research Council. Eat for Health: Australian Dietary Guidelines. Canberra (AUST): NHMRC; 2013.

27. STATA Multilevel Mixed Effects Reference Manual. Release 13. College Station (TX): Stata Press; 2013.

28. Rabe-Hesketh S, Skrondal A. Multilevel Mixed-effects Logistic Regression (QR Decomposition). Multilevel Mixed-Effects Reference Manual. Release 15. College Station (TX): Stata Press; 2017.

29. Schembri L, Curran J, Collins L, Pelinovskaia M, Bell H, Richardson C, et al. The effect of nutrition education on nutrition‐related health outcomes of Aboriginal and Torres Strait Islander people: A systematic review. Aust N Z J Public Health. 2016;40 Suppl 1:42-7.

30. Peiris D, Wright L, Corcoran K, News M, Turnbull F. 1 Deadly Step: Process Evaluation of a Chronic Disease Screening Program in NSW Aboriginal Communities. Chatswood (AUST): State Government of New South Wales Agency for Clinical Innovation; 2014.

31. Tucker P, Gilliland J. The effect of season and weather on physical activity: A systematic review. Public Health. 2007;121(12):909-22.

32. Moroshko I, Brennan L, O’Brien P. Predictors of dropout in weight loss interventions: A systematic review of the literature. Obes Rev. 2011;12(11):912-34.

33. Graffagnino CL, Falko JM, Londe M, Schaumburg J, Hyek MF, Shaffer LE, et al. Effect of a community‐based weight management program on weight loss and cardiovascular disease risk factors. Obesity. 2006;14(2):280-8.

34. Minniti A, Bissoli L, Di Francesco V, Fantin F, Mandragona R, Olivieri M, et al. Individual versus group therapy for obesity: Comparison of dropout rate and treatment outcome. Eat Weight Disord. 2007;12(4):161-7.

35. Doyle J, Firebrace B, Reilly R, Crumpen T, Rowley K. What makes us different? The role of Rumbalara Football and Netball Club in promoting Indigenous wellbeing. Aust Community Psychol. 2013;25(2):7-21.

36. Australian Bureau of Statistics. 4727.0.55.006 - Australian Aboriginal and Torres Strait Islander Health Survey: Updated Results, 2012–13. Canberra (AUST): ABS; 2014.

37. Fink JT, Smith DR, Singh M, Ihrke DM, Cisler RA. Obese employee participation patterns in a wellness program. Popul Health Manag. 2016;19(2):132-5.

38. Wenitong M, Adams M, Holden CA. Engaging Aboriginal and Torres Strait Islander men in primary care settings. Med J Aust. 2014;200(11):632-3.

39. Andrology Australia. Engaging Aboriginal and Torres Strait Islander Males in Primary Care Settings. Melbourne (AUST): Monash University School of Public Health and Preventive Medicine Andrology Australia; 2018.

40. Magkos F, Fraterrigo G, Yoshino J, Luecking C, Kirbach K, Kelly SC, et al. Effects of moderate and subsequent progressive weight loss on metabolic function and adipose tissue biology in humans with obesity. Cell Metab. 2016;23(4):591-601.

41. Vita P, Cardona-Morrell M, Bauman A, Singh MF, Moore M, Pennock R, et al. Type 2 diabetes prevention in the community: 12-month outcomes from the Sydney Diabetes Prevention Program. Diabetes Res Clin Pract. 2016;112:13-19.

42. O’Hara BJ, Phongsavan P, Eakin EG, Develin E, Smith J, Greenaway M, et al. Effectiveness of Australia’s Get Healthy Information and Coaching Service®: Maintenance of self-reported anthropometric and behavioural changes after program completion. BMC Public Health. 2013;13(1):175-88.

43. Williamson DA, Bray GA, Ryan DH. Is 5% weight loss a satisfactory criterion to define clinically significant weight loss? Obesity. 2015;23(12):2319-20.

44. Latner JD, Ciao AC. Weight-loss history as a predictor of obesity treatment outcome: Prospective, long-term results from behavioural, group self-help treatment. J Health Psychol. 2014;19(2):253-61.

45. Venditti EM, Bray GA, Carrion-Petersen ML, Delahanty LM, Edelstein SL, Hamman RF, et al. First versus repeat treatment with a lifestyle intervention program: Attendance and weight loss outcomes. Int J Obes. 2008;32(10):1537-44.

46. Worick A, Petersons M. Weight loss contests at the worksite: Results of repeat participation. J Am Diet Assoc. 1993;93(6):680-1.

47. Ortega FB, Ruiz JR, Labayen I, Lavie CJ, Blair SN. The Fat but Fit paradox: What we know and don’t know about it. Br J Sports Med. 2018;52(3):151-3.

48. Milat AJ, Bauman AE, Redman S, Curac N. Public health research outputs from efficacy to dissemination: A bibliometric analysis. BMC Public Health. 2011;11(1):934-42.



Supplementary Table 1: Summary of data collection.

Supplementary Table 2: Adjusted beta coefficients (β) and adjusted odds ratios (AOR) for comparisons of participant health characteristics at start of each competition across competitions 2012-2015.

Supplementary Table 3: Adjusted beta coefficients (β) and adjusted odds ratios (AOR) for pre- post difference compared with reference competition.

Supplementary Table 4: Change in participant health characteristics from pre to post competition 2013-2015 among new participants within each competition.



The national population-based breast screening program, BreastScreen Australia, was implemented in 1991

with the objective of reducing breast cancer mortality through early detection of asymptomatic breast cancers using mammography. The target participation rate for biennial screening is 70% of all women aged 50–74 years, with the program having extended the upper target age range from 69 to 74 years from 2013. However, this objective has not been met with rates remaining around 18% to 15% lower than the target since 1996 when uptake data began to be calculated.1-3 In addition, inequities between subgroups of women exist with Aboriginal and Torres Strait Islander women, referred to hereafter as Indigenous women, and women living remotely having the lowest rates of attendance at screening and relatively poorer breast cancer outcomes.4 The efficacy of breast screening in reducing mortality depends upon adequate population coverage,5 therefore improving participation across all groups of women is crucial to effective national cancer control.

The BreastScreen program was implemented in the Northern Territory (NT) in 1994. The NT has the smallest population in Australia dispersed across a geographical area that makes up 17.5% of the country’s total land mass.6 It has the lowest population density in Australia with the capital city, Darwin, classified as ‘outer regional’ based on the Australian Statistical Geographical

Standards Remoteness Areas classification (ASGSRA).7 Furthermore, the NT has the highest proportion of Indigenous residents compared to other Australian states or territories, with up to 30% of the population identifying as Aboriginal and/or Torres Strait Islander. Regarding breast screening, the NT consistently has the lowest participation rates in the nation. NT Indigenous attendance is less than half of the national rate (24% vs. 55%, respectively) and the overall NT

attendance (41%) is lower than all states and territories combined.2,8 Breast cancer is the most common non-melanoma cancer in Australian women and the number one cancer in NT women.8 Therefore, with evidence that screening prevents 43 deaths in 10,000 women screened,9 concerted efforts to improve screening attendance are needed.

A lower uptake of mammographic screening is not uncommon among Indigenous peoples around the world and the reasons

Breast screening attendance of Aboriginal and Torres Strait Islander women in the Northern Territory of AustraliaKriscia A. Tapia,1 Gail Garvey,1,2 Mark F. McEntee,3 Mary Rickard,1,4 Lorraine Lydiard,5 Patrick C. Brennan1

1. Faculty of Health Sciences, The University of Sydney, New South Wales2. Charles Darwin University, Wellbeing and Preventable Chronic Diseases Division Menzies School of Health Research, Northern Territory3. Department of Medicine, University College Cork, Ireland4. BreastScreen New South Wales5. BreastScreen Northern TerritoryCorrespondence to: Ms Kriscia A. Tapia: Faculty of Health Sciences, The University of Sydney, New South Wales; e-mail: [email protected]: December 2018; Revision requested: April 2019; Accepted: May 2019 The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,



Abstract

Objective: To compare breast screening attendances of Indigenous and non-Indigenous women.

Methods: A total of 4,093 BreastScreen cases were used including 857 self-identified Indigenous women. Chi-squared analysis compared data between Indigenous and non-Indigenous women. Logistic regression was used for groupings based on visits-to-screening frequency. Odds ratios and 95% confidence intervals were calculated for associations with low attendance.

Results: Indigenous women were younger and had fewer visits to screening compared with non-Indigenous women. Non-English speaking was mainly associated with fewer visits for Indigenous women only (OR 1.9, 95%CI 1.3-2.9). Living remotely was associated with fewer visits for non-Indigenous women only (OR 1.3, 95%CI 1.1-1.5). Shared predictors were younger age (OR 12.3, 95%CI 8.1-18.8; and OR 11.5, 95%CI 9.6-13.7, respectively) and having no family history of breast cancer (OR 2.1, 95%CI 1.3-3.3; and OR 1.8, 95%CI 1.5-2.1, respectively).

Conclusions: Factors associated with fewer visits to screening were similar for both groups of women, except for language which was significant only for Indigenous women, and remoteness which was significant only for non-Indigenous women.

Implications for public health: Health communication in Indigenous languages may be key in encouraging participation and retaining Indigenous women in BreastScreen; improving access for remote-living non-Indigenous women should also be addressed.

Key words: breast cancer, screening, participation, Aboriginal and Torres Strait Islander

INDIGENOUS HEALTH


for this are multifaceted. While some reasons for non-participation by Indigenous and non-Indigenous women may be similar, cultural beliefs around breast cancer and fatalistic views on health are reported as having significant influence on Indigenous screening behaviours.10 In addition, systematic barriers to screening are also evident for Indigenous Australians, including geographic isolation, lack of means of transportation to attend services, and a shortage of culturally competent facilitators to screening.11,12 Evidence in the US, Canada, Alaska, Pacific Islands, and New Zealand show that Indigenous women are more likely to be underrepresented in breast screening13-15 and have higher breast cancer mortality rates compared with non-Indigenous women.14-16 A similar scenario is reported for Indigenous Australian women who, despite lower breast cancer incidence, have poorer health outcomes and higher rates of death from breast cancer when compared with other Australian women.17 Australian Indigenous women are also younger and more likely to have advanced tumours at the time of diagnosis compared with non-Indigenous women,18-20 making treatment options limited and the tumours potentially more aggressive. National and local strategies to improve accessibility have been implemented with some success, such as the mobile screening vans that travel to remote communities,21 the process of block bookings of appointments for Indigenous women,11 evidence-based and culturally sensitive materials developed by Indigenous health experts,22 and involvement of Indigenous health care workers.23 Despite these efforts, Indigenous Australians’ attendance at screening remains around 16% lower compared with non-Indigenous women (37% vs. 53%, respectively, in 2015) .

There is currently limited information on the screening characteristics of women in the NT as they relate to indigeneity and attendance. The aim of the current work is to investigate variables associated with attendance at BreastScreen for women in the NT. With significant differences shown previously between NT women in the screening population,24 a further aim is to measure the variations between Indigenous and non-Indigenous women’s screening attendances.

Methods

Data collectionEthical approval was obtained from the Human Research Ethics Committee (HREC)

of the NT Department of Health and Menzies School of Health Research (HREC 2016-2627). Written consent to use personal information for evaluation and research was collected from women prior to having a screening mammogram and only consenting women’s data was made available to researchers. This consent request is written in the client information form routinely collected by BreastScreen NT, a population screening program.

The study was performed retrospectively using a client data sample retrieved from BreastScreen NT. The sample consisted of 4,093 women (857 self-identified as Indigenous and 3,236 reported being non-Indigenous) aged between 40 and 85 years who were screened between 30 March and 24 November 2015. BreastScreen NT sends postal invitations to women aged 50 to 74 years old to attend screening every two years; however, screening is free for women from 40 years of age. Mammograms were performed at permanent screening facilities in the NT in Darwin, Palmerston, and Alice Springs; women located in remote to very remote communities in the NT were screened via the BreastScreen NT mobile bus unit. Digital image files were sent electronically to Sydney Breast Clinic (SBC) in NSW for radiologist interpretation.

Women’s radiologist-reported findings and self-reported personal details such as Indigenous status, date of birth, residential address, main language spoken, family history of breast cancer, previous breast cancer, current breast lump, and use of hormone replacement therapy (HRT) in the past six months were stored on the NT Department of Health computerised database. NT Department of Health personnel extracted the data and provided de-identified information to researchers. The number of screening rounds that a woman has attended was generated by the NT database system based on number of entries. Women’s residential postcodes were categorised by the researchers based on the ASGSRA classification. In the NT, only three categories are available: outer regional, remote and very remote.

Data analysisIn the first stage of analysis, base-line differences between Indigenous and non-Indigenous women’s characteristics and screening attendances were explored. Next, Receiver Operating Characteristics (ROC) curve analyses were employed for number of

visits to screening and ages to determine cut-off points for these variables for Indigenous and non-Indigenous women. Using these cut-off points, chi-squared tests were used to derive odds ratios (OR) and 95% confidence intervals (CI). A p-value <0.05 was considered significant.

In the second stage of analysis, we investigated Indigenous and non-Indigenous women separately and focused on potential associations with numbers of visits to screening above and below the cut-off point. Categorical variables such as age group, previous breast cancer diagnosis, family history of breast cancer, main language, geographic remoteness, current breast lump, and case outcome were analysed using chi-squared tests. Logistic regression was used to derive odd ratios (OR) and 95% and confidence intervals (CI). Next, multivariate stepwise logistic regression was performed on variables with univariate significance to determine predictive factors for low screening attendance. Factors with p values <0.02 were retained in the model.

BM SPSS© version 24 statistical software was used for the analyses.

Results

Women in this study had visited screening between one and 21 times in their lifetime. That is, for some women, this was their first time attending screening, for those on the highest end of the range it was their twenty-first visit, and other women ranged somewhere in between. Figure 1 displays the proportions of Indigenous and non-Indigenous women according to the number of times they have attended screening.

Table 1 shows that Indigenous women had fewer visits to screening compared with non-Indigenous women with medians of two visits (IQR 1-3) and three visits (IQR 2-7), respectively. Indigenous women were younger than non-Indigenous women with median ages of 54 years (IQR 48-60 years) and 57 years (IQR 52-63 years), respectively.

There was a higher proportion of Indigenous women residing in remote areas (67.7%) compared with outer regional areas (13.3%), and 71.3% of Indigenous women mainly spoke another language at home. In contrast, non-Indigenous women had more than half of the population (56.9%) living in outer regional areas than in remote locations (36.2%), and 84.2% mainly spoke English at home. Both groups of women reported similar experiences with HRT use

Indigenous Health Breast screening attendance in the Northern Territory


and personal and family histories of breast cancer with greater proportions answering negatively. Further similarities were shown in that larger proportions of Indigenous and non-Indigenous women reported no breast lump at screening, and majorities in both groups had normal mammograms at this round of screening.

ROC curve analyses determined that 55 years and 3 visits were the cut off-points for age and frequency of attendance at screening, respectively, as shown in Table 2. Indigenous women were 1.8 times as likely to be under 55 years of age (OR 1.8, 95%CI 1.5–2.0; p<0.001) and more than 3 times as likely to have fewer than 3 visits to screening (OR 3.3, 95%CI 2.8–3.9; p<0.001) than non-Indigenous women.

Table 3 shows the unadjusted results of the two-tailed tests based on screening visits above and below the cut-off point. The following describes the variables associated with low screening attendance (<3 visits). Women were likely to have attended screening less than 3 times if: they were younger than 55 years compared with older women (Indigenous: OR 10.8, 95%CI 7.4–15.7; p<0.001; and non-Indigenous: OR 10.7, 95%CI 9–12.6; p<0.001), had not had a previous breast cancer diagnosis compared with women who had a been diagnosed with breast cancer in the past (Indigenous: OR 7.5, 95%CI 2.0–28.0; p<0.001; and non-Indigenous: OR 2.0, 95%CI 1.2–3.4; p<0.05); live in remote locations compared with non-remote women (Indigenous: OR 1.5, 95%CI 1.1–2.4; p<0.05; and non-Indigenous: OR 1.5, 95%CI 1.3–1.7; p<0.001), and if they do not have a family history of breast cancer compared with women who do (Indigenous: OR 2.3, 95%CI 1.6–3.3; p<0.001; and non-Indigenous: OR 1.3, 95%CI 1.1–1.5; p<0.001).

Speaking a main language other than English was associated with low attendance for Indigenous women (OR 2.3, 95%CI 1.7–3.2; p<0.001) but not for non-Indigenous women, while presenting with a current breast lump was significant for non-Indigenous women (OR 1.8, 95%CI 1.3–2.7; p<0.05) but not for Indigenous women.

Both Indigenous and non-Indigenous women whose cases were recalled to assessment at the time of data collection were likely to have had fewer visits to screening than cases that were reported as normal (Indigenous: OR 5.4, 95%CI 2.1–13.6; p<0.001; non-Indigenous: OR 1.9, 95%CI 1.4–2.6; p<0.001).

Multiple logistic regression analysis reported that significant predictors for low screening

Figure 1 Distribution of Indigenous (n=857) and non-Indigenous women (n=3236) per

number of visits to BreastScreen NT in 2015.

45 40 35 30 25 20 15 10 5 0 5 10 15 20 25 30 35 40 45

123456789

101112131415161718192021

Proportion of women (%)

Vis

it nu

mbe

r

Screening attendance of Indigenous and non-Indigenous women

Indigenous

Non‐Indigenous

Figure 1: Distribution of Indigenous (n=857) and non-Indigenous women (n=3236) per number of visits to BreastScreen NT in 2015.

Table 1: Characteristics of Indigenous and non-Indigenous women screened in the NT in 2015.Variables Indigenous

N=857Non-Indigenous

N=3,236Median age (y) at screening (Q1, Q3) 54 (48,60) 57 (52,63) min, max age (y) at screening 40,79 40,85Median number of screening visits (Q1, Q3) 2 (1,3) 3 (2,7) min, max number of screening visits 1,19 1,21Place of residencea (%) Outer regional 118 (13.3) 1,823 (56.9) Remote 579 (67.7) 1,172 (36.2) Very remote 159 (18.6) 209 (6.5)Main languageb (%) English 245 (28.7) 2,721 (84.2) Other-language 610 (71.3) 511 (15.8)HRT use within 6 monthsc (%) No 840 (98) 2,950 (91.4) Yes 17 (2) 278 (8.6)Family history of breast cancerd (%) No 613 (80.7) 1,961 (66.1) Yes 147 (19.3) 1,007 (33.9)Previous breast cancer (%) No 485 (98.6) 3,161 (97.7) Yes 12 (1.4) 75 (2.3)Current breast lump (%) No 828 (96.6) 3,121 (96.4) Yes 29 (3.4) 115 (3.6)Case decision (%) Normal 729 (92.4) 3,039 (93.9) Recalled 65 (7.6) 197 (6.1)Notes: a: Visitors (n=33) were excluded b: Not known (n=6) was excluded c: Not known (n=1) was excluded d: Not known (n=365) were excluded

Tapia et al. Article



Table 2: OR and 95% CI of Indigenous and non-Indigenous women’s cut-off ages and number of visits to screening.Variables Indigenous

N=857non-Indigenous

N=3,236P value OR (95% CI)

<55 years 458 (53.4%) 1,269 (39.2%) < 0.0001 1.8 (1.5-2.0)≥55 years 399 (46.6%) 1,967 (60.8%)<3 visits 607 (70.8%) 1,360 (42%) < 0.0001 3.3 (2.8-3.9)

≥3 visits 250 (29.2%) 1,876 (58%)Notes:P values obtained from Chi-squared test

Table 3: Association of factors with screening attendance for Indigenous (n=857) and non-Indigenous (n=3,236) women in BreastScreen NT.Variable Indigenous Non-Indigenous

Less than 3 visits N (%)

3 or more visits N (%)

P value OR (95% CI) reference is 1

Less than 3 visits N (%)

3 or more visits N (%)

P value OR (95% CI) reference is 1

Age <55 years 416 (90.8) 42 (9.2) < 0.001 10.8 (7.4-15.7) 942 (74.2) 327 (25.8) < 0.001 10.7 (9-12.6) ≥55 years 191 (47.9) 208 (52.1) 418 (21.3) 1,549 (78.7)Previous breast cancer No 604 (71.5) 241 (28.5) < 0.001 7.5 (2.0-28.0) 1,340 (42.4) 1,821 (57.6) < 0.05 2.0 (1.2-3.4) Yes 3 (25) 9 (75) 20 (26.7) 55 (73.3)Case decision Recalled 60 (92.3) 5 (7.7) < 0.001 5.4 (2.1-13.6) 113 (57.4) 84 (42.6) < 0.001 1.9 (1.4-2.6) Normal 547 (69.1) 245 (30.9) 1,247 (41) 1,792 (59)Family history of breast cancer No 461 (75.2) 152 (24.8) < 0.001 2.3 (1.6-3.3) 856 (43.7) 1,105 (56.3) < 0.05 1.3 (1.1-1.5) Yes 84 (57.1) 63 (42.9) 374 (37.1) 633 (62.9)Main language Other language 464 (76.1) 146 (23.9) < 0.001 2.3 (1.7-3.2) 199 (38.9) 312 (61.1) 0.129 0.8 (0.7-1.0) English 141 (57.6%) 104 (42.4) 1,158 (42.6) 1,563 (57.4)Place of residence Remote 533 (72.2) 205 (27.8) < 0.05 1.5 (1.1-2.4) 647 (46.8) 735 (53.2) < 0.001 1.5 (1.3-1.7) Non-remote 74 (62.2) 45 (37.8) 694 (37.9) 1,139 (62.1)Current breast lump Yes 22 (75.9) 7 (24.1) 0.544 1.3 (0.6-3.1) 65 (56.5) 50 (43.5) < 0.05 1.8 (1.3-2.7) No 585 (70.7) 243 (29.3) 1,295 (41.5) 1,826 (58.5)HRT within 6 months No 598 (71.2) 242 (28.8) 0.101 2.2 (0.8-5.8) 1,245 (42.1) 1,712 (57.9) 0.723 1.0 (0.8-1.3) Yes 9 (52.9) 8 (47.1) 114 (41.1) 164 (59)Note:P values derived from Chi-squared test

attendance for Indigenous women were younger age (OR 12.3, 95%CI 8.1–18.8; p<0.001), being recalled to assessment during this screening round (OR 5.4, 95%CI 1.8–13; p<0.001), no family history of breast cancer (OR 2.1, 95%CI 1.3–3.3; p<0.02), and mainly speaking a language other than English (OR 1.9, 95%CI 1.3–2.9; p<0.02). Remoteness and having no past breast cancer diagnosis, which had univariate significance, were non-significant and removed from the model. For non-Indigenous women, predictors for low screening attendance were younger age (OR 11.5, 95%CI 9.6–13.7; p<0.001), being recalled to assessment during this screening round (OR 1.8, 95%CI 1.3–2.6), no family history of breast cancer (OR 1.8, 95%CI 1.5–2.1; p<0.001), and living remotely (OR 1.3, 95%CI 1.1–1.5; p<0.02). Presenting with a current

breast lump and no previous breast cancer diagnosis, which had univariate significance, were non-significant and removed from the model. These results are shown in Figure 2.

Discussion

It is widely reported that Indigenous Australian women have a lower survival rate and a younger age profile when diagnosed with breast cancer in comparison with their non-Indigenous counterparts.4 It is also well documented that Indigenous women have consistently lower attendance at screening for breast and cervical cancers than the rest of the population.1,17 However, in recent years, a 5% increase in national breast screening participation has been reported for Indigenous women aged 50–69 years,

whereas the overall rate for the entire target population has remained steady.2 Also, more of the screening rounds for Indigenous women were initial screens than for other women,25 a report in line with the findings of this study. The increased national Indigenous participation rate, although still about 19% lower compared with the general population, is a positive step towards better Indigenous population coverage. The improvement may be attributed to the BreastScreen National Accreditation Standards (NAS) which recommend BreastScreen services to implement strategies that increase access and participation for underserved populations. Some of the ways in which BreastScreen is trying to reduce systematic barriers to screening are, improved record-keeping of Indigenous data, targeted and culturally appropriate health promotion, growing the Indigenous health workforce, and more access points for consumers to screening sites.1,26,27 However, closing the gap on breast screening participation is a complex challenge that requires deep understanding of the logistical, cultural and health communication needs of Indigenous Australians.11


Our results show that younger Indigenous women were more likely to have attended BreastScreen for the first and second time in 2015 than older Indigenous women and non-Indigenous women. This suggests that young Indigenous women, arguably a critical target group given the age profile of breast cancer in Indigenous Australians,4 are demonstrating initial engagement with the screening program. What our data do not show, however, is whether these Indigenous women are likely to have continuous and regular attendance at screening beyond their second visit. While limited with the absence of women’s data over time, our findings could imply that although young Indigenous women are engaging with initial screening, significant attrition may be occurring after the second visit. With BreastScreen Australia’s aim of reducing breast cancer mortality through early detection, women’s ongoing participation in the program is critical.

Association between low attendance and living remotely are shown for both Indigenous and non-Indigenous women; however, it only remains significant in the multivariate model for Indigenous women. Geographic remoteness is widely reported to present barriers to screening in Australia1,17 and is of particular importance in the NT, with approximately two-thirds of the overall population living in remote to very remote locations.28 There have been national efforts to improve access to screening via the mobile screening van that travels to remote areas and indeed may account for the increase in participation of NT women in recent years.8 However, overall participation rates are still lower compared with other states and territories of Australia, and evidence of poorer overall health continues to be reported with increasing geographic remoteness.29-32

The characteristic that was significant for Indigenous women but not for others in our dataset was language, in that Indigenous women who mainly spoke a language other than English were likely to have fewer than three attendances compared with English-speaking Indigenous women and compared with other women. This result, coupled with the earlier finding that Indigenous compared with non-Indigenous women were more likely to attend BreastScreen for the first and second time in 2015, suggests that Indigenous women are engaging with initial screening in the NT, where culturally and linguistically appropriate strategies have been implemented to meet the needs of Indigenous peoples.8 However, our results

also imply that continued attendance beyond three visits may wane with those for whom English is not the main spoken language, a finding supported by other researchers who have shown that language can be a significant barrier to health for many Indigenous cultures around the world.10,13,33,34 For example, it is reported that there is no word for cancer in many Indigenous languages, including Australian languages, which immediately presents difficulties when promoting screening.35 While it is well established that women of diverse cultural backgrounds have historically lower uptake to screening in Australia,1,27 tracking whether these women remain in the program beyond their initial attendance as a true measure of appropriate engagement in NT should be the focus of further work.

Non-Indigenous women presenting with a current breast lump were associated with having fewer than three visits to screening in the unadjusted results. While BreastScreen Australia (BSA) targets asymptomatic women, there is a small group of women who present to screening with symptoms, particularly in the early screening rounds. The reported rate of symptomatic screening according to BreastScreen screening data from 1996 to 2005 is slightly higher for Indigenous women than non-Indigenous women.25 Our study, however, found the opposite – that the association with a current breast lump only had univariate significance for non-Indigenous women. A study in Finland which included self-reported or radiographer-

reported breast lumps in 1.3% of women screened in a population-based program found that the risk of breast cancer was sevenfold for women with lumps reported at screening compared to women with other symptoms (including nipple retractions and secretions).36 Given that the risk is high for women with breast lumps reported at screening, and with screening attendance for all women being lower than the national average in the NT, improving participation in BreastScreen NT is critically important.

A factor previously associated with low screening re-attendance in Australia and elsewhere is when a woman was previously recalled to assessment with a false-positive result.2,37,38 In the first round of screening, with high recall rates (up to 10.8% reported in 2012),29 and low positive predictive values (1.1% of women attending a first screen in 2015 had an invasive breast cancer or DCIS detected),2 low return attendance of recalled women in subsequent rounds is unsurprising. Maintaining high sensitivity and specificity in the BreastScreen program is one of the overarching goals of the NAS and therefore this must be evident, even with the first screening round.27 Our results show that women who had attended fewer screenings happened to have been recalled to assessment in this round; however, limitations in the data mean we can only estimate how a false positive would affect these women’s decisions about subsequent screening rounds. The current work reaffirms the importance of diagnostic efficacy to

Figure 2. Factors associated with less than 3 screening visits for Indigenous (n=857) and

non-Indigenous women (n=3236) in the NT.

Figure 2: Factors associated with less than 3 screening visits for Indigenous (n=857) and non-Indigenous women (n=3,236) in the NT.

Tapia et al. Article


long-term BreastScreen engagement and provides new evidence that this finding is not unique to any single grouping of women.

To improve access and program retention for diverse groups within the screening population, the NAS recommend Screening and Assessment Services (SAS) to provide equitable service to women who are culturally and linguistically diverse, are Indigenous, live in rural and remote areas and are from lower socioeconomic backgrounds.26 Although an evaluation of the program in 2014 found that BreastScreen Australia SAS uniformly performed well across the high-priority standards and performance indicators (for benign biopsy rates, cancer detection rates and interval cancer rates), researchers reported that SAS with high numbers of diverse participants failed to meet the standard for time between screening and assessment.27 That is, attendance at assessment within the recommended 28 days after being recalled was lower for services with high cultural diversity. The reasons are unclear and warrant further investigation; however, the concern is that longer times to assessment may affect health outcomes for these women. While strategies have been implemented at state levels to try to increase screening participation, attendance at post-screening assessment should be carefully considered at the SAS level, particularly for Indigenous women who are reportedly less likely to attend post-screening assessment within the recommended 28 days.25

There were a few limitations in this study. Longitudinal data would have allowed us to provide a broader scope of women’s attendance at BreastScreen in the NT. As data were only collected from an eight-month period in a program of biennial screening, the information about women’s screening behaviours over time could not be surmised. A further limitation of this study is the lack of historical clinical information on women, such as the result of prior screening rounds, as a previous false positive finding is shown to affect future screening attendance.

Conclusion

The current work corroborates previously reported variations between Indigenous and non-Indigenous women’s screening characteristics and provides evidence of factors strongly associated with low program attendance. Given the disparity in participation rates and known variations

in breast cancer incidence, mortality and survival for Indigenous and non-Indigenous women, strategies to optimise engagement with the screening program should be targeted to meet the logistical, cultural and health communication needs of Australian women.

References1. Australian Department of Health and Ageing.

Evaluation of the BreastScreen Australia Program – Evaluation Final Report – June 2009. Canberra (AUST): Government of Australia; 2009.

2. Australian Institute of Health and Welfare. BreastScreen Australia Monitoring Report 2014–2015. Canberra (AUST): AIHW; 2017.

3. Australian institute of Health and Welfare. BreastScreen Australia Monitoring Report 1998–1999 and 1999–2000. Canberra (AUST): AIHW; 2003.

4. Tapia KA, Garvey G, Mc Entee M, Rickard M, Brennan P. Breast cancer in Australian Indigenous women: Incidence, mortality, and risk factors. Asian Pac J Cancer Prev. 2017;18(4):873-84.

5. World Health Organization. WHO Guidelines Approved by the Guidelines Review Committee. WHO Position Paper on Mammography Screening. Geneva (CHE): WHO; 2014.

6. Australian Bureau of Statistics. 1270.0.55.005 - Australian Statistical Geography Standard (ASGS): Volume 5 - Remoteness Structure, July 2011. Canberra (AUST): ABS; 2011.

7. Australian Institute of Health and Welfare. Rural, Regional and Remote Health: A Guide to Remoteness Classifications. Canberra (AUST): AIHW; 2004.

8. Zhang X, Condon J, Douglas F, Bates D, Guthridge S, Garling L, et al. Women’s Cancers and Cancer Screening in the Northern Territory. Darwin (AUST): Northern Territory Department of Health; 2012.

9. Marmot MG, Altman DG, Cameron DA, Dewar JA, Thompson SG, Wilcox M. The benefits and harms of breast cancer screening: an independent review. Br J Cancer. 2013;108(11):2205-40.

10. Kolahdooz F, Jang SL, Corriveau A, Gotay C, Johnston N, Sharma S. Knowledge, attitudes, and behaviours towards cancer screening in indigenous populations: A systematic review. Lancet Oncol. 2014;15(11):e504-16.

11. Pilkington L, Haigh MM, Durey A, Katzenellenbogen JM, Thompson SC. Perspectives of Aboriginal women on participation in mammographic screening: A step towards improving services. BMC Public Health. 2017;17(1):697.

12. Roder D, Webster F, Zorbas H, Sinclair S. Breast screening and breast cancer survival in Aboriginal and Torres Strait Islander women of Australia. Asian Pac J Cancer Prev. 2012;13(1):147-55.

13. Martins T, Hamilton W, Ukoumunne OC. Ethnic inequalities in time to diagnosis of cancer: A systematic review. BMC Fam Pract. 2013;14(1):197.

14. Moore SP, Antoni S, Colquhoun A, Healy B, Ellison-Loschmann L, Potter JD, et al. Cancer incidence in indigenous people in Australia, New Zealand, Canada, and the USA: A comparative population-based study. Lancet Oncol. 2015;16(15):1483-92.

15. Seneviratne S, Campbell I, Scott N, Shirley R, Lawrenson R. Impact of mammographic screening on ethnic and socioeconomic inequities in breast cancer stage at diagnosis and survival in New Zealand: A cohort study. BMC Public Health. 2015;15:46.

16. Smith-Bindman R, Miglioretti DL, Lurie N, Abraham L, Barbash RB, Strzelczyk J, et al. Does utilization of screening mammography explain racial and ethnic differences in breast cancer? Ann Intern Med. 2006;144(8):541-53.

17. Australian Institute of Health and Welfare and Cancer Australia. Cancer in Aboriginal and Torres Strait Islander Peoples of Australia: An Overview. Canberra (AUST): AIHW; 2013. p. 165.

18. Gibberd A, Supramaniam R, Dillon A, Armstrong BK, O’Connell DL. Are Aboriginal people more likely to be diagnosed with more advanced cancer? Med J Aust. 2015;202(4):195-9.

19. Roder D, Webster F, Zorbas H, Sinclair S. Breast screening and breast cancer survival in Aboriginal and Torres Strait Islander women of Australia. Asian Pac J Cancer Prev. 2012;13(1):147-55.

20. Dasgupta P, Baade PD, Youlden DR, Garvey G, Aitken JF, Wallington I, et al. Variations in outcomes for Indigenous women with breast cancer in Australia: A systematic review. Eur J Cancer Care (Engl). 2017;26(6):e12662.

21. Australian Health Ministers’ Advisory Council. Aboriginal and Torres Strait Islander Health Performance Framework 2014 Report. Canberra (AUST): AHMAC; 2015.

22. Garvey G, Cunningham J, Valery PC, Condon J, Roder D, Bailie R, et al. Reducing the burden of cancer for Aboriginal and Torres Strait Islander Australians: Time for a coordinated, collaborative, priority-driven, Indigenous-led research program. Med J Aust. 2011;194(10):530-1.

23. Karvelas P. Sect. Indigenous Health News: Programs Close the Gap on Indigenous Breast Cancer Screening. The Australian. 2014 October 14.

24. Tapia KA, Garvey G, McEntee MF, Rickard M, Lydiard L, Brennan PC. Mammographic densities of Aboriginal and non-Aboriginal women living in Australia’s Northern Territory. Int J Public Health. 2019;64. doi.org/10.1007/s00038-019-01237-w

25. Cancer Australia. Study of Breast Cancer Screening Characteristics and Breast Cancer Survival in Aboriginal and Torres Strait Islander Women of Australia. Surry Hills (AUST): Cancer Australia; 2012. p. 83.

26. BreastScreen Australia. Breastscreen Australia National Accreditation Standards, October 2015. Canberra (AUST): Cancer Australia; 2016.

27. Roder DM, Ward GH, Farshid G, Gill PG. Influence of service characteristics on high priority performance indicators and standards in the BreastScreen Australia program. Asian Pac J Cancer Prev. 2014;15(14):5901-8.

28. Australian Bureau of Statistics. 2075.0 - Census of Population and Housing - Counts of Aboriginal and Torres Strait Islander Australians, 2011. Canberra (AUST): ABS; 2011.

29. Australian Institute of Health and Welfare. BreastScreen Australia Monitoring Report 2011–2012. Canberra (AUST): AIHW; 2014. p. 99.

30. Australian Institute of Health and Welfare. BreastScreen Australia Monitoring Report 2013–2014. Canberra (AUST): AIHW; 2016.

31. Australian Institute of Health and Welfare. Rural, Regional and Remote Health: Indicators of Health System Performance. Canberra (AUST): AIHW; 2008.

32. Roder D, Zorbas HM, Kollias J, Pyke CM, Walters D, Campbell ID, et al. Analysing risk factors for poorer breast cancer outcomes in residents of lower socioeconomic areas of Australia. Aust Health Rev. 2014;38(2):134-41.

33. Shahid S, Bessarab D, Howat P, Thompson SC. Exploration of the beliefs and experiences of Aboriginal people with cancer in Western Australia: A methodology to acknowledge cultural difference and build understanding. BMC Med Res Methodol. 2009;9:60.

34. Flood D, Rohloff P. Indigenous languages and global health. Lancet Glob Health. 2018;6(2):e134-e5.

35. Shahid S, Thompson SC. An overview of cancer and beliefs about the disease in Indigenous people of Australia, Canada, New Zealand and the US. Aust N Z J Public Health. 2009;33(2):109-18.

36. Singh D, Malila N, Pokhrel A, Anttila A. Association of symptoms and breast cancer in population-based mammography screening in Finland. Int J Cancer. 2015;136(6):E630-7.

37. Sim MJ, Siva SP, Ramli IS, Fritschi L, Tresham J, Wylie EJ. Effect of false-positive screening mammograms on rescreening in Western Australia. Med J Aust. 2012;196(11):693-5.

38. Klompenhouwer EG, Duijm LE, Voogd AC, den Heeten GJ, Strobbe LJ, Louwman MW, et al. Re-attendance at biennial screening mammography following a repeated false positive recall. Breast Cancer Res Treat. 2014;145(2):429-37.



The long-standing health and mortality disadvantage of Aboriginal and Torres Strait Islander Australians is

well known1-3 and is generally worse in the Northern Territory than in other states and territories. As part of efforts to monitor progress with mortality and life expectancy and help inform appropriate policy and service responses, it is useful to construct life tables on a regular basis to obtain life expectancy and other mortality statistics. Life expectancy at birth is a valuable summary indicator – not only of mortality conditions, but also of the general health and wellbeing of a population – and it is one of the key ‘Closing the Gap’ indicators.4 When the Closing the Gap targets were announced by the Council of Australian Governments in 2008, the aim was to close the life expectancy gap within a generation (by 2031) and halve the under-five mortality rate within 10 years (by 2018).5

Unfortunately, creating life tables for the Aboriginal and Torres Strait Islander population is not straightforward. The disadvantage of Aboriginal and Torres Strait Islander Australians with regards to health and mortality is compounded by a statistical disadvantage in which the coverage, consistency and quality of population data is far from ideal. The Australian Bureau of Statistics (ABS) advises that most births and deaths of Aboriginal and Torres Strait Islander Australians are likely to be recorded, but not all are specifically identified as Indigenous.3 Aboriginal and Torres Strait Islander

Estimated Resident Populations (ERPs), which form the denominators of a wide range of demographic and socioeconomic measures, are also likely to have problems. They originate from census counts that miss about one-in-six Aboriginal and Torres Strait Islander people and, although ERPs are corrected for census net undercount, they still contain some limitations including age heaping, a limited available time series, and inconsistency over time.

Creating robust life tables is therefore challenging in this difficult data environment and adjustments must be made to data to rectify the deficiencies. In calculating

2010–12 Aboriginal and Torres Strait Islander life tables for Australia, the ABS applied adjustment factors to account for the under-identification of Aboriginal and Torres Strait Islander deaths.3 The adjustment factors were calculated from the Census Data Enhancement Indigenous Mortality Project, a study undertaken by the ABS that linked 2011 Census records with deaths recorded in the 12 months following the Census.6 Unfortunately, adjustment factors could not be applied to individual states and territories due to small numbers by age group.

The Northern Territory is the one jurisdiction where the coverage and quality of Aboriginal

Limited progress in closing the mortality gap for Aboriginal and Torres Strait Islander Australians of the Northern TerritoryTom Wilson,1 Yuejen Zhao,2 John Condon3

Abstract

Objectives: To assess whether progress is being made towards reducing Aboriginal and Torres Strait Islander inequality in life expectancy and under-five mortality in the Northern Territory.

Methods: Life tables for five-year periods from 1966–71 to 2011–16 were calculated using standard abridged life table methods with Aboriginal and Torres Strait Islander deaths and population estimates as inputs. The latter were calculated using reverse cohort survival.

Results: In 2011–16, life expectancy at birth for the Aboriginal and Torres Strait Islander population was 68.2 years for females and 64.9 years for males. Limited progress in under-five mortality rates has been made in recent years.

Conclusions: Although Aboriginal and Torres Strait Islander life expectancy has increased in the long run, the gap with all-Australian life expectancy has not narrowed. The gap in under-five mortality rates is much lower than it was in the 1960s and 1970s, but progress has been limited over the past decade.

Implications for public health: The ‘Closing the Gap’ target of halving the gap in under-five mortality by 2018 will not be met in the Northern Territory, and there is no evidence yet of progress on the target to eliminate the gap in life expectancy by 2031.

Key words: Aboriginal and Torres Strait Islander, life expectancy, Northern Territory, mortality, Closing the Gap

1. Northern Institute, Charles Darwin University, Northern Territory2. Northern Territory Department of Health3. Menzies School of Health Research, Charles Darwin University, Northern TerritoryCorrespondence to: Dr Tom Wilson, Northern Institute, Charles Darwin University, Ellengowan Drive, Darwin, NT 0909; e-mail: [email protected]. Submitted: October 2018; Revision requested: April 2019; Accepted: May 2019The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,



INDIGENOUS HEALTH


and Torres Strait Islander demographic data is good, and the application of under-identification adjustment factors is unnecessary. Available evidence suggests there is very good coverage of Aboriginal and Torres Strait Islander deaths.3,7 The ABS Indigenous Mortality Project discovered the Northern Territory had the highest rate of consistent reporting between census and death registrations, compared to the other states and territories; 95% of linked records had consistent Indigenous identification across both sources (compared to 62%, nationally). The number of deaths recorded as Indigenous in death registrations was actually 2% higher than the number of deaths identified as Indigenous from the linked census records. Northern Territory death registration forms have recorded Indigenous status since September 1988 and Aboriginal and Torres Strait Islander deaths data are available from the ABS from this time. Death counts for earlier years back to 1967 were estimated by Condon et al.7 by inferring Indigenous status from other information supplied on death registration forms. A blind test of this method for a period where Indigenous status was officially recorded in death records found that 95% of recorded Indigenous deaths were correctly inferred as such.

The time series of deaths for the Northern Territory Aboriginal and Torres Strait Islander population now stretches half a century. The aim of this paper is to summarise life expectancy and under-five mortality trends for the Northern Territory Aboriginal and Torres Strait Islander population over this period, in part updating earlier work.2,8,9 Our statistics differ from those of the ABS and other researchers by covering a longer period – from mid-1966 to mid-2016 – and by the creation of life tables based on a set of population estimates that are consistent over the whole study period. We also examine life expectancy across the age spectrum and not just at birth.

Data and Methods

Life tables were calculated using the standard abridged life table method,10 which requires age-sex-specific deaths and populations as input data. We used the usual ages of 0, 1–4, and then five-year age groups up to the highest open-ended age group of 85-plus. To reduce random variation, we calculated life tables for five-year periods from 1966–71

to 2011–16 (from 1 July in one year to 30 June five years later). The number of deaths was obtained directly from the ABS and the Condon et al. dataset. ABS Aboriginal and Torres Strait Islander ERPs were used as population estimates for 2016, but those for earlier years were calculated specifically for this study. The reason for doing so is because Aboriginal and Torres Strait Islander ERPs are inconsistent from one census year to another and are not available prior to 1986. For comparative purposes, abridged life tables were calculated for the Australian population as a whole for the same five-year time periods. We also decomposed Aboriginal and Torres Strait Islander life expectancy improvements between 1966–71 and 2011–16 to determine which age-specific mortality rates contributed the most to life expectancy changes. The decomposition method of Arriaga was applied.11

Counts of deaths to Aboriginal and Torres Strait Islander residents of the Northern Territory by sex and ages 0, 1–4, and then five-year age groups up to 85+ for individual years 1967 to 2001 were obtained from the Condon et al. dataset. Deaths for the second half of 1966 were assumed to be the same as those for the first half of 1967. Aboriginal and Torres Strait Islander deaths for five-year periods from 1 July to 30 June for 2001–06, 2006–11 and 2011–16 were obtained from the ABS for the same age and sex groups. Deaths from the ABS were also obtained in five-year period-cohort form (e.g. the cohort aged 25–29 in 2011 and aged 30–34 in 2016) for population estimation calculations. Aboriginal and Torres Strait Islander ERPs for 30 June 2016 by sex and five-year age group were obtained from the ABS.12 Census net interstate migration counts for five-year

periods from 1966–71 to 2011–16 by sex and five-year period-cohort were also obtained from the ABS.

Aboriginal and Torres Strait Islander population estimates were calculated using reverse survival from 30 June 2016 ERPs going back in five-year increments to 30 June 1966. Zero net overseas migration and zero net change in Indigenous identification were assumed. Figure 1 illustrates the principle of the calculations for one cohort over one five-year period. In this example, the population aged 40–44 in 2011 was calculated as the population of the same cohort in 2016 (then aged 45–49) plus the deaths that occurred to the cohort over the 2011–16 period minus cohort net migration. The shaded parallelogram represents the cohort deaths and net migration while the thick vertical lines are the cohort populations in 2011 and 2016. This calculation was applied to all cohorts by sex for all periods from 2011–16 back to 1966–71 – with one exception. An adjustment was made to the ‘starting’ 2016 populations aged 0–4. The number of Aboriginal and Torres Strait Islander 0–4-year-olds obtained from reverse survival for earlier years was found to be higher than Aboriginal and Torres Strait Islander ERPs published for those years, indicating a systematic under-estimation of the population in the 0–4 age group. We therefore increased the 0–4-year-old ERP in 2016 by a factor of 1.02 for females and 1.05 for males; the average ratios of reverse survived population estimates to ERPs for this age group over the years 1996 to 2011. The effect of this adjustment on life expectancy at birth was less than 0.1 years.

The resulting population estimates dataset spans 50 years and is fully consistent with the 2016 ERPs for ages 5+ and the demographic components of change over that period. However, it should be noted that the population estimates and life tables used for this study are not perfect. The 2016 ERPs that provide the starting point for the earlier population estimates are not precise figures but are estimates based on census counts adjusted for net under-enumeration, people temporarily away from their usual residence on census night, plus the small timing difference between the census in August and the 30 June reference date of the ERP. There is almost certainly some degree of error in these data. Population estimates for earlier years are based on the assumption that all components of demographic change in the calculations are accurate, including the

Figure 1: Illustration of reverse survival Note: P denotes population, D deaths and N net migration

2011

2016

Figure 1: Illustration of reverse survival.

Note: P denotes population, D deaths and N net migration

Indigenous Health Closing the Gap in Indigenous mortality


assumption of zero net overseas migration and zero net identification change. This last assumption is the most uncertain for recent years. Available evidence from the Australian Census Longitudinal Dataset (ACLD) suggests net identification change between 2011 and 2016 was low for the Northern Territory13 and much lower than in any other jurisdiction in Australia. Due to the small sample size and census errors, we are not fully convinced of the existence of this phenomenon in the Northern Territory, but careful attention will need to be paid to future updates to the ACLD when 2021 Census data are added. ACLD data suggests no net identification change occurred in the Northern Territory between 2006 and 2011.

It is also assumed that all data inputs for the population estimates and life tables refer to the Aboriginal and Torres Strait Islander population defined in exactly the same way, although it is known that indigeneity is identified or recorded in slightly different ways in different data collections.14 Our assumptions almost certainly do not hold precisely.

Results

Life expectancy at birthFigure 2 presents life expectancy at birth estimates for the Northern Territory Aboriginal and Torres Strait Islander population from 1966–71 to 2011–16. Equivalent statistics for the Australian population as a whole are also shown (by the dashed lines). In 2011–16, life expectancy stood at 68.2 years for Aboriginal and Torres Strait Islander females in the Northern Territory and 64.9 years for males. For Australia as a whole, life expectancy in 2011–16 reached 84.9 years for females and 80.8 years for males. Over the entire study period of 1966–71 to 2011–16, Aboriginal and Torres Strait Islander life expectancy at birth in the Northern Territory increased by an average of 3.1 years per decade for females and 2.7 years per decade for males. Most recently, between 2006–11 and 2011–16, male life expectancy increased by an impressive 1.9 years, although there was no significant change for females.

The new Northern Territory Aboriginal and Torres Strait Islander life expectancy figures presented here are very close to estimates published previously for 1966–71 to 2006–11 using the same methods.8 Minor differences are due to slight changes in the population

estimates that form the denominators of age-specific death rates. The new Aboriginal and Torres Strait Islander life expectancy figures are also similar to equivalent statistics published by the ABS. ABS Aboriginal and Torres Strait Islander life expectancy at birth estimates for the Northern Territory for 2010–12 were 68.7 years for females and 63.4 for males.3 Our estimates, interpolated to the same reference dates, are close to these at 68.3 and 64.0 years, respectively.

A decomposition of increases in Aboriginal and Torres Strait Islander life expectancy at birth between 1966–71 and 2011–16 was undertaken to reveal which age-specific mortality rates contributed most to the increases (Table 1). For females, 45% of the increase was due to improvements in under-five mortality, while for males it was about 50%. Other contributions to life expectancy increases were distributed across other age groups, although with relatively smaller

Figure 2: Northern Territory Aboriginal and Torres Strait Islander and all-Australian life expectancy at birth, 1966-71 to 2011-16 Source: authors’ calculations Note: 95% confidence intervals shown for Aboriginal and Torres Strait Islander life expectancy

Figure 2: Northern Territory Aboriginal and Torres Strait Islander and all-Australian life expectancy at birth, 1966-71 to 2011-16.

Source: authors’ calculationsNote: 95% confidence intervals shown for Aboriginal and Torres Strait Islander life expectancy

Figure 3: Northern Territory Aboriginal and Torres Strait Islander life expectancy at selected ages by sex, 1966-71 to 2011-16 Source: authors’ calculations

Note: 95% confidence intervals shown

Figure 3: Northern Territory Aboriginal and Torres Strait Islander life expectancy at selected ages by sex, 1966-71 to 2011-16.

Source: authors’ calculationsNote: 95% confidence intervals shown

Wilson, Zhao and Condon Article



contributions from the childhood ages of 5–14 and younger adult age groups. In contrast, for the Australian population as a whole, most life expectancy gains over this period have been contributed by mortality declines in the older adult age groups.

Aboriginal and Torres Strait Islander life expectancy at ages other than birth have also increased over the long run. Figure 3 illustrates trends in life expectancy at birth (age 0) and at ages 25, 50 and 75 over the study period. For example, female life expectancy at age 50 has increased from 18.1 years in 1966–71 to 24.4 years by 2011–16 (+6.3 years), while for males of the same age it has increased from 17.8 years to 22.1 (+4.3 years). For both genders, most of the increase in the adult ages has occurred since the early 1990s, reflecting sustained declines in mortality rates across most age groups above age 50 since then.

These increases in Aboriginal and Torres Strait Islander life expectancy are undoubtedly positive developments but, if the Closing the Gap goal is to be met, life expectancy not only needs to be increasing, but increasing at a much faster rate than all-Australian life expectancy. Figure 4 illustrates the gap in life expectancy at birth and selected other ages. In 2011–16, the difference in life expectancy at birth between the Northern Territory Aboriginal and Torres Strait Islander population and the Australian population as a whole stood at 16.7 years for females and 15.9 years for males. Aside from a narrowing of the gap for Aboriginal and Torres Strait Islander females between 1966–71 and 1971–76, both male and female Aboriginal and Torres Strait Islander life expectancy at birth has remained

between 15 and 19 years below all-Australian life expectancy.

As the graphs in Figure 4 show, the life expectancy gaps vary by age. For females, the life expectancy gap at age 25 has fluctuated over time but was not appreciably different in 2011–16 and 1966–71. At ages 50 and 75, the gap has increased slightly. In contrast, male life expectancy gaps for ages 25, 50 and 75 were larger in 2011–16 than in 1966–71. In part, this is due to substantial gains in all-Australian male life expectancy over the past 50 years.

Childhood mortalityThe other mortality-related target of Closing the Gap was to halve the gap in under-five mortality within 10 years. Figure 5 shows the mortality rates for 0–4-year-old Aboriginal and Torres Strait Islander children in the Northern Territory from 1966–71 to 2011–16 alongside equivalent rates for Australia as a whole (shown by the dashed lines). The steep decline in childhood mortality rates in the 1960s, 1970s and 1980s represents great success in combating perinatal, respiratory and infectious causes of death.15 However, progress in lowering under-five mortality appears to have stalled between 2006–11 and 2011–16, with mortality rates for females changing from 0.00290 (95%CI 0.00248-0.00332) to 0.00328 (0.00285-0.00371), and from 0.00322 (0.00277-0.00366) to 0.00317 (0.00275-0.00360) for males. In 2011–16, these rates were still approximately four times those of the Australian population as a whole, which were 0.00076 for females and 0.00087 for males. Over this most recent period, the gap in the under-five mortality rate widened,

increasing from 0.00198 to 0.00253 for females and from 0.00204 to 0.00230 for males.

Discussion and conclusions

The updated life expectancy and under-five mortality rates for the Northern Territory Aboriginal and Torres Strait Islander population presented here mostly show limited amounts of improvements in recent years, especially for females. Although Aboriginal and Torres Strait Islander life expectancy has increased in the long run, the gap has failed to narrow. In fact, for males, the gap across the adult ages has increased over time. The original target of completely closing the life expectancy at birth gap within a generation (by 2031) was ambitious at the time of its announcement, and probably unachievable. It not only requires improvement in absolute terms but improvement at a faster rate than all-Australian life expectancy; this would need to happen at a very rapid pace to completely close the gap by 2031.

Among the disappointing statistics overall, there is some good news. Male life expectancy at birth in the Northern Territory increased by 1.9 years between 2006–11 and 2011–16. In addition, the Aboriginal and Torres Strait Islander infant mortality rate has fallen dramatically over the past half century – by 86% from 1966–71 to 2011–16. The probability of a newly born Aboriginal and Torres Strait Islander child in the Northern Territory dying before their first birthday is now 1.2% for females and 1.3% for males. Nonetheless, these statistics are still above

Table 1: Contribution to life expectancy at birth increases between 1966-71 and 2011-16 from mortality changes by age group, Northern Territory Aboriginal and Torres Strait Islander population and Australian population.

Ages

Northern Territory Aboriginal and Torres Strait Islander population

Australian population

Females Males Females MalesΔe0 (%) 95% CI Δe0 (%) 95% CI Δe0 (%) 95% CI Δe0 (%) 95% CI

0–4 6.35 (45.2%) 6.16-6.54 6.04 (50.1%) 5.87-6.22 1.27 (12.2%) 1.26-1.28 1.58 (12.2%) 1.58-1.595–14 0.09 (0.7%) 0.02-0.17 0.16 (1.3%) 0.08-0.23 0.16 (1.6%) 0.16-0.17 0.25 (1.9%) 0.24-0.2515–24 0.63 (4.5%) 0.51-0.76 0.45 (3.7%) 0.33-0.57 0.22 (2.1%) 0.22-0.23 0.65 (5.0%) 0.64-0.6625–34 0.66 (4.7%) 0.54-0.77 0.92 (7.6%) 0.8-1.04 0.23 (2.2%) 0.22-0.23 0.35 (2.7%) 0.35-0.3635–44 1.45 (10.3%) 1.33-1.56 1.23 (10.2%) 1.12-1.34 0.48 (4.6%) 0.47-0.49 0.64 (4.9%) 0.64-0.6545–54 1.27 (9.0%) 1.2-1.33 1.18 (9.8%) 1.13-1.23 0.98 (9.5%) 0.97-0.99 1.51 (11.7%) 1.51-1.5255–64 2.12 (15.1%) 2.08-2.16 0.75 (6.2%) 0.7-0.79 1.62 (15.7%) 1.62-1.63 2.85 (21.9%) 2.84-2.8565–74 1.05 (7.5%) 0.88-1.22 1.09 (9.0%) 0.89-1.29 2.44 (23.6%) 2.44-2.44 3.17 (24.4%) 3.17-3.1775+ 0.44 (3.1%) 0.08-0.79 0.25 (2.1%) -0.1-0.61 2.94 (28.5%) 2.93-2.96 1.97 (15.2%) 1.96-1.98Total 14.06 (100%) 13.57-14.56 12.07 (100%) 11.57-12.57 10.35 (100%) 10.33-10.37 12.99 (100%) 12.96-13.01Source: authors’ calculationsNote: Δe0 denotes change in life expectancy at birth


the equivalent all-Australian probabilities (0.3% for females and 0.4% for males), leaving room for improvement.

How can life expectancy at birth be increased? If Aboriginal and Torres Strait Islander childhood mortality (across ages 0–14 years) were to fall to the same rates experienced by the Australian population as a whole in 2011–16, the gain in life expectancy at birth would be about one year for both males and females. So, future life expectancy gains must come from the adult age groups. Halving the difference in 2011–16 age-specific mortality rates between the Northern Territory Aboriginal and Torres Strait Islander population and the Australian population as a whole at ages 15 and above would add about six years to both male and female Aboriginal and Torres Strait Islander life expectancy at birth. That is the scale of improvement that would clearly indicate a move towards closing the gap.

Our findings suggest that current approaches to tackling Aboriginal and Torres Strait Islander mortality inequalities are struggling to make headway. Previous research shows a combination of six selected risk factors explain 60–70% of the Aboriginal and Torres Strait Islander life expectancy gap in the Northern Territory: socioeconomic disadvantage, smoking, alcohol abuse, obesity, pollution and intimate partner violence.16 Socioeconomic disadvantage (such as relatively poor education, housing, hygiene, nutrition and income) is the largest contributor, responsible for between one-third and one-half of the gap in life expectancy at birth.

The other Closing the Gap targets focused on education and employment, which are fundamentally necessary (albeit not sufficient on their own) to an individual’s capacity to control their own destiny and contribute to the wellbeing of their family, community and environment. Substantial progress in housing, education, hygiene, nutrition, employment and income is required to improve the mortality outcomes of the Northern Territory’s Aboriginal and Torres Strait Islander population as measured by life expectancy at birth and under-five mortality. The inequality in mortality between Aboriginal and Torres Strait Islander and other Australians is the cumulative effect of disadvantage in almost every stage and aspect of life. Closing the

mortality gap depends on, and can only follow, substantial and sustained progress in the social, economic and environmental circumstances of Aboriginal and Torres Strait Islander Australians.

Acknowledgement

This work was funded by a Charles Darwin University Start-Up grant.

Ethical standardsThe study was ruled exempt from ethics review by the Human Research Ethics Committee of Charles Darwin University.

(a) Females (b) Males

Figure 4: The Northern Territory Aboriginal and Torres Strait Islander life expectancy gap at selected ages by sex, 1966-71 to 2011-16 Note: Gap calculated as Northern Territory Aboriginal and Torres Strait Islander life expectancy minus that for Australia as a whole. 95% confidence intervals shown.

Figure 4: The Northern Territory Aboriginal and Torres Strait Islander life expectancy gap at selected ages by sex, 1966-71 to 2011-16.

Note: Gap calculated as Northern Territory Aboriginal and Torres Strait Islander life expectancy minus that for Australia as a whole. 95% confidence intervals shown.

Figure 5: The mortality rates of 0–4 year-old Aboriginal and Torres Strait Islander children in the Northern Territory, 1966-71 to 2011-16.

Figure 5: The mortality rates of 0-4 year old Aboriginal and Torres Strait Islander children in the Northern Territory, 1966-71 to 2011-16 Source: authors’ calculations Note: 95% confidence intervals shown for Aboriginal and Torres Strait Islander mortality rates

Source: authors’ calculationsNote: 95% confidence intervals shown for Aboriginal and Torres Strait Islander mortality rates

Wilson, Zhao and Condon Article


References1. Australian Institute of Health and Welfare. The Health

and Welfare of Australia’s Aboriginal and Torres Strait Islander Peoples. Canberra (AUST): AIHW; 2015.

2. Wilson T, Condon J, Barnes T. Improvements in Northern Territory Indigenous life expectancy, 1967-2004. Aust N Z J Public Health. 2007;31(2):184-8.

3. Australian Bureau of Statistics. 3302.0.55.003 - Life Tables for Aboriginal and Torres Strait Islander Australians, 2010-2012. Canberra (AUST): ABS; 2013.

4. Australian Institute of Health and Welfare. Aboriginal and Torres Strait Islander Health Performance Framework 2017 Report. Canberra (AUST): AIHW; 2017.

5. Council of Australian Governments. National Indigenous Reform Agreement (Closing the Gap). Canberra (AUST): COAG; 2009.

6. Australian Bureau of Statistics. 3302.0.55.005 - Death Registrations to Census Linkage Project – Key Findings for Aboriginal and Torres Strait Islander Peoples, 2011-2012. Canberra (AUST): ABS; 2013.

7. Condon JR, Barnes T, Cunningham J, Smith L. Demographic Characteristics and Trends of the Northern Territory Indigenous Population, 1966 to 2001. Darwin (AUST): Cooperative Research Centre for Aboriginal Health; 2004.

8. Wilson T. New population and life expectancy estimates for the Indigenous population of Australia’s Northern Territory, 1966-2011. Plos One. 2014;9(5):e97576.

9. Georges N, Guthridge SL, Li SQ, Condon JR, Barnes T, Zhao Y. Progress in closing the gap in life expectancy at birth for Aboriginal people in the Northern Territory, 1967-2012. Med J Aust. 2017;207(1):25-30.

10. Preston SH, Heuveline P, Guillot M. Demography: Measuring and Modelling Population Processes. Oxford (UK): Blackwell; 2001.

11. Arriaga EE. Measuring and explaining the change in life expectancies. Demography. 1984;21:83-96.

12. Australian Bureau of Statistics. 3238.0.55.001 - Estimates of Aboriginal and Torres Strait Islander Australians, June 2016. Canberra (AUST): ABS; 2018.

13. Biddle N, Markham F. Indigenous Identification Change Between 2011 and 2016: Evidence from the Australian Census Longitudinal Dataset. CAEPR Topical Issue No.: 1. Canberra (AUST): Australian National University Centre for Aboriginal Economic Policy Research; 2018.

14. Australian Bureau of Statistics. 4726.0 - Perspectives on Aboriginal and Torres Strait Islander Identification in Selected Data Collection Contexts. Canberra (AUST): ABS; 2013.

15. Tay EL, Li SQ, Guthridge S. Mortality in the Northern Territory, 1967-2006. Darwin (AUST): Northern Territory Department of Health; 2013.

16. Zhao Y, Wright J, Begg S, Guthridge S. Decomposing Indigenous life expectancy gap by risk factors: A life table analysis. Popul Health Metr. 2013;11(1):1-9.



Poor dietary intakes and obesity are leading risk factors for preventable non-communicable diseases such as

diabetes, heart disease and some cancers.1 In Australia, two-thirds of adults and one-quarter of all children were overweight or obese in 2014-15.2 The food environment is a key driver of these public health issues due to the ubiquitous availability and marketing of cheap energy-dense, nutrient-poor foods and beverages that contain excessive amounts of sugar, salt and saturated fats.3 Non-alcoholic beverages, including sugar-sweetened beverages (SSBs), are the largest contributors to added sugars in the daily diets of Australians (37%)4,5 and have thus been identified as a key policy target to improve population diets.6

Price promotions (also referred to as ‘temporary price discounts’ or ‘specials’) are widely used by retailers and food manufacturers to influence consumer purchasing patterns. Price promotions result in a short-term sales uplift of a particular product by enticing consumers to purchase in greater quantities and/or temporarily switch brands or shopping habits.7 Accordingly, the UK government and public health groups in Australia have recently called for regulations restricting price promotions on unhealthy foods and beverages as part of a broader regulatory strategy to address childhood obesity.8-10 Beverage price promotions are of particular interest given the potential of price promotions to undermine SSB taxes, which have now been introduced in more

than 30 jurisdictions.6 SSB taxes aim to reduce demand for SSBs via an increase in their prices. In contrast, price promotions aim to increase demand via a temporary reduction in prices and may thereby attenuate the effects of a SSB tax. Similar policies to restrict the influence of price promotions on alcohol have previously been recommended in Australia,11 with legislative bans on multi-buys implemented in Scotland in 2011.12

The limited evidence examining the extent of beverage price promotions to date suggests that SSBs are more commonly price promoted compared to non-sugary beverages. A cross-sectional in-store audit of price promotions across a nation-wide sample of food stores (including 955 supermarkets) in the United States during 2010-12 revealed that there was a higher prevalence of price promotions among SSBs

The frequency and magnitude of price-promoted beverages available for sale in Australian supermarketsChristina Zorbas,1 Beth Gilham,1 Tara Boelsen-Robinson,1,2 Miranda R.C. Blake,1,2 Anna Peeters,1 Adrian J. Cameron,1 Jason H.Y. Wu,3 Kathryn Backholer1

1. Global Obesity Centre, Institute for Health Transformation, School of Health and Social Development, Deakin University, Geelong, Victoria2. School of Public Health and Preventive Medicine, Monash University, Victoria 3. The George Institute for Global Health, Faculty of Medicine, University of New South Wales, Sydney, New South WalesCorrespondence to: Ms Christina Zorbas, Deakin University, Locked Bag 20000, Geelong, VIC 3220; e-mail: [email protected]: October 2018; Revision requested: February 2019; Accepted: March 2019The authors have stated the following conflict of interest: AJC is an academic partner on a healthy supermarket intervention trial that includes Australian local government

and supermarket retail (IGA) collaborators. No funding was received from IGA for this trial, which was funded by the Australian National Health and Medical Research Council, VicHealth and Deakin University.



Abstract

Objective: Price promotions are used to influence purchases and represent an important target for obesity prevention policy. However, no long-term contemporary data on the extent and frequency of supermarket price promotions exists. We aimed to evaluate the frequency, magnitude and weekly variation of beverage price promotions available online at two major Australian supermarket chains over 50 weeks.

Methods: Beverages were categorised into four policy-relevant categories (sugar-sweetened beverages, artificially-sweetened beverages, flavoured milk and 100% juice, milk and water). The proportional contribution of each category to the total number of price proportions, the proportion of price promotions within the available product category, the mean discount, and weekly variation in price promotions were calculated.

Results: For Coles and Woolworths respectively, 26% and 30% of all beverages were price promoted in any given week. Sugar-sweetened beverages made up the greatest proportion of all price promotions (Coles: 46%, Woolworths: 49%). Within each product category, the proportion of sugar-sweetened and artificially-sweetened beverages that were price promoted was similar, higher than the other categories and reasonably constant over time. Diet drinks and sugar-sweetened soft drinks were most heavily discounted (by 29-40%).

Conclusions: Beverage price promotions are used extensively in Australian supermarkets, undermining efforts to promote healthy population diets.

Implications for public health: Policies restricting price promotions on sugar-sweetened beverages are likely to be an important part of strategies to reduce obesity and improve population nutrition.

Key words: Sugar-sweetened beverages, food policy, price promotions, obesity

FOOD AND BEVERAGE


(18.2%) compared to non-sugary beverages (12.1%).13 Similarly, a four-week examination of beverage price promotions in New Zealand during 2007 highlighted that less healthy beverages, such as SSBs (44.1%), were more likely to be price promoted compared to healthier beverages (14.9%).14 However, these studies were short-term and were conducted eight and eleven years ago, respectively. With significant week-to-week fluctuation in price promotions, a current assessment to quantify price promotions throughout the year is required to understand which products are promoted, the degree of price discounting and how trends vary across seasons.

In this study, we conducted a weekly systematic audit of all non-alcoholic beverage price promotions available for sale online at two major Australian supermarket chains (accounting for 67% of the grocery market share),15 over 52 weeks. We additionally audited all non-alcoholic beverages available for sale online at each supermarket (with and without a price promotion) to calculate the proportion of each beverage category that was price promoted each week. We aimed to examine the frequency and magnitude of beverage price promotions, and whether this differed by beverage category or season.

Methods

Data collectionData was collected weekly for 52 weeks from November 2016 to November 2017 from the online websites of the two major Australian supermarket chains, Coles and Woolworths. Weekly data collection was selected to align with the price promotion cycle in these supermarkets (updated weekly on Wednesdays). The following data was collected weekly for all non-alcoholic beverage product types (single purchase items that may include, for example, a single can or a 24-pack of cans; hereafter referred to as ‘beverage/s’) where the sale price was less than the regular retail price: product name, volume, pack size, regular retail price, promotional price and whether the promotion was a ‘multi-buy’ promotion. A price promotion was defined as a temporary price reduction. Products advertised as ‘everyday low price’ were not considered a price promotion as the prices for these items did not vary across weeks. A ‘multi-buy’ price promotion was defined as a price promotion that required consumers to purchase more than one unit to receive the discount (i.e.

two for $15, three for $10; two for the price of one). Data was not collected for beverages requiring significant preparation before consumption, such as tea, coffee beans, chocolate syrups and drink powders (with the exception of cordial, a concentrated sugar-sweetened beverage requiring water for preparation, being a popular children’s beverage in Australia). A complete audit of the price of all ready-to-drink beverages and cordials (regardless of whether they were price promoted or not) was conducted in May 2017 by one member of the research team (BG). This audit was conducted manually by recording the data into a Microsoft Excel™ spreadsheet and combined with the weekly data on price-promoted beverages to determine the proportion of each beverage category that was price promoted each week.

Four trained researchers collected the data on a rotating roster. For the first 26 weeks, price data was manually collected by entering the product information into an excel spreadsheet. For the remaining 26 weeks, data collection was conducted using an automated online scraping tool, which extracted and exported the necessary information into a spreadsheet. This data was manually checked each week to ensure information was extracted for the correct number of products, with a random 50 products checked for data accuracy (all of which indicated 100% accurate data extraction). Two weeks of data were excluded due to data collection errors, leaving 50 weeks of data for analysis.

The validity of using online data for this project was confirmed in a prior study where we tested the correlation between food and beverage availability and price, online and in-store, for both Coles and Woolworths. In that study, we randomly selected 96 products from four categories (breakfast cereals, cereal based bars, juices and sugar-sweetened beverages) using the Australian Food Switch database (>40,000 supermarket food and beverage products).16 We found a high correlation (>90%) for the availability of products and the presence of price promotion for a given product, online and in-store (unpublished results).

Beverage classificationEach beverage was classified into one of four policy-relevant6 categories (‘SSBs’, ‘Artificially-Sweetened Beverages’ (ASBs), ‘flavoured milk and 100% juice’, ‘milk and water’; see Table 1). Flavoured milk and 100% fruit or vegetable juices were not included in the SSB category because, although these products contain sugar, they typically have a higher nutritional value compared to other SSBs, and consequently are often exempt from interventions and policies targeting sugary drinks, including most SSB taxes.6 Milk and water were purposely classified as distinct from ASBs because of the nutritional importance of these products within a healthy diet.17

Data analysisThe proportion of beverages on price promotion in any given week within the available product category (number of price-promoted beverages within a product category/total number of beverages within that beverage category), and the proportional contribution of each beverage category to the total number price-promoted beverages (number of all price-promoted beverages within a product category/total number of price-promoted beverages), was calculated. We additionally calculated the mean discount (%) for each beverage category across the 50 weeks for each beverage category.

Weekly variation in the proportion of each price-promoted beverage category and the proportion of multi-buys for each beverage category was assessed graphically over the one-year of data collection.

Analyses were conducted using Microsoft Excel™.

Table 1: Categories for beverages sold at the two major Australian supermarkets (Coles and Woolworths) between November 2016 and November 2017.

Major category Sub-major category

Sugar-Sweetened Beverages (SSBs)

Soft drink Flavoured water, ice tea, sports or energy drinksFruit-flavoured drinks (<99% juice)Flavoured mineral water (sugar-sweetened)Cordial

Artificially-Sweetened Beverages (ASBs)

Diet soft drink Diet flavoured water, ice tea, sport drinks or energy drinksFlavoured mineral water (no sugar)Diet cordial

Flavoured milk and 100% juice

Flavoured milk100% fruit or vegetable juice

Milk and water Plain full- or low-fat milkPlain still or sparkling water

Food and Beverage Beverage price promotions


Results

Price promotionsAcross both supermarkets, an average of 971 beverages product types were available for sale each week (Coles n=960; Woolworths n=982), of which 40% were SSBs (Coles n=381; Woolworths n=397), 13 and 15% for ASBs (Coles n=120; Woolworths n=143), 28 and 24% for flavoured milks and 100% juice (Coles n=270; Woolworths n=236) and 20 and 21% for plain milk and water (Coles n=189; Woolworths n=206) (Table 2).

On average, in any given week 26% and 30% of all beverages were price promoted for Coles and Woolworths, respectively (Table 2). When examining price promotions within each policy-relevant beverage category, findings from both supermarkets indicated that the proportions of price promotions within beverage categories was similar for SSBs and ASBs (Coles: 30% of all SSBs vs. 33% of all ASBs; Woolworths: 37% of all SSBs vs. 38% of all ASBs), with this finding being consistent across the 50 weeks of the study. The proportion of price-promoted products was lowest for the ‘milk and water’ category with a weekly average of 14% for Coles and 15% for Woolworths (Table 2).

Across all price-promoted beverages (not within beverage categories), the greatest number of price promotions were for SSBs (46% and 49% for Coles and Woolworths, respectively), followed by flavoured milk and 100% juice (27% and 22% of all price-promoted beverages for Coles and Woolworths, respectively), ASBs (16% and 18% of all price-promoted beverages for Coles and Woolworths, respectively) and water and plain milk (11% and 10% of all price-promoted beverages for Coles and Woolworths, respectively). In total, 73% and 71% of price promotions (across all price-promoted beverages) were for sugary drinks (SSBs and flavoured milk and 100% juice combined), at Coles and Woolworths, respectively.

Across the year, the mean price reduction for all beverages was similar for both supermarkets at -33% for Coles and -26% for Woolworths. Price-promoted diet soft drinks (Coles: -40%; Woolworths: -34%) and diet flavoured water, ice tea, sports and energy drinks (Coles: -40%; Woolworths: -29%) were most heavily discounted, followed by sugar-sweetened soft drinks (Coles: -39%; Woolworths: -34%).

Table 2: Weekly mean number and proportion of beverages on price promotion, by beverage category, for the two major Australian Supermarkets (Coles and Woolworths) between November 2016 and November 2017.Beverage Category Beverages in

product line, n (% of all

beverages)

Mean number of price

promoted beverages per week, n (SD)

Mean % of product line price promoted

(SD)

Mean % of all beverage

price promotions

(SD)

Mean price change, %

(SD)

ColesTotal 960 (100) 247 (30) 26 (3) 100 -33 (9)SSBs 381 (40) 115 (18) 30 (5) 46 (4) -36 (11) Cordial 47 (5) 11 (7) 23 (15) 4 (3) -27 (9)

Flavoured water, ice tea, sports and energy drinks

94 (10) 27 (7) 28 (7) 11 (2) -38 (10)

Fruit-flavoured drink (<99%) 79 (8) 25 (8) 31 (11) 10 (3) -33 (14) Flavoured mineral water (sugar-sweetened) 21 (2) 7 (4) 32 (21) 3 (2) -36 (6) Soft drink 140 (15) 46 (9) 33 (6) 19 (3) -39 (10)ASBs 120 (13) 40 (7) 33 (6) 16 (2) -39 (9) Diet cordial 8 (1) 1 (2) 14 (27) 0 (1) -32 (11)

Diet flavoured water, ice tea, sports and energy drinks

27 (3) 8 (3) 30 (10) 3 (1) -40 (8)

Flavoured mineral water (no sugar) 31 (3) 7 (4) 22 (14) 3 (1) -35 (8) Diet soft drink 54 (6) 24 (4) 44 (8) 10 (2) -40 (10)Flavoured milk and 100% juice 270 (28) 66 (14) 24 (5) 27 (5) -26 (9) Flavoured milk 73 (8) 19 (8) 26 (12) 8 (3) -25 (8) 100% fruit or vegetable juice 197 (21) 47 (13) 24 (6) 19 (4) -27 (9)Milk and Water 189 (20) 27 (7) 14 (4) 11 (3) -32 (10) Milk 137 (14) 16 (5) 12 (4) 7 (2) -30 (11) Water 52 (5) 11 (4) 21 (7) 4 (2) -34 (9)WoolworthsTotal 982 (100) 297 (54) 30 (6) 100 -26 (11)SSBs 397 (40) 145 (27) 37 (7) 49 (4) -28 (12) Cordial 62 (6) 15 (8) 24 (13) 5 (2) -21 (8)

Flavoured water, ice tea, sports and energy drinks

106 (11) 45 (9) 43 (9) 15 (3) -28 (12)

Fruit-flavoured drink (<99%) 84 (9) 31 (8) 36 (10) 10 (2) -22 (10) Flavoured mineral water (sugar-sweetened) 39 (4) 10 (6) 27 (14) 3 (2) -29 (7) Soft drink 106 (11) 44 (12) 42 (11) 15 (3) -34 (10)ASBs 143 (15) 55 (11) 38 (8) 18 (3) -30 (11) Diet cordial 13 (1) 6 (4) 49 (30) 2 (1) -19 (5)

Diet flavoured water, ice tea, sports and energy drinks

39 (4) 15 (4) 39 (10) 5 (1) -29 (11)

Flavoured mineral water (no sugar) 31 (3) 8 (4) 26 (12) 3 (1) -28 (8) Diet soft drink 60 (6) 25 (7) 41 (12) 8 (2) -34 (10)Flavoured milk and 100% juice 236 (24) 66 (19) 28 (8) 22 (4) -20 (8) Flavoured milk 75 (8) 17 (8) 23 (11) 6 (3) -21 (8) 100% fruit or vegetable juice 161 (16) 49 (16) 30 (10) 16 (4) -20 (8)Milk and Water 206 (21) 30 (10) 15 (5) 10 (3) -23 (9) Milk 140 (14) 15 (5) 11 (4) 5 (2) -22 (9) Water 66 (7) 15 (6) 23 (10) 5 (2) -24 (9)Note: Mean % of each product line promoted each week (third data column) was calculated by dividing the total number of price promoted products within a

product category by the total number of products available within the category; % of all price promotions (fourth data column) was calculated by dividing the total number of price promoted beverages within a category by the total number of price promoted beverages.

The proportion of each beverage category that was price promoted each week was relatively constant over time for both supermarkets, with both Coles and Woolworths demonstrating a peak during the week of December 14th for both SSBs and ASBs (Figure 1). The proportion of SSBs and

ASBs price promoted in any given week was similar across the year.

Multi-buy price promotionsOn average, in any given week, 4% and 8% of all beverages were available as a multi-buy promotion (a subset of price-promoted

Zorbas et al.



price promotions in four New Zealand supermarkets reported that the majority of all price promotions were for ‘red’ (drink less) beverages (44.1%) compared to ‘amber’ (drink in moderation; 40.9%) and ‘green’ (drink most) beverages (14.9%).(14) Our study further revealed that a much higher proportion of all price promotions were for sugary beverages (73% and 71% for Coles and Woolworths, respectively) compared to non-sugar beverages. Similarly, a 2010-12 cross-sectional audit of price promotions in 955 US supermarkets showed a greater prevalence of price promotions among SSBs (18.2%) compared to non-sugary beverages (12.1%) products.13 However, our contemporary results suggest that this proportion is much higher and, on average, approximately one-third of all SSB products are price promoted.

The strengths of our study include the comprehensive nature of data collection, covering 50 weeks of price promotions cycles within a year, across all seasons and holiday events. Our data is further strengthened by our audit of all beverages available for sale, which allowed us to examine the extent of price promotions relative to their availability. However, this complete audit of all available beverages was also limited to just one collection point, mid-way through the data collection period. Our study is further limited to the availability of price promotions and does not reflect customer purchasing behaviour. The health implications of beverage price promotions depend on their influence on healthy and unhealthy beverage choices – a function of both the frequency and magnitude of price promotions on healthy and unhealthy beverages and consumer responses to such price promotions. While studies from the UK and US show that the impact of price promotions on purchasing behaviour is similar for healthy and less healthy foods,18,19 comparable analyses are not available in the Australian context. Finally, it is important to acknowledge that ‘everyday low prices’ were not included as a price promotion in our study as we were interested in temporary (not ‘everyday’) price reductions. Australian supermarkets use ‘everyday low prices’ on items such as plain milk as a tactic to increase market competitiveness, which may explain the lower proportion of price promotions in the water and plain milk category.

Figure 1: Weekly variation in the proportion of each beverage category price promoted at Coles and Woolworths.

beverages) in Coles and Woolworths, respectively (Table 3). A similar proportion of all SSBs and ASBs were promoted as multi-buys at each store (Coles: 6% and 7% of all SSBs and ASBs, respectively; Woolworths: 11% and 12% for all SSBs and all ASBs).

Of all multi-buy promotions in a given week, the majority were for the SSB category (52% and 59% for Coles and Woolworths, respectively). When combining all sugary drinks (SSBs, flavoured milks and 100% juice), the multi-buys for these beverages made up more than three-quarters of all multi-buy offers (Coles: 74%, Woolworths: 75%). The proportion of multi-buys offered within each beverage category was variable across beverage categories and across supermarkets. Within the beverage categories available at Coles, multi-buys were most common within the flavoured mineral water (sugar-sweetened) category (Coles: 16%, Woolworths: 12%), whereas for Woolworths, beverages within the categories flavoured water, ice teas, sports and energy drinks (Coles: 3%, Woolworths: 14%) and artificially sweetened water, ice teas, sports drinks were most commonly promoted as a multi-buy (Coles: 4%; Woolworths: 15%).

Discussion

This is the first study to systematically and comprehensively quantify the extent and magnitude of price-promoted beverages available for sale, over a 12-month period, in Australian supermarkets. We demonstrate that the frequency of price promotions for sugary drinks (SSBs, flavoured milk and 100% juice combined) is approximately proportional to their availability. On average, sugary drinks constitute two-thirds of all beverage product types available for sale and around two-thirds of all price-promoted beverages in any given week. Within each beverage category, the proportion of all beverage products available for sale with a price promotion did not markedly differ for SSBs and ASBs (approximately one-third of all SSBs and ASBs are price promoted in any given week). The mean discount for price-promoted beverages is also similar across beverage types, with an overall mean price discount of 33% and 26% for Coles and Woolworths, respectively.

Our conclusions are similar to previous international studies of shorter duration. A 2007 four-week audit of beverage

Note: shading on graph represents seasons: December-February (Summer); March-May (Autumn); June-August (Winter); September-November (Spring)


Policy implicationsWe show that, in any given week, the proportion of price-promoted SSBs and ASBs is similar (Coles: 30% of all SSBs, vs. 33% of all ASBs; Woolworths: 37% of all SSBs vs. 38% of ASBs), indicating that these supermarkets do not distinguish between healthy and less healthy beverages when setting price promotions. Rather, it is likely that these supermarkets use price promotions as a way of increasing store traffic and overall sales. Nevertheless, the ubiquity of price promotions on sugary drinks supports recent calls by public health coalitions and governments for a ban on unhealthy food and beverage price promotions.8-10 A modelling study from the UK further supports these policy recommendations, finding that, on average, one-fifth of the volume of price-promoted food and beverages sold can be considered to be in addition to what would be sold were the promotion not in place (i.e. on top of the substitution effect from non-price-promoted products).20 We are not aware of any empirical studies examining behavioural responses to removing price promotions on sugary drinks. Such evidence would help refine these policy recommendations.

This research highlights that public health SSB pricing interventions may need to extend beyond a tax on SSBs and consider policies that reduce the influence of price promotions on consumer purchasing behaviour. With international SSB taxes commonly set at 10-20%, the magnitude and regularity of SSB price promotions may attenuate the impact of any future SSB tax in Australia.6 Policies that reduce the influence of SSB price promotions, such as restrictions on unhealthy beverages (and food), would create an even pricing playing field across all supermarkets and may ameliorate any financial impact to industry – a core concern for industry lobbyists. Alternative policy options may include a restriction on the advertising of price promotions in-store, as has been suggested by the Scottish government,21 however, more research is required to understand the impact of such policies on beverage choices and population health.

Our results demonstrating that the availability of sugary drinks is proportional to price promotion frequency, suggest that interventions to increase the relative availability of healthier beverages, compared to unhealthy beverages, may also

Table 3: Weekly mean number and proportion of beverages on ‘multi-buy’ price promotion, by product category, for the two major Australian Supermarkets (Coles and Woolworths) between November 2016 and November 2017.Beverage Category Mean number of

multi-buy beverages per week, n (SD)

Mean % of product line (SD)

Mean % of all multi-buys (SD)

Coles (Total) 41 (25) 4 (3) 100 SSBs 22 (16) 6 (4) 52 (16) Cordial 2 (4) 3 (8) 2 (6) Flavoured water, ice tea, sports and energy drinks 3 (3) 3 (4) 9 (13) Fruit-flavoured drink (<99%) 5 (6) 7 (7) 12 (12) Flavoured mineral water (sugar-sweetened) 3 (4) 16 (21) 9 (11) Soft drink 9 (8) 6 (6) 21 (16)ASBs 8 (6) 7 (5) 21 (11) Diet cordial 1 (2) 7 (20) 1 (3) Diet flavoured water, ice tea, sports and energy drinks 1 (2) 4 (6) 4 (7) Flavoured mineral water (no sugar) 2 (3) 8 (11) 6 (8) Diet soft drink 4 (3) 7 (6) 10 (8)Flavoured milk and 100% juice 9 (8) 3 (3) 22 (21) Flavoured milk 0 (0) 0 (0) 0 (0) 100% fruit or vegetable juice 9 (8) 5 (4) 22 (21)Milk and water 1 (2) 1 (1) 5 (15) Milk 0 (1) 0 (1) 3 (14) Water 1 (1) 1 (2) 2 (6)Woolworths (Total) 79 (30) 8 (3) 100SSBs 47 (20) 11 (5) 59 (6) Cordial 3 (5) 5 (8) 4 (6) Flavoured water, ice tea, sports and energy drinks 16 (7) 14 (7) 20 (8) Fruit-flavoured drink (<99%) 11 (6) 13 (8) 15 (9) Mineral water (sugar sweetened) 5 (5) 12 (12) 6 (5) Soft drink 12 (8) 11 (8) 14 (8)ASBs 17 (9) 12 (6) 21 (6) Diet cordial 2 (3) 13 (20) 2 (3) Diet flavoured water, ice tea, sports and energy drinks 6 (3) 15 (8) 8 (4) Flavoured mineral water (no sugar) 3 (3) 9 (10) 4 (4) Diet soft drink 7 (5) 11 (8) 8 (5)Flavoured milk and 100% juice 11 (5) 5 (2) 16 (7) Flavoured milk 2 (1) 2 (2) 2 (2) 100% fruit or vegetable juice 10 (4) 6 (3) 13 (6)Milk and water 3 (3) 2 (2) 4 (3) Milk 1 (2) 1 (1) 1 (2) Water 3 (3) 4 (4) 3 (3)

inadvertently reduce the number of price promotions for sugary drinks. However, any such changes would need to be monitored carefully to determine if the changes are likely to have the intended public health impact.

Conclusion

Price promotions are used extensively for beverages sold in Australian supermarkets, with the vast majority of available price promotions for sugary drinks, undermining efforts to promote healthy population diets. Policies to restrict price promotions on SSBs are likely to be an important part of any approach to reduce obesity and improve

population nutrition. Empirical studies to evaluate the likely impact of such a policy are clearly required.

FundingThis work is supported by a National Heart Foundation Vanguard grant (101674). AJC is supported by a DECRA fellowship (DE160100141) from the Australian Research Council and is a researcher within a NHMRC Centre for Research Excellence in Obesity Policy and Food Systems (APP1041020). JW is supported by a UNSW Scientia Fellowship. The funding sources had no role in the design, analysis or writing of this article.

Zorbas et al.


References1. Guh DP, Zhang W, Bansback N, Amarsi Z, Birmingham

CL, Anis AH. The incidence of co-morbidities related to obesity and overweight: A systematic review and meta-analysis. BMC Public Health. 2009;9:88.

2. Australian Institute of Health and Welfare. Overweight and Obesity in Australia: A Birth Cohort Analysis. Catalogue No.: PHE 215. Canberra (AUST): AIHW; 2017.

3. Swinburn BA, Sacks G, Hall KD, McPherson K, Finegood DT, Moodie ML, et al. The global obesity pandemic: Shaped by global drivers and local environments. Lancet. 2011;378(9793):804-14.

4. Australian Bureau of Statistics. Australian Health Survey: Nutrition First Results - Foods and Nutrients, 2011-12. Consumption of Sweetened Beverages. Canberra (AUST): ABS; 2015.

5. Australian Bureau of Statistics. Australian Health Survey: Consumption of Added Sugars, 2011-12. Consumption of Added Sugars - A Comparison of 1995 to 2011-12. Canberra (AUST): ABS; 2017.

6. Backholer K, Blake M, Vandevijvere S. Sugar-sweetened beverage taxation: An update on the year that was 2017. Public Health Nutr. 2017;20(18):3219-24.

7. Hawkes C. Sales promotions and food consumption. Nutr Rev. 2009;67(6):333-42.

8. Obesity and Price Promotions (Briefing). Glasgow (SCO): Obesity Action Scotland; 2016.

9. Obesity Health Alliance. Joint Policy Position on Obesity. London (UK): Obesity Health Alliance; 2017.

10. Sacks G, Robinson E, Cameron A for INFORMAS. Inside our Supermarkets: Assessment of Company Policies and Commitments Related to Obesity Prevention and Nutrition, Australia 2018. Melbourne (AUST): Deakin University, 2018.

11. Johnston R, Stafford J, Pierce H, Daube M. Alcohol promotions in Australian supermarket catalogues. Drug Alcohol Rev. 2017;36(4):456-63.

12. Scottish Parliament. Alcohol etc. (Scotland) Act 2010 [Internet]. 2010 [cited 2019 Fb 12]. Available from: http://www.legislation.gov.uk/asp/2010/18/pdfs/asp_20100018_en.pdf

13. Powell LM, Kumanyika SK, Isgor Z, Rimkus L, Zenk SN, Chaloupka FJ. Price promotions for food and beverage products in a nationwide sample of food stores. Prev Med. 2016;86:106-13.

14. Pollock S, Signal L, Watts C. Supermarket discounts: Are they promoting healthy non‐alcoholic beverages? Nutr Diet. 2009;66(2):101-7.

15. Youl T. IBISWorld Industry Report G4111: Supermarkets and Grocery Stores in Australia. Los Angeles (CA): IBISWorld; 2018 October.

16. Haskelberg H, Neal B, Dunford E, Flood V, Rangan A, Thomas B, et al. High variation in manufacturer-declared serving size of packaged discretionary foods in Australia. Br J Nutr. 2016;115(10):1810-18.


18. Taillie LS, Ng SW, Xue Y, Harding M. Deal or no deal? The prevalence and nutritional quality of price promotions among U.S. food and beverage purchases. Appetite. 2017;117:365-72.

19. Nakamura R, Suhrcke M, Jebb SA, Pechey R, Almiron-Roig E, Marteau TM. Price promotions on healthier compared with less healthy foods: A hierarchical regression analysis of the impact on sales and social patterning of responses to promotions in Great Britain. Am J Clin Nutr. 2015;101(4):808-16.

20. Smithson M, Kirk J, Capelin C. An Analysis of the Role of Price Promotions on the Household Purchases of Food and Drinks High in Sugar: A Research Project for Public Health England Conducted by Kantar Worldpanel UK. London (UK): Public Health England; 2015.

21. A Healthier Future - Scotland’s Diet & Healthy Weight Delivery Plan. Edinburgh (SCO). Government of Scotland; 2018.



Working with the food industry for public health good presents challenges and opportunities.

Differing fundamental foci, for example on profit versus health, mean that food industry actions can directly contribute to public health (e.g. supporting growers producing fruit and vegetables) or undermine it (such as allowing the proliferation of cheap, unhealthy commodities).

Front-of-pack nutrition labelling systems (FoPL) are recommended by the World Health Organization as a tool to promote healthier diets.1 Their development requires multi-stakeholder negotiation. However, as FoPL can change purchasing intent,2 they are opposed by some industries whose profits rely on foods detrimental to health.

This paper deals specifically with the process leading to the adoption of the Health Star Rating (HSR) FoPL in Australia and New Zealand up to 2014. The controversies that followed the HSR adoption are outside the scope of this paper.3 We reflect on the Public Health Association of Australia (PHAA) actions to improve nutrition for more than a decade leading up to the development of the HSR. These include prioritising both a National Nutrition Policy and the development of a health advocacy tool based largely on 10 sequential steps for planning or evaluating public health advocacy4 (see Figure 1). The lessons we draw are consistent with the findings of Kumar et al.5 who conclude:

Strong leadership, policy entrepreneurship and a coherent alliance between public health and consumer groups enabled the development of a FoPL system in Australia and could contribute to advancing FoPL standards at the international level.5

Background

The Australian Federal Government commissioned former Federal Labor Health Minister and academic, Dr Neal Blewett,

to lead a review into food labelling law and policy in 2011. Consistent with PHAA’s prior call for a colour-coded multiple traffic lights (MTL) system, Blewett’s final ‘Labelling Logic’ report found “MTL systems were the most effective in facilitating consumers’ understanding of the nutrient profiles across foods within and across food categories”.6 It recommended: an interpretative FoPL system be developed reflective of a comprehensive Nutrition Policy (Recommendation 50); a MTL FOPL system be introduced that was

Development of Australia’s front-of-pack interpretative nutrition labelling Health Star Rating system: lessons for public health advocates Michael Moore,1,2 Alexandra Jones,1 Christina M. Pollard,3 Heather Yeatman4

1. The George Institute for Global Health, UNSW Sydney, New South Wales2. UC Health Research Institute, University of Canberra, Australian Capital Territory3. School of Public Health, Curtin University, Western Australia4. Faculty of Social Science, University of Wollongong, New South WalesCorrespondence to: Adjunct Professor Michael Moore, The George Institute for Global Health, UNSW, Sydney, New South Wales; e-mail: [email protected]: January 2019; Revision requested: March 2019; Accepted: April 2019The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,



Abstract

Objectives: To draw advocacy lessons from actions undertaken by public health groups to assist the development of Australia and New Zealand’s Health Star Rating (HSR) front-of-pack nutrition labelling system.

Methods: The advocacy approaches undertaken by the Public Health Association of Australia leading up to the time of the adoption of the HSR is examined using a 10 step advocacy framework. Key roles in advocacy planning and implementation are described, along with coordinating efforts by health and consumer groups during the HSR development processes.

Results: HSR aims to support consumers to make informed choices to protect from diet-related diseases, including obesity. The HSR launched despite a number of major obstacles, owing to a strategic, coordinated advocacy effort undertaken by a guiding coalition.

Conclusions: Actions to improve nutrition are often highly contested, particularly if the desired outcome competes with commercial interests. However, by deploying a structured approach to public health advocacy it is possible to influence government despite opposition from commercial interests.

Implications for public health: A shared vision and a coordinated effort by public health professionals enabled advocates to overcome undue commercial influence.

Key words: advocacy, nutrition, public health, Front of Pack Labelling (FoPL), Health Star Rating (HSR)

FOOD AND BEVERAGE


initially voluntary but mandatory for general or high-level health claims or equivalent (Recommendation 51); that government provides advice and support for producers adopting the MTL and educates consumers (Recommendation 52); and monitoring industry compliance and evaluating food supply and consumer food choice improvements (Recommendation 53).

Thwarted on traffic lights and next steps The Legislative and Governance Forum on Food Regulation (Forum), later to be the Australia and New Zealand Ministerial Forum on Food Regulation (FoFR), rejected Recommendation 51, specifically ruling out MTL.

The FoFR did accept the more general Recommendation 50: “an interpretative front-of-pack labelling system be developed that is reflective of a comprehensive Nutrition Policy and agreed public health priorities”.6 However, there was as yet no Nutrition Policy. Ministers delegated the process to the Food Regulation Standing Committee (FRSC), which is made up of senior public servants. FRSC determined the specific members of the FoPL Steering and Project Committee (SPC) who were drawn from industry, public health and consumer stakeholders. The development of a FoPL was to be a collaborative process, following a set of objectives and principles provided by Ministers that were already a balancing act between health and profitability.7 The choice of stakeholders by FRSC reflects the importance of Advocacy Step 3 “building and maintaining influential relationships”.

The guiding coalitionPrior to the first meeting of the FoPL SPC, 16 public health and consumer organisations (Figure 2) held a strategy meeting to generate a sense of urgency, form a ‘guiding coalition’, strengthen relationships, and develop a

shared vision for change (Advocacy Steps 1, 2 and 4, Figure 1).8 Throughout the process, the PHAA and others continued to advocate for an interpretative MTL FoPL scheme to be initiated as part of a National Nutrition Policy.

The consumer and public health guiding coalition agreed on a series of principles, announced in a media statement (Advocacy Step 5, “communicating the vision for buy-in”) released on the day of the SPC’s first meeting. It concluded with calling for:

... an interpretive system that includes colours and symbols that are easy to understand, provides a quick comparison between different products, and makes healthy choices easy.9

The guiding coalition also established its bottom line, the compromises they would be willing to make – beyond which they would walk away – and an agreed public position. Each member of the guiding coalition acted as a representative of their organisation and conduit for feedback on negotiations. The process moved quickly and there was little time for standard consultation processes and procedures. Each organisation relied on their current policy positions for guidance, which in the case of the PHAA were developed through the Food and Nutrition Special Interest Group (FANSIG). Resource limitations and government procedural processes meant only a small number of technical experts were present during complex political and technical negotiations. The contested and time-bound nature of policy development meant that some individuals with extensive relevant nutrition science expertise who had originally advised government were no longer involved in direct negotiations.

Challenges of working with industryWithin the SPC, an agreed outcome was challenging as the committee comprised of multiple stakeholders. Health and consumer advocates sought clear messages for public health, while industry advocates remained protective of their profit motive. At the first SPC meeting, the concept of star ratings – similar to those already in the Australian market to rate hotels and movies – was agreed. A label format and suitable criteria for rating individual food and drink products to align with the Australian Dietary Guidelines was required. Collectively, the SPC agreed to “aim for a gold medal – but accept a position on the podium”.10

A Technical Design Working Group (TDWG) was established to seek the most effective, defensible and consistent approach to applying the Health Stars as the system developed.11 Additionally, an Implementation, Evaluation and Education Working Group (IEEWG) examined regulatory options. Both groups had wide representations but limited time for deliberations. Vigorous discussion ensued before reaching agreement for an HSR scoring system based on a pre-existing nutrient profiling scoring criteria (NPSC) already used to for health claims. The information about the adaptation of the NPSC has been recently published as part of HSR’s five-year review.12

The greatest challenge in development of the HSR was having industry renege on agreed positions.

Industry renegesThere was initial agreement by industry groups to adopt the scheme, but some industry members reneged on the position to adopt the use of stars and the algorithm. The guiding coalition moved quickly, consistent with Advocacy Step 7: Be Opportunistic. Parallel to the development of the HSR the guiding coalition members continued to take actions to strengthen outcomes for public health benefit, as did industry for commercial benefit. Although the HSR system was a collaboratively agreed product, sources revealed industry players were approaching Ministers prior to the FoFR meetings intent on blocking the agreement. In response to these actions, the PHAA ‘opportunistically’ approached Ministers on the morning of the Forum meeting, reiterating support for the HSR. Ministers rejected industry lobbyists’ approaches, viewing them as ‘reneging’ on

Figure 1: The Advocacy Tool.Step 1: Establishing a Sense of UrgencyStep 2: Creating the Guiding Coalition Step 3: Developing and Maintaining Influential RelationshipsStep 4: Developing a Change VisionStep 5: Communicating the Vision for Buy-inStep 6: Empowering Broad-based ActionStep 7: Be OpportunisticStep 8: Generating Short-term WinsStep 9: Never Letting Up Step 10: Incorporating Changes into the Culture

Figure 2: The ‘Guiding Coalition’.Australian Chronic Disease Prevention AllianceAustralian Medical Association Australian Division of World Action on Salt and HealthCancer Council Australia Cancer Council NSW CHOICE Diabetes Australia Diabetes Australia VicDietitians Association AustraliaNational Heart FoundationKidney Health AustraliaNational Stroke FoundationThe George Institute for Global HealthPhysical Activity, Nutrition and Obesity Research GroupObesity Policy Coalition Public Health Association AustraliaUniversity of Wollongong

Food and Beverage Lessons for public health advocates


an agreement. They approved the HSR ‘in principle’ at the Forum meeting in Sydney in June 2013.

Some supportive food companies were waiting for the algorithm to be made public via an HSR website to begin using the HSR. Once the system was ‘live’, any person could assess individual food products online for their relative healthfulness according to the algorithm. Other manufacturers, with products of limited health value, were nervous about its impact and sought to lessen the scheme’s effectiveness, including seeking to have the HSR website removed.

Industry players continued lobbying to undermine agreed HSR positions, particularly following the official launch of the HSR website in early February 2014. The Australian Federal Food Minister, at the behest of her then Chief of Staff and without consulting all other ministers, ordered the HSR website taken down within hours of its launch online. It was later discovered the Chief of Staff had a conflict of interest, having previously worked as a consultant to a major confectionery manufacturer and not severed all ties.13

Timely advocacyThe guiding coalition responded quickly to the website removal, meeting and agreeing to take turns creating media opportunities to keep the issue on the agenda (Advocacy Step 8: Generating short term wins). The Sydney Morning Herald health editor wrote the first story.13 A week of questioning followed in the media, in the Senate and through public questioning of government. Examples of HSR on foods were published, 66 professors of health called for reinstatement of the website and public health professionals published advocacy pieces.14 Eventually, Ministers agreed to reinstate the website with a compromise to allow all packaged foods to be included and the HSR be on a voluntary basis for five years, subject to a two-year review of progress. They later agreed the system would be subject to a comprehensive formal five-year review, due in 2019.

The HSR represents an important improvement in nutrition labelling for consumers but concerns remain about the performance of its algorithm in guiding consumers towards genuinely healthier choices.15 The HSR represents an important improvement in nutrition labelling for consumers. A predominant focus of the review has been to assess whether it

adequately aligns with evidence-based dietary advice, particularly that of the Australian Dietary Guidelines. During the HSR development, it was agreed that the uptake needed to be ‘widespread and consistent’ and there was a condition that it remain voluntary unless this did not occur, at which point it would be made mandatory. By June 2018, in Australia, the HSR was on more than 10,300 products and over 3,900 in New Zealand.16 However, HSR remains on less than one-third of products overall, and these are mostly those that score well.17 Moreover, Australia still does not have a wider National Nutrition Policy.

It is incumbent on public health professionals to maintain their persistence and work to improve the efficacy of the HSR system (Advocacy Step 9: Never letting up). It also is critical the HSR is just one of the tools in improving nutrition and health outcomes. Advocates continue to pursue a National Nutrition Policy18 to guide the development and implementation of a comprehensive set of public health interventions for improved dietary patterns ‘incorporated into the culture’ (Advocacy Step 10).

Conclusion

Successful advocacy requires systematic and objective reflection on past actions. While different approaches are required in different circumstances, advocacy does have common elements. The ten sequential steps applied in the development process of the HSR system on packaged food for public health benefit provide an important case study in public health advocacy.

References1. World Health Organization Regional Office for Europe.

European Food and Nutrition Action Plan 2015-2020: Time Frame. Copenhagen (DNK): Euro WHO; 2015. p. 19.

2. National Heart Foundation of Australia. Report on the Monitoring of the Implementation of the Health Star Rating System: Key Findings for Area of Enquiry Two – Consumer Awareness and Ability to Use the Health Star Rating System Correctly. Canberra (AUST): Australian Department of Health; 2017 July.

3 Lawrence MA, Dickie S, Woods JL. (2018) Do nutrient-based front-of-pack labelling schemes support or undermine food-based dietary guideline recommendations? Lessons from the Australian Health Star Rating System. Nutrients. 2018;10(1). pii: E32. doi: 10.3390/nu10010032

4. Moore M, Yeatman H, Pollard C. Evaluating success in public health advocacy strategies. Vietnam J Public Health. 2013;1(1):66-75.

5 Kumar M, Gleeson D, Barraclough S. Australia’s Health Star Rating policy process: Lessons for global policy-making in front-of-pack nutrition labelling. Nutr Diet. 2018;75(2):193-199.

6. Blewett N, Goddard N, Pettigrew S, Reynolds C, Yeatman H. Labelling logic: Review of Food Labelling Law and Policy. Canberra (AUST): Australian Department of Health and Ageing. 2011.

7. Australia New Zealand Food Regulation Ministerial Council. Front of Pack Labelling Policy Statement. Canberra (AUST): Australian Department of Health Food Regulation Secretariat; 2009.

8. FoPL Project Committee. Front of Pack Labelling Project Committee - Objectives and Principles for the Development of a Front-of-pack Labelling (FoPL) System. Canberra (AUST): Australian Department of Health Food Regulation Secretariat; 2012.

9. Australian Chronic Disease Prevention Alliance. Media Release: Health and Consumer Alliance Sends Unified Message to Govt: Clear Food Labelling a Key to Healthier Australia. Sydney (AUST): Cancer Coucil Australia; 2012 March 26.

10. Moore M. FoPL Steering and Project Committee Star Ratings Concept Meeting Notes 2012 March. Unpublished observations.

11. Australia and New Zealand Food Regulation Secretariat. Front-of-pack Labelling Committee and Working Group Meetings. Canberra (AUST): Australian Department of Health Food Regulation Secretariat; 2018.

12. MPConsulting. Five Year Review of the HSR System - Technical Paper, History and Development of the Health Star Rating Algorithm. Canberra (AUST): Department of Health Health Star Rating Technical Advisory Group; 2018.

13. Corderoy A, Massola J. Government official who opposed healthy food website owns shares in food lobbying company. Sydney Morning Herald. 2014 February 12.

14. Corderoy A. Anger as federal food guide is pulled from web. Sydney Morning Herald. 2014 February 10.

15. Lawrence M, Pollard C. Food labels are about informing choice, not some nanny state. J Home Econ Inst Aust. 2014;21(1):40.

16. Australia and New Zealand Ministerial Forum. Australia and New Zealand Ministerial Forum on Food Regulation. Communique—29 June 2018 [Internet]. Canberra (AUST): Australian Department of Health Food Regulation Secretariat; 2018 [cited 2019 Mar 21]. Available from: http://foodregulation.gov.au/internet/fr/publishing.nsf/Content/forum-communique-2018-June

17. Jones A, Shahid M, Neal B. Uptake of Australia’s Health Star Rating System. Nutrients. 2018,10(8):997.

18. Public Health Assocation of Australia Food and Nutrition Special Interest Group. National Nutrition Policy. Canberra (AUST): PHAA; 2018.

Moore et al. Article


Unhealthy diets – high in salt, harmful saturated and trans fats, added sugar and energy – are a leading cause

of death and disability globally.1 Australia has some of the highest obesity rates in the world: nearly two-thirds of Australian adults and one in four children are overweight or obese. Unprecedented availability and aggressive marketing of processed and pre-packaged foods and beverages are a key driver of obesity and diet-related conditions including high blood pressure, heart disease, type 2 diabetes and dental caries.2 Obesity is estimated to cost Australia more than $8.6 billion annually.3

Interpretive front-of-pack nutrition labels (FoPL) are recommended by the World Health Organization (WHO) as an evidence-based policy to promote healthier diets.4,5 These types of labels use nutrient profiling to assess the nutritional quality of individual foods and display this in a simplified, visual form. There is growing evidence that FoPL have potential to improve nutrition literacy, guide consumer choice and incentivise industry to improve their product formulations.6,7 While not a complete source of dietary advice, FoPL is recognised by WHO as a helpful tool to use in conjunction with interventions aimed at improving the overall nutritional quality of diets.8 At least 16 government-endorsed schemes in various formats are operating in over 23 countries.9 This proliferation of formats has prompted the international food

standards agency, the Codex Alimentarius Commission, to commence work developing further international guidance on FoPL.10

In June 2014, Australia and New Zealand adopted a voluntary FoPL in the form of the Health Star Rating system (HSR) following a lengthy process of development involving federal, state and territory governments in collaboration with industry, public health

and consumer groups.11 In short, HSR aims to “provide convenient, relevant and readily understood nutrition information and/or guidance on food packs to assist consumers to make informed food purchases and healthier eating choices”.12 Its developers also recognised that the system should aim to be aligned with existing health strategies and guidelines, and provide incentives for

The performance and potential of the Australasian Health Star Rating system: a four-year review using the RE-AIM frameworkAlexandra Jones,1,2 Anne Marie Thow,3 Cliona Ni Mhurchu,1,4 Gary Sacks,5 Bruce Neal1,6

1. George Institute for Global Health, UNSW, Sydney, New South Wales2. Charles Perkins Centre, The University of Sydney, New South Wales3. Menzies Centre for Health Policy, The University of Sydney, New South Wales4. National Institute for Health Innovation, University of Auckland, New Zealand5. School of Health and Social Development, Deakin University, Melbourne, Victoria6. Department of Epidemiology and Biostatistics, School of Public Health, Faculty of Medicine, Imperial College London, London, United KingdomCorrespondence to: Alexandra Jones, George Institute for Global Health – Food Policy, 1 King St, Newtown, Sydney, NSW 2042; e-mail: [email protected] Submitted: December 2018; Revision requested: April 2019; Accepted: April 2019The authors have stated the following conflict of interest: Alexandra Jones was a member of the Health Star Rating’s Technical Advisory Group (TAG) between 2017 and 2018.

Cliona Ni Mhurchu is a member of the New Zealand Health Star Rating Advisory Group (HSRAG). Neither TAG nor HSRAG had any role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.



Abstract

Objective: The Health Star Rating (HSR) is a front-of-pack nutrition labelling system, implemented voluntarily in Australia and New Zealand since 2014. Our aim was to evaluate HSR’s performance.

Method: We used data from peer-reviewed publications and government-commissioned monitoring and evaluation, websites and communiqués to evaluate HSR’s performance between June 2014 and October 2018 using the RE-AIM (Reach, Efficacy, Adoption, Implementation and Maintenance) framework.

Results: Thirty-three peer-reviewed publications, 21 government and three independent reports informed the assessment. Awareness and trust in HSR was increasing, though campaign reach remained low. Consumers liked, could understand and use the HSR logo, though effects on purchasing were largely unknown. The algorithm was the focus of a formal review. HSR was present on 20-28% of products but biased to those that scored better (HSR≥3.0). Necessary stakeholders were mostly engaged.

Conclusions: A substantial body of work supports continuation and strengthening of HSR. Reasonable refinements to HSR’s star graphic and algorithm, action to initiate mandatory implementation, and strengthened HSR governance present the clearest opportunities for improving public health impact.

Implications for public health: Development and implementation of government-led front-of-pack nutrition labelling systems have the potential to improve public health, while engaging a diverse set of stakeholders.

Key words: food labelling, nutrition, food policy, health star rating, obesity

FOOD AND BEVERAGE


Jones et al. Article

improvements to the healthiness of the food supply.13

The HSR System has three components: an underlying algorithm, the label graphic and an accompanying education campaign.

The algorithm assigns a rating from 0.5 (least healthy) to 5.0 stars (most healthy) in ten half-star increments, assessing both ‘risk’ components (total energy, total sugars, saturated fat, sodium) and ‘positive’ components of food (fibre, fruit, vegetable, nut and legume content (FVNL) and in some cases, protein). It derives from an existing model used to regulate health and nutrient content claims in both countries, embedded in the Australia New Zealand Food Standards Code.14 It was adapted for HSR in consultation with Food Standards Australia New Zealand (FSANZ) and technical and nutrition experts, including industry representatives.15,16

Where they elect to utilise the system, food manufacturers are responsible for correct and accurate use of government guidance

material specifying how to display one of several permitted variants of the HSR graphic.17 No fee or charge is payable to any party for HSR use, with manufacturers bearing the cost of updating their own packages. Roll-out has been accompanied by government-funded education campaigns specific to each country.

At its adoption, Australian and New Zealand Food Ministers agreed HSR would remain voluntary for five years, and subject to a two-year review of progress.18 They later agreed the system would be subject to a comprehensive formal review, due to be delivered by mid-2019.19

The aim of this study was to evaluate the extent to which the HSR had achieved its objectives since implementation and to contribute recommendations on how its public health impact may be enhanced. This evaluation was independent and separate from the formal review commissioned by government.

Methods

We evaluated HSR with the RE-AIM framework, a method widely used to assess the public health impact of health promotion programs.20,21 The five dimensions of the framework (Reach, Efficacy, Adoption, Implementation and Maintenance) are particularly appropriate for evaluating the implementation of population health policy, allowing assessment of both the process and outcomes. In Table 1 we define metrics for evaluating each of the RE-AIM dimensions as they apply to implementation of HSR.

Data sources and criteria for inclusionWe conducted the evaluation using two sources of information:

• Government-issued information on HSR implementation (e.g. official websites, communiqués, monitoring reports and commissioned research)

Table 1: Operationalising the RE-AIM Framework for evaluation of the HSR system.Dimension Description Definition in context Metrics for assessment Data sources identified through search (n)a

Reach Proportion, and representativeness of the target population that participates in the policy

Extent to which the Australian and New Zealand population has access to HSR

Fraction of population that:• Is aware of the HSR system (unprompted and prompted)• Trust HSR• Has been exposed to the HSR campaign

Government-commissioned nationally representative surveys on awareness, understanding and use (12)Government-commissioned campaign evaluations (10)

Efficacy Extent to which the policy has delivered outcomes in the target population

Extent to which HSR is guiding consumers towards healthier choices

Efficacy of HSR label graphic• Consumer understanding and use• Impact on choice and purchasing• Impact in driving industry reformulationEfficacy of HSR algorithm• Alignment with current nutrition, medical and behavioural

sciences literature (content validity)• Alignment with other health and nutrition policies

(construct validity)• Alignment with health outcomes (predictive validity)Efficacy of HSR campaign• Consumer understanding• Impact on call to action

Independent, peer-reviewed research: RCTs, randomized online surveys, choice experiments, focus groups, cross-sectional examination of food supply (28)Government-commissioned nationally representative surveys (11)Government-commissioned reports on alignment with other policies and reformulation (2)Government-commissioned campaign evaluations (10)Independent report benchmarking HSR against international best practice (2)

Adoption The degree to which the necessary settings have been engaged in the policy

Degree to which HSR is operating as an Australian and New Zealand governments initiative in partnership with industry, public health and consumer groups

Representation of each stakeholder in governance structures for HSR implementationStakeholder analysis of involvement, interest, power and impact of HSR on each actor

Government websites: HSR; Food Regulation; AusTender, Department of Health (Aus), Ministry of Primary Industries (NZ)Government-commissioned media analysis (1)Website and reports of the Independent Reviewer

Implementation Extent to which the policy actually has been implemented as intended in the real world

Extent to which HSR has actually been implemented as intended including the number of products displaying HSR and compliance of labels with HSR guidance materials

Uptake of HSR on product labels• Number of products displaying, proportion of food supply• HSR status of those displaying• Number of manufacturers displaying HSRCompliance of labels displayed with HSR Guidance materials

Independent, peer-reviewed research, cross-sectional examination of food supply (5)Government-commissioned monitoring and evaluation reports (6)

Maintenance How the policy is sustained over time and is evaluated

Initiatives, implemented as a direct consequence of HSR, designed to enforce and sustain the intervention and monitor its effects

HSR governance• Dedicated funding sources• Monitoring and evaluation mechanisms• Enforcement mechanisms, including anomaly and dispute

processes and other mechanisms to ensure compliance, transparency and accountability

Government websites: HSR; Food Regulation; AusTender Department of Health (Aus), Ministry of Primary Industries (NZ)Government-issued communiqués and budget papersReports of the independent reviewer (2)

Notes:

a: Data sources may cover more than one outcome or RE-AIM dimension e.g. reports which consolidate data on general HSR awareness, and operation of the HSR campaign


Food and Beverage A four-year review using the RE-AIM framework

• Secondary data from peer-reviewed and grey literature (e.g. reports produced by industry, consumer or public health stakeholders).

All materials were publicly available and obtained using a systematic search strategy outlined in Supplementary Appendix 1.

We limited our search to materials produced after HSR’s adoption and, given our focus on implementation, excluded any materials published about HSR’s development before and after this date. We also excluded materials concerning use of HSR in other jurisdictions and settings (e.g. on labels in other countries, or on foods or products for which it wasn’t intended). We focused on original analysis, and therefore excluded commentaries and editorials that repeated information already included through original research. To keep the exercise manageable, we excluded individual media items, but included summary media analysis released by government. We included materials published up to and including 1 October 2018.

Extraction and coding of dataWe created a database of materials on HSR implementation. For each item, we extracted standard information including: author(s), title, date and place of publication, type of publication (e.g. peer-reviewed research, government-commissioned report) and jurisdiction covered (Australia and/or New Zealand). We also extracted information on study design, population and/or data relied upon to assist in evaluating the strength of the evidence obtained. Finally, we coded materials by component of the HSR System reviewed (algorithm, label and/or education campaign), outcome evaluated (awareness, understanding, use, uptake, alignment with existing policies), and relevant RE-AIM dimension. This database is included in Supplementary Appendix 1.

AnalysisOutcomes of the literature review were summarised by each RE-AIM dimension and synthesised where possible in tables and figures to provide an overall view of the degree to which each dimension has been achieved. To evaluate Adoption, we also adapted a stakeholder analysis approach used by Brugha and Varvasovzsky,22 using findings of the literature review and consultation among the authors to assign

a rating to the average interest, influence and position of key HSR stakeholders, and impact of HSR implementation on them. Our findings for all dimensions combined were used to assess HSR’s overall achievement of its objectives, and to make recommendations on where its public health impact could be improved.

Results

We identified 33 relevant peer-reviewed publications, 21 government-commissioned and three independent reports, most of which contained quantitative data relevant to one of more of the RE-AIM dimensions of Reach, Efficacy and Implementation: see Supplementary Appendix 1. Adoption and Maintenance were primarily assessed through information provided by the Australian and New Zealand governments through websites and communiqués, facilitating analysis of stakeholder engagement in HSR’s current operation, governance and funding.

ReachReach was assessed by the proportion of the population that were aware of HSR, trust it, and had been exposed to the education campaign.

HSR awareness had been evaluated in nine nationally representative surveys in Australia and three in New Zealand.23-34 They suggested low, but consistently improving, unprompted awareness (3% April 2015,

to 21% July 2018), and steadily increasing prompted awareness of the HSR system (33% April 2015 to 84% July 2018) (Figure 1). Females, younger people, those with higher education, higher income and normal weight were consistently more likely to be aware of HSR.

In Australia, these surveys showed that trust in HSR among the total population had steadily increased from 38% in April 2015 to 61% in July 201823-31 (Figure 1). In New Zealand, trust was 39% in January 2017,33 and steady at 40% in June 2018.34

Ten of these surveys evaluated exposure to the education campaign. Australia’s campaign ran over four waves between 2014 and 2017 with eight surveys conducted up until July 2018 showing campaign recognition fluctuating between 13 and 25% (Figure 1).24-31 Evaluators noted funding was ‘modest’ compared to other government and private sector campaigns.25 In New Zealand, reported recognition rose from 12% in December 2016 to 45% in June 2018 following addition of television to the marketing mix.33,34

EfficacyEfficacy was assessed by the extent to which HSR was guiding consumers towards healthier choices. HSR’s efficacy had been the subject of more than 29 peer-reviewed research papers and 15 government-commissioned reports covering performance of one or more of the HSR’s three components: the label graphic, underlying

Figure 1: Unprompted awareness of HSR, prompted awareness of HSR, trust in HSR and exposure to the HSR campaign in Australia.

Figure 1: Unprompted awareness of HSR, prompted awareness of HSR, trust in HSR and exposure to the HSR campaign in Australia

Data is provided from the date of the first availability in Australia (April 2015).

0

20

40

60

80

100

Apr‐15 Oct‐15 Apr‐16 Oct‐16 Apr‐17 Oct‐17 Apr‐18

% Australian po

pulatio

n

Unprompted awareness Prompted Awareness Trust Have seen campaign

Note:Data is provided from the date of the first availability in Australia (April 2015).


algorithm, and accompanying education campaign.

Twenty-six papers and reports had assessed the efficacy of the HSR graphic, including two randomised controlled trials (RCTs), seven randomised choice experiments, eleven nationally representative surveys, three intercept surveys, two focus group studies, and one policy analysis paper.

HSR ‘star’ graphic

Understanding and use: Most research identified the HSR graphic as easy to understand and use. The HSR star logo was found to be more likely to be understood and to influence product selection than the Nutrition Information Panel (NIP),35,36 health and nutrient content claims,37-40 and alternative FoPL designs including the Multiple Traffic Light (MTL)35,36,38,41 and industry-preferred Daily Intake Guide.35,36,38,41-44 Several studies confirmed these results in children.40-42,44 These experimental findings were consistent with government-commissioned monitoring surveys, where between two-thirds and three-quarters of consumers consistently self-reported HSR was easy to understand and use.23-34

Two New Zealand studies (one intercept survey and one online experiment) produced disparate findings45,46 though both were conducted shortly after HSR adoption, using label designs different from the HSR graphic used in practice. In 2018, the intercept survey was repeated with an updated label, producing results more consistent with other findings on consumer understanding and use.47

No experiments had assessed use and understanding of HSR’s ‘energy icon only’ variant of the label, which displayed only kilojoule (and not star rating) information. In government surveys, only 1% of consumers found it easy to understand and use.26

Choice and purchasing: Consumers consistently self-reported being influenced by HSR when shopping23-31,33,34,37,48 but studies assessing HSR’s impact on choice and real world purchases were less clear.

Several studies inferred a shift towards purchasing of more healthy food or beverage choices when compared to no FoPL,43,49 and suggested that HSR remained a significant attribute in driving product choice even when there were co-existing health claims39,40 or other forms of nutrition information and marketing on the label.50 In the disparate New

Zealand studies noted above, HSR was shown to be ineffective in influencing unprompted consumer choice between two breakfast cereals46,47 and consumers made similar purchases using HSR and MTL.45 Randomised controlled trials examining the impact of FoPL in the real world identified no effect of HSR on the healthiness of food purchases,35,36 despite participants’ stated preference for the HSR label format.

Ability to incentivise reformulation: Several companies reported HSR was guiding reformulation activities51 but only two papers systematically assessed HSR’s impact on reformulation across the food supply.52,53 Research in New Zealand found small but statistically significant favourable changes in mean energy density, sodium and fibre in HSR labelled products compared with their composition prior to adoption of HSR.52 In Australia, these methods were replicated and used to model cost-effectiveness, with researchers determining HSR a cost-effective strategy for delivering food reformulation under both voluntary and mandatory implementation scenarios.53

HSR algorithm

Thirteen peer-reviewed publications and two government reports assessed the performance of the HSR algorithm using different validation methods.54,55

Alignment with current scientific literature (i.e. content validity): Food components included in the algorithm were largely consistent with those in government-led FoPL elsewhere.56-58 HSR’s components ‘to limit’ were the four most common elements reported in FoPL globally: energy (used in 41% of systems), sodium (43%), saturated fat (35%) and total sugars (41%).56 Not all FoPL contained ‘positive’ components. Those used in HSR (FVNL, fibre and protein) were used in several other FoPL elsewhere,57 but lack of transparency in FVNL and fibre values relied upon to calculate HSR, and changes to the ‘tipping point’ for determining eligibility to receive protein points were raised by public health and consumer stakeholders as concerns in the five-year review.59

Two papers focused on incorporating added or free sugars into HSR60,61 to accord with evidence-based recommendations of Australian and New Zealand food-based dietary guidelines and updated WHO Guidelines on Sugars Intake.62 A 2017 audit suggested added or free sugars were included in 14% of FoPL globally.56 Treatment

of sugars was being considered in the five-year review.59

Alignment with existing policies and other measures of healthiness (i.e. construct validity): We identified eight studies that assessed alignment between HSR and the Australian Dietary Guidelines (ADGs). This work consistently found that healthy ‘core’ or ‘Five Food Group’ (FFG) foods received higher HSRs on average (HSR 3.7-4.0) than ‘discretionary’ foods (HSR 1.9-2.5).61,63-68

The two papers focused on added sugars demonstrated that alignment with the ADGs could be improved by incorporating added sugars into the HSR algorithm.60,61

Three papers and one government report attempted to specify overall alignment with the ADGs. Two large cross-sectional examinations of the food supply calculating HSRs for all products (n=34,000; 65,600) regardless of whether they displayed HSR, found between 82-87% of products had HSRs corresponding with a pre-defined ‘appropriate’ range for core or discretionary using a cut-point of HSR 3.5 (i.e. core foods scored equal or above this and discretionary foods below this).63,65 Two smaller studies (n=1,269; 3,940) reviewed the algorithm using information from labels on which HSR was displayed. The findings of these studies highlighted that between 39-57% of discretionary foods displayed a HSR≥2.5, assessed by the study authors as an unacceptable ‘pass’ mark.64,67 Each of these works highlighted HSR ‘outliers’, attributed in some cases to the algorithm and in others to imprecise definitions of unhealthy food.63 Recommendations made for improving the algorithm including its treatment of sugar, protein, juices, and unpackaged fruits and vegetables were being considered in HSR’s five-year review.59-61

HSR alignment with Australia and New Zealand’s existing health claims legislation was found to be good at a cut-point of HSR≥3.5; with 97.3% of products over this threshold eligible to display a health claim.69

While HSR was explicitly designed to focus on packaged and processed foods,11 there is increasing international interest in the impact of industrial food processing on health, particularly the association between high levels of consumption of ultra-processed foods (UPF) and poor diets.70 Three papers assessed HSR against the NOVA food classification system.67,71,72 In a sample of dairy foods, HSR correctly classified milks, but not yoghurt and cheeses, based on degree



of processing.72 In a sample of supermarket own-brand foods voluntarily displaying HSR (n=3,940), unprocessed and minimally processed foods had a higher mean (HSR 4.4) than processed (HSR 3.5) or UPF (HSR 2.5), however, 55% of UPF displayed a HSR≥2.5, assessed by the authors as a failure to fall below a designated cut-off of HSR 2.0.67 A summary of submissions to the five-year review acknowledged stakeholder comments on degree of processing,73 but at the time of writing the independent reviewer had determined it was outside the reasonable scope of the review.59

Alignment with health outcomes (i.e. predictive validity): No papers were identified that assessed the ability of the HSR algorithm to predict health outcomes, reported as the strongest method for assessing the validity of nutrient profile models.54,72

HSR Campaign

Eight government-commissioned surveys in Australia and two in New

Zealand evaluated the performance of the HSR campaign.24-31,33,34 The majority of respondents reported that they understood campaign messages, though Australian evaluation noted some persistent confusion.24,25,27 In both countries, those who had seen the campaign self-reported higher awareness, trust, understanding and use of the HSR, and consistently reported carrying out at least one behavioural objective of the campaign with around two-thirds self-reporting they had purchased a new product because of its HSR.24,25,27,33,34

AdoptionAdoption was measured as the degree to which the necessary stakeholders engaged in HSR implementation. Available data were used to map HSR governance structures (Figure 2) and summarise involvement of each stakeholder (Table 2). Stakeholder analysis was conducted through iterative consultation among the authors, assessing the average level of interest, influence

and position of each stakeholder in HSR implementation, and HSR implementation’s resulting impact on them (Table 2).

Key stakeholders with high interest and a supportive position included the Australian (Commonwealth, State and Territory) and New Zealand governments, each of whom contributed funding and together retained ultimate decision-making power on the future of the system through voting rights exercised in the Ministerial Forum on Food Regulation (Forum). Decisions by politicians in the Forum are supported by the work of senior government officials in the Food Regulation Standing Committee (FRSC). In New South Wales, State Government integrated HSR into its food procurement criteria in schools and hospitals.74 Despite this formal influence, media analysis up to 2016 noted government representatives rarely participated in public commentary on HSR implementation.75

Food manufacturers and retailers have high interest and influence, though their

Table 2: Assessment of average interest, influence and position of stakeholders involved in HSR implementation, and impact of HSR on them.Stakeholder Characteristics

Involvement in the issue Interest in HSR Influence/Power Position* Impact of HSR on stakeholder

Australian Commonwealth Government

Participate in Trans-Tasman Food Regulatory Committees with remit over HSR

Host FoPL Secretariat – primary public point of contact

Facilitate government coordination – e.g. chair Jurisdictional Group and TAG

Run Australian education and awareness campaign

Administer tender for HSR monitoring and evaluation in Australia and overall

Contribute funding to support implementation

High High Supportive Medium

New Zealand Government Participate in Trans-Tasman Food Regulatory Committees with remit over HSR

Contribute funding to support implementation, including NZ specific campaign

Coordinate and manage NZ HSRAG

Coordinate and collate NZ monitoring and evaluation

High Medium Supportive Medium

Australian State and Territory Governments

Participate in Trans-Tasman Food Regulatory Committees with remit over HSR

Selected representatives on HSRAC and TAG

Contribute funding to support implementation, including campaign

Consider integration of HSR into State-based policies e.g. school canteen guidelines

Medium Medium Supportive Low

Food manufacturers and retailers Formal representation on HSRAC,TAG and NZ HSRAG

Responsibility to voluntarily apply HSR on products

Provide in-store placement of HSR campaign materials (retailers)

High High Somewhat supportive

High

Public health community Formal representation on HSRAC, TAG and NZ HSRAG

Conduct and publish research on HSR efficacy and implementation

Build awareness of HSR among peers, patients and public in Australia and globally

Advocate for improvements to HSR to improve public health impact

Medium Medium Somewhat supportive**

Medium

Consumer groups Formal representation on HSRAC and NZ HSRAG

Conduct and publish consumer research on HSR efficacy and implementation

Build awareness of HSR with consumers and consumer organisations globally

Advocate for improvements to HSR to improve consumer utility

Medium Medium Somewhat supportive

Medium

Notes:Key to abbreviations used: HSR, Health Star Rating; FoPL, Front-of-Pack Label; NZ, New Zealand; TAG, Technical Advisory Group; NZ HSRAG New Zealand Health Star Rating Advisory Group; HSRAC, Health Star Rating Advisory Committee*Possible values for position include: supportive, somewhat supportive, somewhat opposed and opposed**While the majority of papers, policy statements, submissions and media representations from this group were generally supportive, a small number of vocal opponents were noted



Figure 2: Health Star Rating system governance. Figure 2: Health Star Rating System Governance

Australia and New Zealand Ministerial Forum on Food Regulation (Forum) Membership: Federal, state and territory Ministers responsible for food from Australia and New Zealand Chaired by Australian Government Minister for Rural Health Responsibilities: Develop domestic food regulatory policy and policy guidelines for setting domestic food standards Ultimate authority to make HSR decisions where no HSRAC consensus; ultimate authority to decide whether HSR made mandatory Decision by consensus where possible, otherwise by vote with six votes required for a decision

Food Regulation Standing Committee (FRSC) – A Forum subcommittee Membership: Senior officials of departments for which Ministers represented on the Forum have portfolio responsibility Responsibilities: Coordinate policy advice to Forum, ensure nationally consistent approach to implementation and enforcement of food standards Absorbs work of previous multi-stakeholder FoPL Steering Committee that developed HSR

Front-of-Pack Labelling Secretariat (Secretariat)

Membership: Commonwealth Department of

Health employees Responsibilities: Public contact point for HSR Maintain HSR website and

newsletter Refer matters for interpretation to

HSRAC Facilitate jurisdictional coordination Administer tender for monitoring and

evaluation

Trans-Tasman HSR Advisory Committee (HSRAC) Membership: Nine Australian representatives: government (3), industry and retailers (3), public health and

consumer groups (3); one New Zealand representative: chair of NZ HSRAG Responsibilities Oversee voluntary implementation, including social marketing and monitoring and evaluation Assess potential anomalies identified within the HSR algorithm Provide advice to FRSC on implementation Foster ongoing collaboration between stakeholders Decision making by consensus, otherwise referral to FRSC and Forum

Technical Advisory Group (TAG) Membership: Government (4), industry (2), public health

(2) Chaired by Commonwealth Department of

Health Responsibilities: Analyse and review performance of HSR

calculator and algorithm as directed by HSRAC, using data provided by industry

Provide evidence to support consideration of options for the five year review (no recommendations)

New Zealand HSR Advisory Group (NZ HSRAG) Membership: Government, (2),

industry (3) public health (3), consumer groups (1), independent food consultant (1)

Responsibilities Support voluntary

implementation in NZ

Independent Reviewer (MP Consulting) Membership: Policy evaluation experts Responsibilities Conduct multi-stakeholder

consultations Review modelling by TAG Produce formal five year

review report considered by HSRAC, FRSC, Forum

New Zealand Ministry of Primary Industries (MPI) Membership: New Zealand

Ministry of Primary Industries employees (unknown number)

Responsibilities: Administer and

monitor HSR implementation in NZ

Commissioned monitoring providers Report to relevant advisory committees by agreed timelines and frameworks Heart Foundation (use, understanding and uptake, AUS) Pollinate (campaign evaluation, AUS) Colmar Brunton (campaign evaluation, NZ) National Institute of Health Innovation, (uptake, NZ) Isentia (Media analysis, AUS)

Jurisdictional Group Membership: Representatives from state and

territory governments Led by FoPL Secretariat Responsibilities: Facilitate information sharing

between jurisdictions Brief members on issues being

considered by HSRAC

Key:

Government: Political appointees

Government: Public sector employees Multi-stakeholder body

Commissioned service provider

Reporting line as specfied in box text

Note: The Two Year Report prepared by HSRAC refers to additional committees: a multi-stakeholder ‘Social Marketing and Advisory Group’ (SMAG) providing feedback and guidance on the education campaign; and a Monitoring and Evaluation Reference Network (MERN) providing opportunity discussion between government jurisdictions and monitoring organisations. As no further information is publicly available on these groups, they have not been included in this diagram.



participation on HSR governance committees was limited to those below the governmental Forum and FRSC. In HSR’s voluntary form, their power comes primarily from agreeing to apply HSR to product labels. Given the potential business impact of mandatory labelling requirements, peak industry bodies had indicated their support for HSR, conditional on it remaining voluntary.76 Media analysis found industry the most frequently cited stakeholder conveying favourable messages about HSR, including that it helps promote products, drives innovation and that industry were acting to introduce it.75

Health and consumer groups participated in HSR governance committees. On the trans-Tasman Health Star Rating Advisory Committee (HSRAC) they had combined numbers equal to industry. They influenced and supported implementation by conducting independent research and disseminating information to their own networks, interest groups and the wider public. Common messages conveyed by these groups in media analysis were that HSR could be an effective tool to communicate with consumers, but also that it was being used by industry in ways that favoured their own interests.75 In formal consultations and policy statements, health and consumer groups broadly indicated their support for HSR while advocating for it to be strengthened, made mandatory and complemented with other nutrition policies.73,77-82

ImplementationImplementation was measured by the extent to which HSR was appearing on labels as intended, assessed by both commissioned monitoring and independent publications involving cross-sectional examination of the food supply.

Uptake

Uptake had been examined in five peer-reviewed publications64,67,71,83,84 and six government-commissioned reports23,31,85-88 covering both Australia and New Zealand. Results indicated uptake was increasing, with government issuing a communiqué in June 2018 that HSR had been displayed on 10,333 products in Australia and over 3,900 in New Zealand.89 Studies that examined proportionate uptake suggested HSR was on between 20-28% of eligible products in the Australian food supply in 2017.31,83,87 Uptake remained lower in New Zealand, at

20.9% in 2018.88 Only ten per cent of new products entering the Australian market between June 2014 and June 2017 chose to display HSR.64 In Australia, more than 118 manufacturers were using HSR in 2018, but large retailers Coles, Woolworths and Aldi were together responsible for more than half of all uptake.31,83 Uptake was skewed towards products that scored at the upper end of the five-star spectrum.31,67,83,85,88 More than 50% of uptake was on UPF foods.67,71

Compliance

Government-commissioned monitoring suggested at least 90% of HSR labels complied with Style Guide formatting requirements, with errors predominantly of a minor technical nature23,31,86 or related to poor legibility.85 Official monitoring also suggested >90% accuracy of HSR values, with incorrect calculations more frequently under-reporting, rather than over-stating HSR, usually by 0.5 stars.23,31,87 Independent research raised issues concerning inconsistent use of the ‘energy icon only’ variant of HSR, particularly on low-scoring non-dairy beverages.83,84

MaintenanceMaintenance was assessed by measures taken by stakeholders to sustain HSR over time. Data available directly from government websites detailed HSR governance structures, complaints mechanisms, frameworks for monitoring and evaluation, and funding committed.

Governance structures

Figure 2 illustrates the governance of HSR in its voluntary status as at October 2018. The trans-Tasman government bodies of the Forum and FRSC retained ultimate decision-making power on the operation and continuance of HSR. Underneath this, implementation was overseen by the HSRAC, whose remit was to foster ongoing collaboration between government, industry, public health and consumer groups. HSRAC coordinated the HSR education campaign, as well as monitoring and evaluation of the system, reporting outcomes to the Forum and FRSC. HSRAC also received matters submitted through HSR complaint mechanisms for decision making by consensus. Where consensus could not be reached, matters were referred to the Forum and FRSC.

Ancillary support was provided by the FoPL Secretariat (Secretariat) in the Australian

Commonwealth Department of Health. The Secretariat acted as public contact point, maintaining the website and newsletter. They also led a Jurisdictional Group, facilitating information sharing on HSR between Australia’s states and territories. In New Zealand, HSR implementation was administered by the Ministry of Primary Industries (MPI), who received advice from their own multi-stakeholder Health Star Rating Advisory Group (HSRAG).

Legal analysis suggested the Australian Commonwealth Government possessed the requisite authority to make HSR mandatory if desired.90

Complaint mechanisms

Potential algorithm anomalies can be submitted to HSRAC for consideration; by October 2018 there had been 21 submissions, two of which (tinned vegetables and dairy desserts) were determined to meet the specific definition of ‘anomaly’, warranting follow up action.11 An additional dispute resolution procedure exists for challenging HSRs on individual products, though to date no disputes appear to have been registered.11 Outside these processes, HSRAC has dealt with concerns surrounding HSR implementation in an ad hoc manner. For example, ‘the form of the food – as prepared’ rules in the HSR Style Guide were subject to a formal public consultation, additional modelling and additional industry proposals before ultimate referral to the Forum for resolution. The process took more than 18 months, with compliance not required by industry until after 2019.11

Monitoring and evaluation framework

Conduct of monitoring in Australia was tendered to the National Heart Foundation shortly after implementation.91 It included regular reports on consumer awareness and use, as well as label implementation, consistency, and nutrient status of products carrying HSR.11 Similar activities occurred in New Zealand, coordinated by MPI with input from academic research organisations.85,88 Regular monitoring of uptake and use was supplemented with commissioned evaluation of the education campaign24,25,27,33,34 and HSR coverage in media.75

In 2016, HSRAC issued a combined two-year monitoring report compiling data from this work.51 Following this, planning commenced for a formal five-year review. An independent reviewer (MP Consulting) was appointed by



tender,59 and a voluntary multi-stakeholder Technical Advisory Group (TAG) created with specific remit to analyse performance of the HSR algorithm and provide technical input.11 The review involved several rounds of written and face-to-face consultation. Feedback consolidated and reported online noted the main concerns raised, namely that some products high in sugar, fat and salt could carry a high rating.73 Results of TAG modelling attempted to provide solutions and were published online with a calculator to test the implications of preferred options on products.92 Recommendations on long-term maintenance of HSR, including whether the system should be made mandatory90 were to be provided in a report for consideration first by HSRAC and then FRSC to inform a decision by Forum Ministers on the future of HSR in mid-2019.11

Funding commitments

Comprehensive information on HSR funding was difficult to obtain due to its federated, trans-Tasman structure. Budget documents recorded $5.3 million committed by Australia’s Commonwealth government to HSR for the period 2016-2019, noting continued involvement and endorsement of government was critical to HSR’s independence.93 Tender databases suggested it distributed about $2 million on monitoring and evaluation services, and about $2.3 million up to October 2018 on campaign development and evaluation.91 This did not

include costs of media buy: in 2017, phase four of the campaign alone had a media buy of $2.2 million.91 Specific contributions or spending by state and territory governments or by New Zealand for HSR related activities were not publicly available and were not included in this sum. It is not clear what resources would be made available for sustaining HSR after delivery of the review report in 2019.

Conclusions

More than four years since voluntary implementation commenced, a significant body of evidence supports continuation and strengthening of HSR. Our systematic analysis points to key areas where HSR’s public health impact can be enhanced (see Box 1).

Awareness and trust were reported as increasing, though unprompted awareness remained modest given HSR’s position as a key pillar of both countries’ responses to addressing the huge burden of diet-related disease. Lower awareness among Australians who were overweight, live in rural areas or experience socioeconomic disadvantage suggests opportunity to improve HSR’s utility among these groups. Successful targeted efforts in New Zealand with ‘priority’ groups suggest similar attention in Australia would be important to address ongoing health inequities.

Exposure to the HSR campaign remained disappointing. While evaluators suggested

campaign funding was ‘modest’, it made up a significant proportion of total spend on HSR. Monitoring suggested most people were aware of HSR from ‘seeing it on pack’, making it arguably more cost-effective for government to focus on increasing HSR uptake, rather than further spending on awareness campaigns.

The bulk of peer-reviewed and government-commissioned research focused on HSR’s efficacy. The ‘star’ graphic was shown to be well-liked by consumers, and superior in utility to the industry-preferred DIG. To maximise the utility of a single FoPL, the DIG and its variants (i.e. Treatwise, energy icon variant of HSR) should now be formally retired. Innovation in FoPL formats worldwide suggest opportunities for strengthening HSR’s graphic design further. Evidence-based features to enhance visibility and consumer utility such as incorporation of colour (for example, France’s Nutriscore94 or the MTL), written government endorsement (as in Chile95 and Singapore96) and Canada’s proposed rules for positioning FoPL in a uniform pack position away from health claims97 provide inspiration for future research and updates to the HSR Style Guide (Box 1).

HSR’s efficacy also depends on its underlying algorithm providing an accurate representation of the healthiness of food. Substantial attention has been placed on the performance of the HSR algorithm, predominantly through content and construct validity assessments that show its similarities with other nutrient profiling algorithms and tend to support its performance as a reasonable, albeit imperfect, tool to assess nutritional quality. Differences in methodologies and ‘cut-points’ have led to variations in results that highlight challenges in assessing alignment with other measures of healthiness without pre-defined indicators by which to measure ‘success’, e.g. a HSR threshold or band of scores appropriate to delineate ‘healthy’ from ‘unhealthy’ or minimally processed from ultra-processed foods. Despite these differences, broadly consistent recommendations have emerged for strengthening algorithm alignment with existing health policies (Box 1).

Our assessment also highlighted that the HSR algorithm has not been subject to more robust forms of validation. HSR is not unique in this respect: a recent systematic review found only 10% of nutrient profile models being used in government-led nutrition

Box 1: Recommendations for improving HSR’s public health impact.Reasonable refinements to improve efficacy• Strengthen utility of the ‘star’ graphic by considering standardised colour, size and placement, specifying separation from health

claims, ending concurrent use of non-interpretive labels (e.g. Daily Intake Guide, Treatwise, ‘energy icon only’ variant)• Implement HSR algorithm improvements to reflect findings of existing research: incorporate added sugars, strengthen treatment

of sodium, review treatment of protein, consider treatment of fresh fruit and vegetables including unpackaged• Conduct further high level validation studies to explore link between the HSR of foods, healthier diets, and health outcomes Responsive regulatory action to improve uptake• Clear targets with specified timelines (e.g. 80% eligible products within two years of 2019 review completion) and commitment

by Forum to make mandatory on specified date where sufficient progress not demonstrated• Improve transparency and accountability of uptake monitoring through use of regularly updated, publicly available branded food

composition database Strengthen government leadership to improve HSR governance• Renewed and unambiguous public commitment and funding to continue HSR beyond five year review• Increased public visibility of government leadership at ministerial level • Authority and resource delegated to FSANZ to provide independent technical advice• Renewed Terms of Reference for multi-stakeholder involvement, controlling for conflicts of interest, particularly in technical

functions such as algorithm review and determining anomalies• Improve transparency of multi-stakeholder committees and public consultations, e.g. agendas and minutes, individual

submissions publicly available• Reform complaint mechanisms to improve utility, provide expeditious resolution of reasonable concerns raised by all stakeholders,

including consumers• Integrate HSR into other government-led nutrition policies e.g. procurement for public settings, criteria for marketing to children,

fast food menu labelling• Situate and support HSR within a comprehensive policy framework e.g. National Obesity or Nutrition Strategy



policies have been subject to ‘predictive’ validity testing to assess associations with health outcomes, e.g. weight gain or cancer risk.58 The most frequently validated of these is the United Kingdom (UK) Ofcom model, from which HSR originated.55,58 Results of studies assessing its performance in UK and French cohorts have found prospective associations with health outcomes in most,98-

103 but not all studies.104 While recognising the significant commonalities between both algorithms, further high-level validation could usefully assess any prospective association between HSR, healthier diets and health outcomes in Australasian populations. It could also assess whether variations in HSR’s design (e.g. creation of extra dairy categories) have impacted these associations.

While refinements to increase efficacy are important, analysis of implementation suggests they are unlikely to drive improvements in impact unless accompanied by radically increased uptake. During its development in 2013, Forum Ministers agreed HSR would remain voluntary subject to there being ‘consistent and widespread’ uptake, otherwise a mandatory approach would be required.105 Even without performance indicators, it is arguable that uptake of less than one-third of eligible products (mostly those that score well), justifies review of HSR’s voluntary status. Mandatory FoPL are increasing globally, including recent initiatives in Mexico, Iran, Chile, Sri Lanka, Peru, Uruguay and Israel.106 Our findings highlight wide support from consumer and public health stakeholders, but not industry, to make HSR mandatory in Australasia. If a mandatory HSR is not yet politically feasible, a ‘responsive regulatory’ approach provides interim suggestions for how uptake must necessarily be enhanced to improve HSR’s utility to consumers107 (Box 1).

The relative engagement and differential power held by HSR stakeholders (Table 2 and Figure 2) provide insights into how HSR’s governance can be enhanced. While HSR is a multi-stakeholder initiative, government retains ultimate responsibility for HSR decision-making and funding. Absence of government comment in media analysis suggests opportunity for greater visible leadership from government Ministers, for example, in reiterating government endorsement of HSR and communicating positive changes for consumers emanating from the five-year review.

Low uptake by industry (particularly on less healthy products), despite their public endorsement of HSR supports review of the Terms of Reference for their engagement. This should take into account increasing global awareness of the need to prevent and manage conflicts of interest in the development of national nutrition policies.108 Notably absent from governance arrangements outlined in Figure 2 are Food Standards Australia New Zealand (FSANZ) who have the expertise and independence to conduct many of the functions performed voluntarily by HSRAC and the TAG to date. While a renewed HSRAC may have a role in continuing to promote multi-stakeholder collaboration in implementation, delegation of greater technical authority to FSANZ to administer and validate the algorithm, monitor uptake, and assess compliance using publicly available branded food composition data, could mitigate real or perceived commercial conflicts of interest in HSR’s governance and facilitate its progressive integration into the formal food regulatory system.

Linkages between HSR and other health policies, as done by NSW in procurement standards, or by countries like Chile in linking FoPL with restrictions on marketing to children,95 provide opportunities to further the utility of HSR. Strategically situating and supporting HSR within a comprehensive policy framework such as a National Obesity Strategy will enhance synergies with existing and future interventions to address diet-related disease.

This paper used a systematic approach to synthesising a growing body of heterogeneous material on HSR’s implementation and efficacy. The strength of the evidence obtained is importantly limited in several areas by study design, the scope of the analyses done and the magnitude of the projects completed. Further investment in high-quality research will provide better insight into the most likely effects of HSR on health outcomes, and how best to maximise them through both technical enhancements and improvements in implementation. Analysis of industry compliance with the HSR algorithm was limited by lack of transparency surrounding some food components (e.g. benefits obtained from Fruit, Vegetable Nut and Legume (FVNL) content) as companies are not required to display the relevant data on the label. Our governance assessment was to some degree limited by reliance on public

information, e.g. no available minutes of HSRAC or TAG meetings.

Implications for public health

Adoption of HSR in 2014 placed Australia and New Zealand among a small but growing number of countries using FoPL as one tool to promote healthier diets. Four years since implementation commenced, available evidence supports the continuation and strengthening of HSR.

As the formal five-year review draws to a close in 2019, reasonable refinements to HSR’s star graphic and algorithm, action to initiate mandatory implementation and strengthened governance – particularly through renewed, visible government leadership – present the clearest opportunities to enhance HSR’s public health impact.

References 1. Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K,

Adair-Rohani H, et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: A systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2224-60.

2. Stuckler D, McKee M, Ebrahim S, Basu S. Manufacturing epidemics: The role of global producers in increased consumption of unhealthy commodities including processed foods, alcohol, and tobacco. PLoS Med. 2012;9(6):e1001235.

3. PricewaterhouseCoopers. Weighing the Cost of Obesity: A Case for Action. Sydney (AUST): PWC; 2015.

4. World Health Organization. ‘Best Buys’ and Other Recommended Interventions for the Prevention and Control of Noncommunicable Diseases, Updated (2017) Appendix 3 of the Global Action Plan for the Prevention and Control of Non-Communicable Diseases 2013-2020. Geneva (CHE): WHO; 2017.

5. World Health Organization. Implementation Plan of the WHO Commission on Ending Childhood Obesity. Geneva (CHE): WHO; 2017.

6. Volkova E, Mhurchu CN. The influence of nutrition labeling and point-of-purchase information on food behaviours. Curr Obes Rep. 2015;4(1):19-29.

7. Vyth EL, Steenhuis IH, Roodenburg AJ, Brug J, Seidell JC. Front-of-pack nutrition label stimulates healthier product development: A quantitative analysis. Int J Behav Nutr Phys Act. 2010;7(1):1.

8. World Health Organization. Nutrient Profiling. Geneva (CHE): WHO; 2010.

9. World Cancer Research Fund International. WCRF International Food Policy Framework for Healthy Diets: NOURISHING. London (UK): WCRF; 2019.

10. Codex Alimentarius Commission Committee on Food Labelling. Report of the Codex Committee On Food Labelling (Forty-fourth Session); 2017 Oct 16-17; Ascunsion, Paraguay. Rome, Itlay: Food and Agriculture Organisation of the United Nations; 2018.

11. Australian Government Department of Health. The Health Star Rating System [Internet]. Canberra (AUST): Front-of-Pack Labelling Secretariat; 2018 [cited 2019 May 2]. Available from: http://healthstarrating.gov.au/

12. Front-of-Pack Labelling Project Committee. Objectives and Principles for the Development of a Front-of-pack Labelling (FOPL) System Endorsed 11 May2012. Canberra (AUST): Australian Government Department of Health; 2018.



13. Australia New Zealand Food Regulation Ministerial Council. Front of Pack Labelling Policy Statement. Canberra (AUST): Food Regulation Secretariat Australian Government; 2009.

14. Food Standards Australia New Zealand Act 1991 (Cth). Australia New Zealand Food Standards Code – Standard 1.2.7 – Nutrition, Health and related Claims.

15. Australian Government Food Regulation Secretariat. Food Regulation Activities. Front-of-Pack Labelling Committee and Working Group Meetings [Internet]. Canberra (AUST): Australian Government Department of Health; 2016 [cited 2016 Nov 21]. Available from: http://foodregulation.gov.au/internet/fr/publishing.nsf/Content/frontofpackcommittee

16. Health Star Rating Technical Advisory Group. Five Year Review of the HSR System - Technical Paper, History and Development of the Health Star Rating Algorithm. Canberra (AUST): Front-of-Pack Labelling Secretariat Australian Government Department of Health; 2018.

17. Health Star Rating Advisory Committee. Health Star Rating System Style Guide. 4th ed. Version 5. Canberra (AUST): Front-of-Pack Labelling Secretariat Australian Government Department of Health; 2017.

18. Australian Department of Health. Media Release: Health Star Rating System on Track. Canberra (AUST): Front-of-Pack Labelling Secretariat; 2014 June 27.

19. Australia and New Zealand Ministerial Forum on Food Regulation. Media Release: Communiqué of Outcomes from the Australia and New Zealand Ministerial Forum on Food Regulation Meeting held on 20 November 2015. Canberra (AUST): Front-of-Pack Labelling Secretariat Australian Government Department of Health; 2015 November 20.

20. Glasgow RE, Vogt TM, Boles SM. Evaluating the public health impact of health promotion interventions: The RE-AIM Framework. Am J Public Health.1999;89(9):1322-7.

21. Jilcott S, Ammerman A, Sommers J, Glasgow REJAoBM. Applying the RE-AIM Framework to assess the public health impact of policy change. Ann Behav Med. 2007;34(2):105-14.

22. Varvasovszky Z, Brugha R. A stakeholder analysis. Health Policy Plan. 2000;15(3):338-45.

23. National Heart Foundation of Australia. Report on the Monitoring of the Implementation of the Health Star Rating System – Year 1. Melbourne (AUST): Heart Foundation; 2015.

24. Pollinate Research. Health Star Rating System Campaign Evaluation Report 2015. Canberra (AUST): Front-of-Pack Labelling Secretariat Australian Government Department of Health; 2015 September.

25. Pollinate Research. Health Star Rating System: Campaign Evaluation Report 2016. Canberra (AUST): Front-of-Pack Labelling Secretariat Australian Government Department of Health; 2016.

26. National Heart Foundation of Australia. Report on Monitoring the Implementation of the Health Star Rating System - Area of Enquiry 2 - Consumer Awareness and Ability to Use the HSR System Correctly. Melbourne (AUST): Heart Foundation; 2016 July.

27. Pollinate Research. Health Star Rating System Campaign Evaluation Report 2017. Canberra (AUST): Front-of-Pack Labelling Secretariat Australian Government Department of Health; 2017 June.

28. National Heart Foundation of Australia. Report on the Monitoring of the Implementation of the Health Star Rating System: Key Findings for Area of Enquiry Two – Consumer Awareness and Ability to Use the Health Star Rating System Correctly – November 2017. Melbourne (AUST): Heart Foundation; 2017.

29. National Heart Foundation of Australia. Report on the Monitoring of the Implementation of the Health Star Rating System: Key Findings for Area of Enquiry 2 - Consumer Awareness and Ability to Use the Health Star Rating System Correctly - April 2018. Melbourne (AUST): Heart Foundation; 2018.

30. National Heart Foundation of Australia. Report on the Monitoring of the Implementation of the Health Star Rating System: Key Findings for Area of Enquiry 2 - Consumer Awareness and Ability to Use the Health Star Rating System Correctly - July 2018. Melbourne (AUST): Heart Foundation; 2018.

31. National Heart Foundation of Australia. Report on the Monitoring of the Implementation of the Health Star Rating System in the First Two Years of Implementation: June 2014 to June 2016. Melbourne (AUST): Heart Foundation; 2017.

32. Colmar Brunton. Health Star Rating Monitoring and Evaluation. Wellington (NZ): Health Promotion Agency; 2016.

33. Colmar Brunton. Health Star Rating Monitoring and Evaluation Year One Follow Up Research Report. Wellington (NZ): Health Promotion Agency; 2017.

34. Colmar Brunton. Health Star Rating Monitoring and Evaluation: Year 2 Follow Up Research Report. Wellington (NZ): Health Promotion Agency; 2018.

35. Neal B, Crino M, Dunford E, Gao A, Greenland R, Li N, et al. Effects of different types of front-of-pack labelling information on the healthiness of food purchases—a randomised controlled trial. Nutrients. 2017;9(12):1284.

36. Ni Mhurchu C, Volkova E, Jiang Y, Eyles H, Michie J, Neal B, et al. Effects of interpretive nutrition labels on consumer food purchases: The Starlight randomized controlled trial. Am J Clin Nutr. 2017;105(3):695-704.

37. Riley MD, Bowen J, Krause D, Jones D, Stonehouse W. A survey of consumer attitude towards nutrition and health statements on food labels in South Australia. Foods Health Dis. 2016;6(12):809-21.

38. Talati Z, Pettigrew S, Dixon H, Neal B, Ball K, Hughes C. Do health claims and front-of-pack labels lead to a positivity bias in unhealthy foods? Nutrients. 2016;8(12):787.

39. Russell CG, Burke PF, Waller DS, Wei E. The impact of front-of-pack marketing attributes versus nutrition and health information on parents’ food choices. Appetite. 2017;116:323-38.

40. Talati Z, Pettigrew S, Hughes C, Dixon H, Kelly B, Ball K, et al. The combined effect of front-of-pack nutrition labels and health claims on consumers’ evaluation of food products. Food Qual Prefer. 2016;53:57-65.

41. Talati Z, Pettigrew S, Ball K, Hughes C, Kelly B, Neal B, et al. The relative ability of different front-of-pack labels to assist consumers discriminate between healthy, moderately healthy, and unhealthy foods. Food Qual Prefer. 2017;59:109-13.

42. Pettigrew S, Talati Z, Miller C, Dixon H, Kelly B, Ball K. The types and aspects of front-of-pack food labelling schemes preferred by adults and children. Appetite. 2017;109:115-23.

43. Talati Z, Pettigrew S, Kelly B, Ball K, Neal B, Dixon H, et al. Can front-of-pack labels influence portion size judgements for unhealthy foods? Public Health Nutr. 2018;21(15):2776-81.

44. Talati Z, Pettigrew S, Kelly B, Ball K, Dixon H, Shilton T. Consumers’ responses to front-of-pack labels that vary by interpretive content. Appetite. 2016;101:205-13.

45. Maubach N, Hoek J, Mather D. Interpretive front-of-pack nutrition labels. Comparing competing recommendations. Appetite. 2014;82:67-77.

46. Hamlin R, McNeill L. Does the Australasian “Health Star Rating” front-of-pack nutritional label system work? Nutrients. 2016;8(6):327.

47. Hamlin R, McNeill LJN. The impact of the Australasian ‘Health Star Rating’, front-of-pack nutritional label, on consumer choice: A longitudinal study. Nutrients. 2018;10(7):906. pii: E906.

48. National Heart Foundation of Australia. Report on the Monitoring of the Implementation of the Health Star Rating System: Key Findings for Area of Enquiry Two – Consumer Awareness and Ability to Use the Health Star Rating System Correctly. Melbourne (AUST): Heart Foundation; 2017.

49. Billich N, Blake MR, Backholer K, Cobcroft M, Li V, Peeters AJA. The effect of sugar-sweetened beverage front-of-pack labels on drink selection, health knowledge and awareness: An online randomised controlled trial. Appetite. 2018;128:233-41.

50. Scrinis G, Parker C. Front‐of‐pack food labeling and the politics of nutritional nudges. Law Policy. 2016;38(3):234-49.

51. Health Star Rating Advisory Committee. Two Year Progress Review Report on the Implementation of the Health Star Rating System - June 2014 - June 2016. Canberra (AUST): Public Health Association Australia; 2017.

52. Mhurchu CN, Eyles H, Choi YH. Effects of a voluntary front-of-pack nutrition labelling system on packaged food reformulation: The Health Star Rating System in New Zealand. Nutrients. 2017;9(8):918.

53. Mantilla Herrera AM, Crino M, Erskine HE, Sacks G, Ananthapavan J, Mhurchu CN, et al. Cost-effectiveness of product reformulation in response to the health star rating food labelling system in Australia. Nutrients. 2018;10(5). pii: E614.

54. Townsend MS. Where is the science? What will it take to show that nutrient profiling systems work? Am J Clin Nutr. 2010;91(4):1109S-15S.

55. Cooper SL, Pelly FE, Lowe JB. Construct and criterion-related validation of nutrient profiling models: A systematic review of the literature. Appetite. 2016;100:26-40.

56. World Health Organization. Global Nutrition Policy Review 2016-2017. Geneva (CHE): WHO; 2018.

57. Kanter R, Vanderlee L, Vandevijvere S. Front-of-package nutrition labelling policy: Global progress and future directions. Public Health Nutr. 2018;21(8):1399-408.

58. Labonté M-È, Poon T, Gladanac B, Ahmed M, Franco-Arellano B, Rayner M, et al. Nutrient profile models with applications in government-led nutrition policies aimed at health promotion and noncommunicable disease prevention: A systematic review. Adv Nutr. 2018;9(6):741-88.

59. MP Consulting. Five Year Review of the Health Star Rating System - Navigation Paper. Canberra (AUST): Front-of-Pack Labelling Secretariat Australian Government Department of Health; 2018.

60. Menday H, Neal B, Wu JHY, Crino M, Baines S, Petersen KS. Use of added sugars instead of total sugars may improve the capacity of the health star rating system to discriminate between core and discretionary foods. J Acad Nutr Diet. 2017;117(12):1921-30.e11.

61. Peters SA, Dunford E, Jones A, Ni Mhurchu C, Crino M, Taylor F, et al. Incorporating added sugar improves the performance of the health star rating front-of-pack labelling system in Australia. Nutrients. 2017;9(7):701.

62. World Health Organization. Guideline: Sugars Intake for Adults and Children. Geneva (CHE): WHO; 2015.

63. Jones A, Rådholm K, Neal B. Defining ‘unhealthy’: A systematic analysis of alignment between the australian dietary guidelines and the Health Star Rating System. Nutrients. 2018;10(4):501.

64. Lawrence M, Dickie S, Woods J. Do nutrient-based front-of-pack labelling schemes support or undermine food-based dietary guideline recommendations? Lessons from the Australian Health Star Rating System. Nutrients. 2018;10(1):32.

65. Dunford E, Cobcroft M, Thomas M, Wu J. Technical Report: Alignment of the NSW Healthy Food Provision Policy with the Health Star Rating System. Sydney (AUST): State Government of New South Wales Ministry of Health; 2015.

66. Carrad AM, Louie JCY, Yeatman HR, Dunford EK, Neal BC, Flood VM. A nutrient profiling assessment of packaged foods using two star-based front-of-pack labels. Public Health Nutr. 2016;19(12):2165-74.

67. Pulker C, Trapp G, Scott J, Pollard CJN. Alignment of supermarket own brand foods’ front-of-pack nutrition labelling with measures of nutritional quality: An Australian perspective. Nutrients. 2018;10(10). pii: E1465.

68. Wellard L, Hughes C, Watson WL. Investigating nutrient profiling and Health Star Ratings on core dairy products in Australia. Public Health Nutr. 2016;19(15):2860-5.

69. Dunford E, Huang L, Peters S, Crino M, Neal B, Ni Mhurchu CJN. Evaluation of alignment between the health claims Nutrient Profiling Scoring Criterion (NPSC) and the Health Star Rating (HSR) nutrient profiling models. Nutrients. 2018;10(8). pii: E1065.

70. Monteiro CA, Cannon G, Moubarac J-C, Levy RB, Louzada MLC, Jaime PC. The UN Decade of Nutrition, the NOVA food classification and the trouble with ultra-processing. Public Health Nutr. 2018;21(1):5-17.

71. Pulker CE, Scott JA, Pollard CMJPhn. Ultra-processed family foods in Australia: nutrition claims, health claims and marketing techniques. Public Health Nutr. 2018;21(1):38-48.



72. Cooper S, Pelly F, Lowe JB. Assessment of the construct validity of the Australian Health Star Rating: A nutrient profiling diagnostic accuracy study. Eur J Clin Nutr. 2017;71(11):1353-1359.

73. MP Consulting. Report on Submissions to the Five Year Review of the Health Star Rating System. Canberra (AUST): Australian Government Department of Health; 2017.

74. NSW Ministry of Health. Healthy Food and Drink in NSW Health Facilities for Staff and Visitors Framework. Sydney (AUST): State Government of New South Wales; 2017.

75. Isentia Insights. Media Analysis Report July 2014 - June 2016 SA Health Star Rating. Adelaide (AUST): State Government South Australia Deaprtment of Health; 2016.

76. Australian Food and Grocery Council. Submission to Senate Select Committee Inquiry into the Obesity Epidemic. Canberra (AUST): AFGC; 2018.

77. Obesity Policy Coalition. Submission to the Senate Select Committee into the Obesity Epidemic in Australia - July 2018. Canberra (AUST): Australian Government Department of the Senate; 2018.

78. Obesity Policy Coalition. Tipping the Scales: Australian Obesity Prevention Consensus. Melbourne (AUST): OPC; 2017.

79. CHOICE. Submission to Select Senate Committee Inquiry into the Obesity Epidemic - July 2018. Sydney (AUST): CHOICE; 2018.

80. Public Health Association of Australia. Submission to the Senate Select Committee Enquiry into the Obesity Epidemic - July 2018. Canberra (AUST): PHAA; 2018.

81. Sacks G for the Food-EPI Australia Project Team. Policies for Tackling Obesity and creating Healthier Food Environments: Scorecard and Priority Recommendations for Australian Governments. Melbourne (AUST): Deakin University; 2017.

82. Vandevijvere S, Mackay S, Swinburn B. Measuring and stimulating progress on implementing widely recommended food environment policies: The New Zealand case study. Health Res Policy Syst. 2018;16(1):3.

83. Jones A, Shahid M, Neal B. Uptake of Australia’s Health Star Rating System. Nutrients. 2018;10(8):997.

84. Brownbill AL, Braunack‐Mayer A, Miller CJHPJoA. Health star ratings: What’s on the labels of Australian beverages? Health Promot J Austr. 2019;30(1):114-18.

85. NZ Ministry for Primary Industries. Health Star Rating: Monitoring Implementation at Year Two. MPI Technical Paper No.: 2017/09. Wellington (NZ): Government of New Zealand; 2017.

86. National Heart Foundation of Australia. Report on the Monitoring of the Implementation of the Health Star Rating System: Key Findings for AOE1 - Consistency in Implementation of the Health Star Rating (HSR) System with the HSR Style Guide – June 2016 to June 2017. Melbourne (AUST): NHF; 2017.

87. National Heart Foundation of Australia. Report on the Monitoring of the Implementation of the Health Star Rating System: Key Findings for AOE1 – Assessment of the Health Star Rating (HSR) Displayed on Pack Using the HSR Calculator – June 2016 to June 2017. Melbourne (AUST): NHF; 2017.

88. NZ Ministry for Primary Industries. The Health Star Rating System in New Zealand 2014-2018 System Uptake and Nutrient Content of Foods by Health Star Rating Status. Wellington (NZ): Government of New Zealand; 2018.

89. Australia New Zealand Ministerial Forum on Food Regulation. Communique of Outcomes from the Ministerial Forum on Food Regulation 29 June 2018. Canberra (AUST): FSANZ; 2018.

90. Brennan M. Is the Health Star Rating System a Thin Response to a Fat Problem: An Examination of the Constitutionality of a Mandatory Front Package Labelling System. Notre Dame Law Rev. 2015;17(5):86.

91. AusTender. Australian Government’s Procurement Information System. Canberra (AUST): Government of Australia; 2018.

92. Matthews Pegg Consulting. Matthews Pegg Consulting: HSR System Enhancements. Canberra (AUST ): MPConsulting; 2018.

93. Australian Department of Health. Budget 2016-2017. Canberra (AUST): Government of Australia; 2016.

94. Santé Publique France. Nutri-Score Corporate Graphic Charter. Saint-Maurice Cedex (FRA): Public Health France; 2018.

95. Corvalán C, Reyes M, Garmendia ML, Uauy R. Structural responses to the obesity and non‐communicable diseases epidemic: The Chilean Law of Food Labeling and Advertising. Obes Rev. 2013;14:79-87.

96. Singapore Health Promotion Board. Healthier Choice Symbol. Singapore (SIN): HPB; 2018.

97. Health Canada. Regulations Amending Certain Regulations Made Under the Food and Drugs Act (Nutrition Symbols, Other Labelling Provisions, Partially Hydrogenated Oils and Vitamin D). Ontario (CAN): Government of Canada; 2018.

98. Deschasaux M, Julia C, Kesse-Guyot E, Lécuyer L, Adriouch S, Méjean C, et al. Are self-reported unhealthy food choices associated with an increased risk of breast cancer? Prospective cohort study using the British Food Standards Agency nutrient profiling system. BMJ Open. 2017;7(6):e013718.

99. Donnenfeld M, Julia C, Kesse-Guyot E, Méjean C, Ducrot P, Péneau S, et al. Prospective association between cancer risk and an individual dietary index based on the British Food Standards Agency Nutrient Profiling System. Br J Nutr. 2015;114(10):1702-10.

100. Julia C, Ducrot P, Lassale C, Fézeu L, Méjean C, Péneau S, et al. Prospective associations between a dietary index based on the British Food Standard Agency nutrient profiling system and 13-year weight gain in the SU. VI. MAX cohort. Prev Med. 2015;81:189-94.

101. Masset G, Scarborough P, Rayner M, Mishra G, Brunner EJ. Can nutrient profiling help to identify foods which diet variety should be encouraged? Results from the Whitehall II cohort. Br J Nutr. 2015;113(11):1800-9.

102. Adriouch S, Julia C, Kesse-Guyot E, Ducrot P, Peneau S, Mejean C, et al. Association between a dietary quality index based on the food standard agency nutrient profiling system and cardiovascular disease risk among French adults. Int J Cardiol. 2017;234:22-7.

103. Adriouch S, Julia C, Kesse-Guyot E, Mejean C, Ducrot P, Peneau S, et al. Prospective association between a dietary quality index based on a nutrient profiling system and cardiovascular disease risk. Eur J Prev Cardiol. 2016;23(15):1669-76.

104. Mytton O, Forouhi N, Scarborough P, Lentjes M, Luben R, Rayner M, et al. Association between intake of less healthy foods defined by the UK’s nutrient profile model and cardiovascular disease: A population-based cohort study. PLoS Med. 2018;15(1):e1002484.

105. Legislative and Governance Forum on Food Regulation. Final Communique 27 June 2014. Canberra (AUST): Australian Government Food Regulation Secretariat; 2014.

106. World Cancer Research Fund International. Building Momentum: Lessons on Implementing a Robust Front-of-pack Food Label. London (UK): WCRF; 2019.

107. Magnusson R, Reeve B. Food reformulation, responsive regulation, and “regulatory scaffolding”: Strengthening performance of salt reduction programs in Australia and the United Kingdom. Nutrients. 2015;7(7):5281–308.

108. World Health Organization. Safeguarding Against Possible Conflicts of Interest in Nutrition Programmes: Draft Approach for the Prevention and Management of Conflicts of Interest in the Policy Development and Implementation of Nutrition Programmes at Country Level. Geneva (CHE): WHO; 2017.



Supplementary Appendix 1: Search strategy and database.


366 Australian and New Zealand Journal of Public Health 2019 vol. 43 no. 4© 2019 Cancer Council Victoria

Sponsorship of elite sport provides a compelling avenue for unhealthy food and sugary drink, alcohol and gambling

companies to promote their products. It is a unique and especially persuasive form of marketing1 and it allows them to advertise to a mass audience, including children and young adults. Branded sponsorship within an elite sport context assumes many forms such as: electronic and fixed signage within the stadium; logos painted on the field; branded uniforms; naming rights to a series, game or stadium; product endorsement by players; pop-up advertisements or verbal commentary during play; and commercial break advertisements.2-4 By embedding marketing within the game, sport sponsorship can cut through advertising clutter, generating immense brand exposure.5 Compared to traditional forms of advertising (e.g. television, radio, print), marketing via sport sponsorship is perceived as a less overt attempt to persuade, is more accepted by consumers and has been shown to meet with less cognitive resistance.1,6 Sponsorship harnesses spectators’ emotional engagement to facilitate transfer of positive and often inextricable associations with popular and valued sports, teams or players to a brand or product.1,7 It can create a ‘health halo’, whereby the image of sport as a healthy activity is transferred to sponsor brands,5,8 and enhance perception of a company’s social responsibility or goodwill in supporting a valued community event, particularly when it is believed the event is reliant on the sponsorship.9,10

From a public health perspective, elite sport sponsorship by unhealthy products

is concerning due to the promotion and normalisation of behaviours associated with adverse health and social outcomes, particularly for vulnerable groups. Advertising of unhealthy food and sugary drinks has been shown to positively influence diet-related attitudes and preferences, intentions and behaviours among children and adolescents.11-15 Likewise, exposure to alcohol advertising is associated with positive alcohol expectancies, attitudes and intentions, which are strong predictors of alcohol use,16-18 and has been shown to expedite initiation of drinking and increase consumption levels in children and adolescents.19-25 While the impact of elite sport sponsorship has been

less directly studied, there is evidence that sponsorship: effectively reaches children and increases awareness of unhealthy commodities26; can influence children’s perceptions of unhealthy food brands and sway family food purchases27; increases alcohol consumption in children28; stimulates harmful levels of drinking alcohol among sportspeople29,30; and can increase awareness, attitudes and preferences for sponsor products among young adults.8 Emerging evidence also suggests that gambling sponsorship may be particularly harmful to problem gamblers or those recovering from being problem gamblers,31,32 young males,32,33 adolescents34 and children.35

Unhealthy sport sponsorship at the 2017 AFL Grand Final: a case study of its frequency, duration and natureTegan Nuss,1 Maree Scully,1 Melanie Wakefield,1 Helen Dixon1

1. Centre for Behavioural Research in Cancer, Cancer Council VictoriaCorrespondence to: Dr Helen Dixon, Centre for Behavioural Research in Cancer, Cancer Council Victoria, 615 St Kilda Road, Melbourne, Victoria 3004;

e-mail: [email protected]: January 2019; Revision requested: April 2019; Accepted: May 2019The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,



Abstract

Objective: To assess the frequency, duration and nature of unhealthy marketing during the highest-rating sporting event in Australia in 2017.

Methods: A content analysis of the 2017 Australian Football League (AFL) Grand Final television broadcast identified episodes of unhealthy food and sugary drink, alcohol and gambling marketing (and pro-health marketing as a comparison).

Results: There were 559 unhealthy marketing episodes (47 minutes 17 seconds). Most (81%) were for unhealthy food and sugary drink products, while alcohol (9%) and gambling (10%) were less frequent. The total duration of unhealthy marketing was delivered primarily via fixed advertising (55%), dynamic advertising (32%) and branded objects (11%). For unhealthy food and sugary drinks, at least one episode was visible 25% of the time. For each of alcohol and gambling, at least one episode was visible 4% of the time. Unhealthy food and sugary drink marketing peaked in Quarter 2. Pro-health marketing was limited, with 26 episodes (2 minutes 59 seconds).

Conclusions: The 2017 AFL Grand Final broadcast featured a high frequency and extensive duration of unhealthy marketing, especially for unhealthy food and sugary drink brands.

Implications for public health: Findings strengthen evidence supporting calls to increase regulation of sport sponsorship by unhealthy brands.

Key words: sport sponsorship, content analysis, food, alcohol, gambling

SPORT

2019 vol. 43 no. 4 Australian and New Zealand Journal of Public Health 367© 2019 Cancer Council Victoria

Restricting the advertising of unhealthy food and sugary drink, alcohol and gambling products, including via sport sponsorship, has consequently been identified as a necessary strategy to reduce harm related to these products.36-40

Research examining the extent and nature of tobacco advertising via sport sponsorship was crucial in building the evidence that led to Federal Legislation (Tobacco Advertising Prohibition Act 1992) banning tobacco sponsorship in Australia.41,42 In the past decade, a small number of studies employing content analysis have identified high volumes of sponsorship by unhealthy food and sugary drink, alcohol and gambling brands in popular, high-profile sporting codes/events in Australia2-4,43,44 and New Zealand.45,46 The present study aimed to contribute to existing studies by quantifying the frequency, duration and nature of unhealthy food and sugary drink, alcohol, and gambling marketing during the highest rating sport event in Australia in 2017 – the Australian Football League (AFL) Grand Final.47 Marketing delivering pro-health messages was also examined to provide a point of comparison.

Method

A content analysis of a digital recording of the 2017 AFL Grand Final television broadcast, including commercials as shown in the Melbourne metropolitan region (televised free-to-air on Channel 7, 30 September 2017), was undertaken to identify marketing

episodes for three types of unhealthy products: unhealthy food and sugary drinks, alcohol and gambling. Food and non-alcoholic drink brands were classified as unhealthy if the majority of products sold and promoted under that brand were deemed discretionary food and drink choices in the Australian Dietary Guidelines;48 that is, food and drink products not required for a healthy diet and typically energy-dense and nutrient-poor (e.g. containing high amounts of saturated fat, sodium/salt, or added sugar).48 Episodes of pro-health marketing (such as moderation messages) were also identified to examine the extent to which in-game advertising may address the potential negative health consequences of consuming unhealthy food, beverage and gambling products.

The 2017 AFL Grand Final broadcast was chosen for study because it was the highest rating sporting event in Australia in 2017, watched by more than 2.7 million viewers, including 322,000 children under the age of 15 years, across major cities in Australia.47,49

The coding framework was developed following previous content analysis studies of marketing during sporting events.2-4,43,50 A marketing episode was coded if the brand name, logo, slogan or readily identifiable imagery or messaging (e.g. McDonald’s Monopoly campaign, Four’n Twenty’s pie image) was clearly visible for at least one second. The length of time each episode was visible was recorded to the nearest second using the timestamp on the video file. For each episode, the brand, type of product

(e.g. unhealthy food and sugary drink) and sub-type (e.g. fast food, soft drink, sports drink), and type of promotion were coded. The types of promotions coded comprised the following five categories, with examples of these shown in Figure 1:

1. Dynamic advertising: Advertising on revolving or electronic banners or signs within the stadium, including the scoreboard.

2. Fixed advertising: Advertising on static banners or signage within the stadium.

3. Integrated advertising: Advertising via on-screen pop-ups and pull-through banners or broadcast announcements.

4. Commercial break advertising: Advertisements during commercial breaks.

5. Branded objects: Logos or other identifiable branding on objects used by players, umpires and other staff on field.

Marketing episodes that occurred simultaneously were coded as separate episodes if they were for a different brand or type of promotion. For example, if a Coca-Cola electronic banner was visible at the same time as Coca-Cola fixed signage, each was coded separately, or if Coca-Cola fixed signage was visible at the same time as McDonald’s fixed advertising, each was coded separately.

The total duration of coded footage was 2 hours 6 minutes 37 seconds. All in-game time was coded, which included the time from the first bounce of each quarter to the start of the between-quarter commercial break for Quarters 1 to 3, and to the final siren for Quarter 4 (Quarter 1: 30 minutes 28 seconds; Quarter 2: 34 minutes 26 seconds; Quarter 3; 30 minutes 24 seconds; Quarter 4: 31 minutes 19 seconds). Commercial breaks within each quarter (i.e. following goals) were coded, while commercial breaks between quarters, including at half-time, were not coded. Pre- and post-match footage was not coded.

The broadcast was independently coded by two of the researchers, using screens of identical size, via playback using VLC Media Player Version 3.0.2.51 Inter-rater reliability was calculated based on 10 minutes of coding using Krippendorff’s alpha, which is suitable for both nominal and ratio data.52 Inter-rater reliability was acceptable (α>0.8) for coding of episodes (α=0.83) and their duration (α=0.94). Any discrepancies in coding were reviewed by the researchers together until consensus was reached.

Figure 1: Examples of types of promotions coded (clockwise from top left): fixed advertising (McDonald’s goal posts, Coca-Cola signage within stadium); dynamic advertising (Four’n Twenty electronic banner); integrated advertising (McDonald’s on-screen pop-up); branded objects (Gatorade padded blocks); commercial break advertising (McDonald’s advertisement).

Sport Unhealthy sport sponsorship at the 2017 AFL Grand Final


Data analysisThe data were analysed using Microsoft Excel and Stata MP 14.2.53 Descriptive statistics were used to analyse the frequency and duration of marketing episodes by type of product, brand and promotion type, where each episode was counted separately. For each unhealthy product type and for unhealthy marketing overall, the total proportion of coded game time when marketing was present was calculated where episodes that occurred simultaneously were not counted separately. That is, the proportion of coded time when at least one marketing episode was visible was calculated. For each unhealthy product type and unhealthy marketing overall, logistic regression with post hoc pairwise comparisons using Bonferroni correction were conducted to examine if the proportion of coded time when at least one episode was visible differed by game quarter.

Results

Frequency and duration of unhealthy marketing overallOverall, there were 559 episodes of unhealthy marketing during the coded time, totalling 47 minutes 17 seconds duration when each episode was counted separately (Table 1).

This corresponds to viewers being exposed to, on average, 4.4 unhealthy marketing episodes per minute. The duration of individual unhealthy marketing episodes ranged in length, from 1 to 39 seconds (median = 3 seconds). More than half (55%) of the total duration of unhealthy marketing was delivered via fixed advertising, almost one-third (32%) via dynamic advertising, and 11% via branded objects (e.g. drink bottles).

Frequency and duration of unhealthy food and sugary drink marketing The majority (81%, n=452) of unhealthy marketing episodes were for fast food, soft drink and sport drink brands (Table 1). Viewers were exposed to an average of 3.6 episodes of unhealthy food and sugary drink marketing per minute. Promotions for McDonald’s made up around half (51%) of the total duration of unhealthy food and sugary drink marketing, amounting to 18 minutes 33 seconds. Coca-Cola accounted for 29% of the duration of unhealthy food and sugary drink marketing, totalling 10 minutes 21 seconds; Gatorade made up 17% (6 minutes 8 seconds); and Four’n Twenty made up 3% (1 minute 15 seconds).

As shown in Table 1, two-thirds (67%) of the total duration of unhealthy food and sugary drink marketing was achieved via fixed

advertising. This was predominantly due to McDonald’s branding on the goal posts (15 minutes 47 seconds), but also included Coca-Cola signage within the stadium (7 minutes 26 seconds), Gatorade signage surrounding the players’ bench (53 seconds) and Four’n Twenty signage within the stadium (11 seconds). Dynamic advertising accounted for 17% of the total duration of unhealthy food and sugary drink marketing and included electronic banners for Coca-Cola (2 minutes 55 seconds), McDonald’s (2 minutes 8 seconds), and Four’n Twenty (1 minute 4 seconds). Branded objects, which included Gatorade-branded drink bottles, portable coolers, towels, mats and padded blocks positioned on the field’s perimeter, made up 14% of the unhealthy food and sugary drink marketing duration. Commercial breaks (30 seconds) and integrated advertising (8 seconds), both for McDonald’s, each contributed 1% or less of the total unhealthy food and sugary drink duration. See Figure 1 for examples of in-game promotions for unhealthy food and sugary drink brands.

Frequency and duration of alcohol marketingPromotions for beer, cider, wine and an alcohol retailer accounted for 9% (n=50) of unhealthy marketing episodes (Table 1).

Table 1: Frequency and duration (minutes, seconds) of unhealthy and pro-health marketing episodes, by product type and promotion type.Fixed

advertisingDynamic

advertisingBranded objects Commercial break Integrated

advertisingTotal

Unhealthy marketing

Unhealthy food and sugary drinksEpisodes (%) 282 (62%) 69 (15%) 98 (22%) 2 (<1%) 1 (<1%) 452Duration (%) 24m 17s (67%) 6m 7s (17%) 5m 15s (14%) 30s (1%) 8s (<1%) 36m 17s

GamblingEpisodes (%) 13 (23%) 44 (77%) – – – 57Duration (%) 38s (11%) 4m 56s (89%) – – – 5m 34s

Alcohol Episodes (%) 13 (26%) 36 (72%) – 1 (2%) – 50Duration (%) 52s (16%) 4m 4s (75%) – 30s (9%) – 5m 26s

Total unhealthy marketingEpisodes (%) 308 (55%) 149 (27%) 98 (18%) 3 (1%) 1 (<1%) 559Duration (%) 25m 47s (55%) 15m 7s (32%) 5m 15s (11%) 1m (2%) 8s (<1%) 47m 17s

Pro-health marketing

Healthy eatingEpisodes (%) – – – – – –Duration (%) – – – – – –

Responsible gamblingEpisodes (%) – – – – – –Duration (%) – – – – – –

Alcohol harm prevention:

Publicly-fundedEpisodes (%) – 25 (100%) – – – 25Duration (%) – 2m 53s (100%) – – – 2m 53s

Industry-fundedEpisodes (%) – 1 (100%) – – – 1Duration (%) – 6s (100%) – – – 6s

Total pro-health marketingEpisodes (%) – 26 (100%) – – – 26Duration (%) – 2m 59s (100%) – – – 2m 59s

Notes: Where multiple marketing episodes occurred simultaneously each episode was coded as a single episode provided it was a different brand and/or type of promotion (i.e., dynamic, fixed, integrated, commercial, branded objects). Duration was

calculated counting each episode separately. Percentages reported to nearest whole number so may not sum to 100.

Nuss et al. Article



On average, viewers were exposed to one episode of alcohol marketing every 2 minutes 32 seconds. The total duration of alcohol marketing was made up of promotions for Carlton Draught (71%; 3 minutes 51 seconds), IGA Liquor (21%; 1 minute 7 seconds), Mercury Cider (6%; 20 seconds) and Wolf Blass (2%; 8 seconds). The total duration of alcohol marketing occurred primarily via dynamic advertising (75%), which included electronic banners for Carlton Draught (2 minutes 37 seconds), IGA Liquor (1 minute 7 seconds), and Mercury Cider (20 seconds). Fixed advertising accounted for 16% of the total alcohol marketing duration, including signage for Carlton Draught (44 seconds) and Wolf Blass (8 seconds), while a commercial break for Carlton Draught (30 seconds) contributed 9%.

Frequency and duration of gambling marketingAs shown in Table 1, one-in-ten (10%, n=57) unhealthy marketing episodes were for gambling brands. On average, viewers were exposed to a gambling episode every 2 minutes 13 seconds. The total duration of gambling marketing was made up of promotions for Ladbrokes (58%; 3 minutes 17 seconds), bet365 (38%; 2 minutes 8 seconds), CrownBet (1%; 5 seconds) and SportsBet (1%; 4 seconds). The majority (89%) of gambling marketing duration was achieved via electronic banners promoting Ladbrokes (2 minutes 48 seconds) and bet365 (2 minutes 8 seconds), while 11% was achieved through fixed signs for Ladbrokes (29 seconds), CrownBet (5 seconds), and SportsBet (4 seconds).

Proportion of coded footage when unhealthy marketing was visibleTable 2 presents the proportion of coded time when at least one episode of each type of unhealthy marketing was visible. Overall, at least one episode of unhealthy marketing (unhealthy food and sugary drinks, alcohol, or gambling) was visible almost one-third (30%) of the coded time. There was a significant association between game quarter and the proportion of time at least one unhealthy marketing episode was visible (X2=44.01, p<0.001). Bonferroni-adjusted post hoc comparisons revealed that unhealthy marketing was significantly higher in Quarter 1 compared to Quarter 3 (p=0.018) and Quarter 4 (p<0.001), and significantly higher in Quarter 2 compared to Quarter 3 (p<0.001) and Quarter 4 (p<0.001).

At least one episode of marketing for unhealthy food and sugary drinks was visible 25% of the coded time (Table 2). There was a significant association between game quarter and the proportion of time at least one episode of unhealthy food and sugary drink marketing was visible (X2=54.18, p<0.001), with unhealthy food and sugary drink marketing higher in Quarter 2 compared to Quarter 1 (p=0.002), Quarter 3 (p<0.001), and Quarter 4 (p<0.001). Marketing for unhealthy food and sugary drink was also significantly higher in Quarter 1 compared to Quarter 4 (p=0.013).

For both alcohol and gambling, at least one episode of marketing was visible 4% of the coded time (Table 2). There was a significant association between game quarter and the proportion of time at least one alcohol marketing episode was visible (X2=29.22, p<0.001), with alcohol marketing higher in

Quarter 1 compared to Quarter 2 (p=0.011) and Quarter 3 (p<0.001), and higher in Quarter 4 compared to Quarter 3 (p=0.026). The proportion of time at least one episode of gambling marketing was visible did not differ across game quarters (X2=4.17, p=0.244).

Pro-health marketingOverall, there were 26 episodes of pro-heath marketing, totalling 2 minutes 59 seconds, when each episode was counted separately, all of which were delivered via dynamic advertising (as per Table 1). The longest pro-health marketing episode was 25 seconds (median=4 seconds). All pro-health marketing was directed toward alcohol harm prevention and mostly comprised promotions for a publicly funded campaign, specifically the Victorian Government’s Transport Accident Commission’s (TAC) “Towards Zero” and “Drinking. Driving. They’re better apart.” messages, which appeared on electronic banners surrounding the stadium.

There was one marketing episode for DrinkWise, an alcohol industry public relations initiative,54,55 which appeared on the electronic scoreboard.

As per Table 2, at least one episode of pro-health marketing was visible 2% of the coded time. There was a significant association between game quarter and the proportion of time at least one episode of pro-health marketing was visible (X2=10.11 p=0.018), with pro-health marketing higher in Quarter 2 compared to Quarter 1 (p=0.025).

Discussion

This study shows that the 2017 AFL Grand Final television broadcast included a high frequency and extensive duration of unhealthy marketing, particularly for unhealthy food and sugary drink brands. Marketing for alcohol and gambling products was less prevalent, accounting for about one-fifth of all unhealthy marketing episodes. Overall, marketing for unhealthy products dwarfed pro-health marketing, with more than twenty times as many episodes of such marketing recorded (559 unhealthy compared to 26 pro-health marketing episodes). To our knowledge, this is the first study to investigate all three products (unhealthy food and sugary drinks, alcohol and gambling) within an AFL match – notably the Grand Final – and to quantitatively examine how marketing for unhealthy brands varies across the game.

Table 2: Proportion of coded time that at least one episode of unhealthy marketing was visible, by product type and game quarter.

Overall Quarter 1a Quarter 2b Quarter 3c Quarter 4(%) (%) (%) (%) (%)

Unhealthy food and sugary drinks 25.0 25.4 30.5a** 22.5b*** 21.1a* b***

Gambling 4.4 4.6 3.6 4.8 4.6Alcohol 4.3 6.2 4.0a* 2.6a*** 4.3c*

Total unhealthy marketing 29.7 31.9 33.8 27.4a* b*** 25.1a*** b***

Healthy eating – – – – –Responsible gambling – – – – –Alcohol harm prevention 2.4 1.6 3.0a* 2.0 2.7Total pro-health marketing 2.4 1.6 3.0a* 2.0 2.7Notes: Where multiple marketing episodes occurred simultaneously the time has only been counted once. Hence, the sum of proportions by product type do not equal

the overall unhealthy marketing proportion.a, b, c: Reference categories for logistic regression analyses.*** denotes a significant difference from the reference category, at p < 0.001; ** denotes a significant difference from the reference category, at p < 0.01;

* denotes a significant difference from the reference category, at p < 0.05.


While it is encouraging that there was some pro-health marketing addressing alcohol harm prevention, its capacity to cut through and impact viewers was likely to have been diluted by the fact that the broadcast featured twice as many incidences of pro-alcohol marketing (50 compared to 26). Most of the alcohol harm prevention marketing came from a reputable publicly funded campaign. A minority of pro-health alcohol messaging came from DrinkWise, an alcohol industry public relations initiative, which has been criticised for potentially benefiting industry more than public health.54,55 Alternative publicly funded alcohol education campaigns are available that demonstrably improve audience awareness of specific alcohol harms56 and motivate drinkers to consume less.57

There were 452 marketing episodes for unhealthy food and sugary drink brands during the Grand Final and, although some of these occurred simultaneously, at least one was visible for 25% of the coded time. The high level of unhealthy food and sugary drink marketing in one of the most watched nationally televised events (sporting or otherwise) is concerning, given that unhealthy food marketing has been identified as promoting poor diet and contributing to the obesity epidemic38 and nearly two-thirds of Australian adults and more than one-quarter of Australian children are above a healthy weight.58 While unhealthy food and sugary drink marketing clearly dominated both alcohol and gambling marketing in the 2017 AFL Grand Final, this same pattern was not evident in an earlier content analysis study that also focused on these three products. Specifically, Lindsay and colleagues2 observed higher levels of alcohol marketing during the 2012 NRL State of Origin series (an average of 66 minutes per game) compared to both gambling (9 minutes per game) and unhealthy food and sugary drink (3 minutes per game) marketing. These contrasting findings are not unexpected, with previous research suggesting that different sports may be targeted by, or more strongly associated with, particular products59,60 or that the products promoted can vary between games50 or between broadcasts of the same game4 within sporting codes.

In the present study, for each of alcohol and gambling marketing at least one episode was visible 4% of the coded time. Although less prevalent than unhealthy food and sugary drink marketing, the presence of promotions

for products that are not legally available to persons under 18 years of age raises serious questions about the kind of messages elite sport is sending young people about alcohol and gambling. Alcohol sponsorship is known to influence children’s product awareness, preferences and consumption,28,61 rendering the 322,000 Australian children who watched the AFL Grand Final49 vulnerable to such marketing effects. (To put this figure in perspective, this is enough children to fill the Melbourne Cricket Ground three times over.) Prior studies indicate that levels of alcohol and gambling marketing can vary considerably within sporting codes. For example, an examination of the 2012 AFL finals series found the frequency of alcohol marketing ranged between games, likely as a function of venue, with the semi-final at ANZ stadium in Sydney featuring 570 episodes compared to the Grand Final at the Melbourne Cricket Ground featuring 67 episodes.50 Further, a content analysis of eight televised AFL matches from 2011 observed between two and 123 episodes (or 0.2 and 11.3 minutes) of gambling marketing per game, with free-to-air matches featuring a greater amount than matches broadcast on Pay TV.4 While the level of alcohol and gambling marketing may appear low in isolated games, sports viewers’ exposure is likely to be cumulative and can build up quickly if they watch several games across every weekend of the season.

More than half of the unhealthy marketing that featured in the Grand Final was fixed advertising, with dynamic advertising and branded objects, such as drink bottles, portable coolers, towels, mats and padded blocks positioned on the field’s perimeter, also common. As with earlier content analyses,2-4,44 these results highlight the myriad strategies sponsors use to create a saturated environment in which it is difficult for viewers to avoid this marketing or to separate it from the experience of the game itself. For example, McDonald’s logos around the goal posts ensure the brand is prominent to spectators at key moments in the game (e.g. when a player is kicking for goal), and dynamic banners flash and revolve, allowing brands to intrude even in the middle of exciting play. Additionally, vision of players drinking from Gatorade-branded bottles potentially signals to the audience their tacit endorsement of the product. Celebrity endorsement has been shown to exert a powerful influence on attitudes and

preferences, particularly for children and adolescents.62-65 Although these unhealthy marketing episodes often occur in brief bursts of short duration and typically feature just a brand or logo, they occur at a high frequency, and there is evidence that this mode of marketing may be more effective than longer, traditional commercial break advertising,66 since exposure effects may be stronger when stimuli are not consciously attended to.67,68 The formation of an environment where marketing is simultaneously pervasive and subtle, and also fused with the experience of the game, is particularly concerning when considering the strong viewership of televised sport by children, who do not recognise the commercial intent of sponsorship or have the cognitive capacity to critically evaluate advertising messages.69-71

Findings revealed that unhealthy food and sugary drink marketing was highest in the second quarter, coinciding with the time viewers may be planning their half-time snack or meal. From our study, it is not possible to determine whether this was a deliberate marketing ploy or a chance occurrence. As the majority of unhealthy food and sugary drink marketing was delivered via fixed advertising, primarily McDonald’s branded goal posts, this pattern of results could be due to particular camera shots appearing more frequently in the second quarter. It is important to note, however, that the number of points scored during this quarter (a time in which the goal posts typically appear on screen for an extended period) was comparable to the other three quarters.72 Alcohol marketing also varied between quarters, with higher prevalence in the first quarter compared to the middle quarters, although the magnitude of these differences was relatively small.

The ubiquitous nature of unhealthy marketing in the 2017 AFL Grand Final underlines the inadequacies of the current system of largely voluntary and mostly self-regulated advertising and marketing in Australia. Both alcohol advertising regulations and guidelines pertaining to food advertising to children exempt marketing that occurs within sport broadcasts.73-80 Given that more advertising for unhealthy food and sugary drink and alcohol products has been observed during televised sport than during any other programming,43,81 this represents clear loopholes in the frameworks intended to protect young people. Similarly, although a ban on betting and gambling commercials during live sport broadcasts between 5.00 am

Nuss et al. Article


and 8.30 pm was introduced in March 2018, the restriction excludes sponsorship (termed “incidental accompaniment”),75 meaning that children and other vulnerable groups remain exposed to gambling advertising embedded within the game. Thus, reforms are needed that place tighter restrictions, or ban altogether, sport sponsorship by unhealthy brands.

The successful removal of tobacco sponsorship in Australia in the early 1990s provides a useful model for how other unhealthy product sponsors could feasibly be banned from sport without compromising the viability of these events.82 Replacing unhealthy sponsorship with health promotion sponsorship, as occurred with tobacco, could form one aspect of multi-component strategies to address overweight and obesity, and harms related to alcohol and gambling.83,84 There are already some examples of this occurring, most notably Western Australia’s Healthway partnerships with the Western Australian Cricket Association (WACA) and West Coast Fever netball team, which include promotion of their Alcohol. Think Again and LiveLighter campaign messages, respectively.85,86 However, the dearth of pro-health marketing in the 2017 AFL Grand Final – of which none were promoting good nutrition – shows there is scope for more of these types of sponsorship relationships, particularly in the AFL context. It should also be noted that in addition to the unhealthy and pro-health brands quantified in our study, many other brands featured throughout the game, from car manufacturers and airlines to electronics producers and retail and media outlets, suggesting that sport sponsorship is an attractive marketing avenue for a wide variety of brands. It is, therefore, likely that the removal of sponsorship by unhealthy food and sugary drink, alcohol and gambling products would allow other brands to contend as sponsors rather than render such sporting events unviable.

A number of study limitations should be noted. First, while content analysis enabled us to document the amount of unhealthy and pro-health sponsor marketing that spectators were potentially exposed to, it did not measure spectators’ actual exposure to, or recall and recognition of, sponsor brands; this would have required a separate study. Second, only promotions that were visible for at least one second during

in-game time were coded; we did not, for example, capture sponsored segments that were televised at half-time (e.g. Gatorade AFL Grand Final Sprint, Macca’s Champion Player) or pre-game footage inside the team changerooms where Gatorade signage was prominently displayed. Consequently, our study likely provides a conservative estimate of the volume of unhealthy marketing that featured in the televised coverage of the 2017 AFL Grand Final. Third, this study focused on a single, broadcast sporting event and did not assess spectators’ exposure to promotions for unhealthy products at the stadium. Future research should investigate the extent and nature of unhealthy and pro-health sponsorship across multiple sports and venues, including within both live stadium and broadcast settings, to provide a clearer understanding of how companies use sport sponsorship to reach and influence consumers with marketing for potentially harmful products.

Conclusions

The televised broadcast of the 2017 AFL Grand Final featured a high frequency and extensive duration of unhealthy marketing, especially for unhealthy food and sugary drink brands. Marketing was delivered in repeated brief bursts and via numerous promotion types embedded within the game, creating a saturated environment in which marketing is not only hard to avoid but difficult to separate from the experience of the game itself. Given the effectiveness of advertising in driving attitudes, preferences and behaviours related to potentially harmful products, these study findings add evidence to support calls for greater regulation of sport sponsorship by unhealthy food and sugary drink, alcohol and gambling brands.

Acknowledgements

This research was funded by the National Health & Medical Research Council’s Targeted Call for Research into Preventing Obesity in 18-24-year-olds (APP1114923). The funding body was not involved in the design of the study or the collection, analysis and interpretation of data, or in writing the manuscript. MW was supported by funding from an NHMRC Principal Research Fellowship.

References1. Meenaghan T. Understanding sponsorship effects.

Psychol Market. 2001;18(2):95-122.2. Lindsay S, Thomas S, Lewis S, et al. Eat, drink and

gamble: Marketing messages about ‘risky’ products in an Australian major sporting series. BMC Public Health. 2013;13:719.

3. Sherriff J, Griffiths D, Daube M. Cricket: Notching up runs for food and alcohol companies? Aust N Z J Public Health. 2010;34(1):19-23.

4. Thomas S, Lewis S, Duong J, et al. Sports betting marketing during sporting events: A stadium and broadcast census of Australian Football League matches. Aust N Z J Public Health. 2012;36(2):145-52.

5. Meenaghan T, Shipley D. Media effect in commercial sponsorship. Eur J Mark. 1999;33(3/4):328-48.

6. Lardinoit T, Quester P. Attitudinal effects of combined sponsorship and sponsor’s prominence on basketball in Europe. J Advert Res. 2001;41(1):48-58.

7. Gwinner KP, Eaton J. Building brand image through event sponsorship: The role of image transfer. J Advert. 1999;28(4):47-57.

8. Dixon H, Scully M, Wakefield M, et al. The impact of unhealthy food sponsorship vs. pro-health sponsorship models on young adults’ food preferences: A randomised controlled trial. BMC Public Health. 2018;18:1399.

9. Lavack AM. An inside view of tobacco sports sponsorship: An historical perspective. Int J Sports Mark Sponsorship. 2003;5(2):105-28.

10. Lipkus IM, Crawford Y, Fenn K, et al. Testing different formats for communicating colorectal cancer risk. J Health Commun. 1999;4(4):311-24.

11. Andreyeva T, Kelly IR, Harris JL. Exposure to Food Advertising on Television: Associations with Children’s Fast Food and Soft Drink Consumption and Obesity. NBER Working Paper No.: 16858. Cambridge (MA): National Bureau of Economic Research; 2011.

12. Borzekowski DL, Robinson TN. The 30-second effect: An experiment revealing the impact of television commercials on food preferences of preschoolers. J Am Diet Assoc. 2001;101(1):42-6.

13. Boyland EJ, Halford JC. Television advertising and branding. Effects on eating behaviour and food preferences in children. Appetite. 2013;62:236-41.

14. Cairns G, Angus K, Hastings G. The Extent, Nature and Effects of Food Promotion to Children: A review of the Evidence to December 2008. Prepared for World Health Organization. Geneva (CHE): World Health Organization; 2009.

15. Harris JL, Bargh JA, Brownell KD. Priming effects of television food advertising on eating behavior. Health Psychol. 2009;28(4):404-13.

16. Stacy AW, Zogg JB, Unger JB, et al. Exposure to televised alcohol ads and subsequent adolescent alcohol use. Am J Health Behav. 2004;28(6):498-509.

17. Kelly KJ, Edwards RW. Image advertisements for alcohol products: Is their appeal associated with adolescents’ intention to consume alcohol? Adolescence. 1998;33(129):47-59.

18. Grube JW. Television alcohol portrayals, alcohol advertising, and alcohol expectancies among children and adolescents. In: Martin S, editor. NIAAA Research Monograph No 28: The Effects of the Mass Media on the Use and Abuse of Alcohol. Rockville (MD): National Institute on Alcohol Abuse and Alcoholism; 1995. p. 105-21.

19. Anderson P, de Bruijn A, Angus K, et al. Impact of alcohol advertising and media exposure on adolescent alcohol use: A systematic review of longitudinal studies. Alcohol Alcohol. 2009;44(3):229-43.

20. Collins RL, Ellickson PL, McCaffrey D, et al. Early adolescent exposure to alcohol advertising and its relationship to underage drinking. J Adolesc Health. 2007;40(6):527-34.

21. Smith L, Foxcroft D. The effect of alcohol advertising, marketing and portrayal on drinking behaviour in young people: Systematic review of prospective cohort studies. BMC Public Health. 2009;9:51.



22. Ellickson PL, Collins RL, Hambarsoomians K, et al. Does alcohol advertising promote adolescent drinking? Results from a longitudinal assessment. Addiction. 2005;100(2):235-46.

23. Jones SC, Magee CA. Exposure to alcohol advertising and alcohol consumption among Australian adolescents. Alcohol Alcohol. 2011;46(5):630-7.

24. Snyder LB, Milici FF, Slater M, et al. Effects of alcohol advertising exposure on drinking among youth. Arch Pediatr Adolesc Med. 2006;160(1):18-24.

25. Rossen I, Pettigrew S, Jongenelis M, et al. Evidence on the Nature and Extent of Alcohol Promotion and the Consequences for Young People’s Alcohol Consumption. Report prepared for the Mental Health Commission. Perth (AUST): Curtin University School of Psychology and Speech Pathology Western Australian Cancer Prevention Research Unit; 2018.

26. Bestman A, Thomas SL, Randle M, et al. Children’s implicit recall of junk food, alcohol and gambling sponsorship in Australian sport. BMC Public Health. 2015;15(1):1022.

27. Kelly B. Food and Beverage Company Sponsorship of Children’s Sport: Publicity or Philanthropy? Sydney (AUST): University of Sydney Sydney Medical School; 2012.

28. Brown K. Association between alcohol sports sponsorship and consumption: A systematic review. Alcohol Alcohol. 2016;51(6):747-55.

29. Andreyeva T, Chaloupka FJ, Brownell KD. Estimating the potential of taxes on sugar-sweetened beverages to reduce consumption and generate revenue. Prev Med. 2011;52(6):413-16.

30. O’Brien KS, Miller PG, Kolt GS, et al. Alcohol industry and non-alcohol industry sponsorship of sportspeople and drinking. Alcohol Alcohol. 2011;46(2):210-13.

31. Hing N, Vitartas P, Lamont M. Gambling sponsorship of sport: An exploratory study of links with gambling attitudes and intentions. Int Gambl Stud. 2013;13(3):281-301.

32. Hing N, Russell AMT, Lamont M, et al. Bet anywhere, anytime: An analysis of internet sports bettors’ responses to gambling promotions during sports broadcasts by problem gambling severity. J Gambl Stud. 2017;33(4):1051-65.

33. Deans EG, Thomas SL, Derevensky J, et al. The influence of marketing on the sports betting attitudes and consumption behaviours of young men: Implications for harm reduction and prevention strategies. Harm Reduct J. 2017;14(1):5.

34. Hing N, Vitartas P, Lamont M, et al. Adolescent exposure to gambling promotions during televised sport: An exploratory study of links with gambling intentions. Int Gambl Stud. 2014;14(3):374-93.

35. Thomas SL, Pitt H, Bestman A, et al. Child and Parent Recall of Gambling Sponsorship in Australian Sport. Melbourne (AUST): Victorian Responsible Gambling Foundation; 2016.

36. Australian Medical Association. Alcohol Marketing and Young People: Time for A New Policy Agenda. Canberra (AUST): AMA; 2012.

37. National Preventative Health Taskforce. Australia: The Healthiest Country by 2020. Technical Report No.: 1. Obesity in Australia: A Need for Urgent Action. Including Addendum for October 2008 to June 2009. Canberra (AUST): Government of Australia; 2009.

38. World Health Organization. Set of Recommendations on the Marketing of Foods and Non-alcoholic Beverages to Children. Geneva (CHE): WHO; 2010.

39. Parliamentary Joint Select Committee on Gambling Reform. Fifth Report: The Advertising and Promotion of Gambling Services in Sport. Broadcasting Services Amendment (Advertising for Sports Betting) Bill 2013. Canberra (AUST): Government of Australia; 2013.

40. World Health Organization. Global Strategy to Reduce the Harmful Use of Alcohol. Geneva (CHE): WHO; 2010.

41. Harris K. What do we see when we watch the cricket? Soc Alternatives. 1988;7(3):65-70.

42. Freeman B, Haslam I, Scollo M, et al. 11.0 Tobacco advertising and promotion: background. In: Scollo M, Winstanley MH, editors. Tobacco in Australia: Facts and Issues. Melbourne (AUST): Cancer Council Victoria; 2012.

43. VicHealth. Alcohol and Junk Food Advertising and Promotion through Sport: Research Highlights. Melbourne (AUST): State Government of Victoria; 2014.

44. Milner L, Hing N, Vitartas P, et al. Embedded gambling promotion in Australian football broadcasts: An exploratory study. Commun Polit Cult. 2013;46:177-98.

45. Chambers T, Signal L, Carter MA, et al. Alcohol sponsorship of a summer of sport: a frequency analysis of alcohol marketing during major sports events on New Zealand television. N Z Med J. 2017;130(1448):27-33.

46. Gee S, Sam MP, Jackson S. Content analyses of alcohol-related images during television broadcasts of major sports events in New Zealand. Int J Sports Mark Sponsorship. 2017;18(3):230-45.

47. AdNews. AdNews Analysis: The Top 20 TV Shows of 2017 [Internet]. Surrey Hills (AUST): Yaffa Media; 2018 [cited 2018 Sep13]. Available from: http://www.adnews.com.au/news/adnews-analysis-the-top-20-tv-shows-of-2017


49. OzTAM. Consolidated Ratings for Week 39, 2017 - Viewing Habits of children 0-15 Years (Capital Cities). North Sydney (AUST): OzTAM; 2017.

50. Jones SC, Barrie L, Chapman M, et al. Alcohol advertising during televised Australian Football finals. Proceedings of the ANZMAC; 2013 Dec 1-4; Auckland, New Zealand. Auckland: University of Auckland; 2013.

51. VideoLAN Organization. VLC Media Player. Version 3.0.2. Paris (FRA): VideoLAN Organization; 2018.

52. Hayes AF, Krippendorff K. Answering the call for a standard reliability measure for coding data. Commun Methods Meas. 2007;1(1):77-89.

53. STATA: statistical software. Release 14. College Station (TX): Stata Corporation; 2015.

54. Brennan E, Wakefield MA, Durkin SJ, et al. Public awareness and misunderstanding about DrinkWise Australia: A cross-sectional survey of Australian adults. Aust N Z J Public Health. 2017;41(4):352-7.

55. Pettigrew S, Biagioni N, Daube M, et al. Reverse engineering a ‘responsible drinking’ campaign to assess strategic intent. Addiction. 2016;111(6):1107-13.

56. Dixon H, Pratt I, Scully M, et al. Using a mass media campaign to raise women’s awareness of the link between alcohol and cancer: Cross-sectional pre-intervention and post-intervention evaluation surveys. BMJ Open. 2015;5:e006511.

57. Wakefield MA, Brennan E, Dunstone K, et al. Features of alcohol harm reduction advertisements that most motivate reduced drinking among adults: An advertisement response study. BMJ Open. 2017;7(4):e014193.

58. Australian Institute of Health and Welfare. A Picture of Overweight and Obesity in Australia 2017. Canberra (AUST): AIHW; 2017.

59. Maher A, Wilson N, Signal L, et al. Patterns of sports sponsorship by gambling, alcohol and food companies: An Internet survey. BMC Public Health. 2006;6:95.

60. Pitt H, Thomas SL, Bestman A, et al. “It’s just everywhere!” Children and parents discuss the marketing of sports wagering in Australia. Aust N Z J Public Health. 2016;40(5):480-6.

61. Kelly B, Baur LA, Bauman AE, et al. Tobacco and alcohol sponsorship of sporting events provide insights about how food and beverage sponsorship may affect children’s health. Health Promot J Austr. 2011;22(2):91-6.

62. Boyland EJ, Harrold JA, Dovey TM, et al. Food choice and overconsumption: Effect of a premium sports celebrity endorser. J Pediatr. 2013;163(2):339-43.

63. Phillipson L, Jones SC. “I Eat Milo to Make Me Run Faster”: How the Use of Sport in Food Marketing may Influence the Food Beliefs of Young Australians. Proceedings of the Australian and New Zealand Marketing Academy Conference; 1-3 December. Sydney, Australia.

64. Bush A, Martin C, Bush V. Sports celebrity influence on the behavioral intentions of Generation Y. J Advert Res. 2004;44(1):108–18.

65. Erdogan BZ. Celebrity endorsement: A literature review. J Mark Manage. 1999;15:291-314.

66. Levin AM, Joiner C, Cameron G. The impact of sports sponsorship on consumers’ brand attitudes and recall: the case of NASCAR fans. J Curr Issues Res Advert. 2001;23(2):23-31.

67. Zajonc RB. Attitudinal effects of mere exposure. J Pers Soc Psychol. 1968;9(2, Pt.2):1-27.

68. Hansen J, Wanke M. Liking what’s familiar: The importance of unconscious familiarity in the mere-exposure effect. Soc Cogn. 2009;27(2):161-82.

69. Kunkel D, Wilcox BL, Cantor J, et al. Psychological issues in the increasing commercialization of childhood. In: Wilcox BL, Kunkel D, Cantor J, Dowrick P, Linn S, Palmer E, editors. Report of the APA Task Force on Advertising and Children. Washington (DC): American Psychological Association; 2004. p. 20-54.

70. Oates C, Blades M, Gunter B. Children and television advertising: When do they understand persuasive intent? J Consum Behav. 2002;1(3):238-45.

71. Oates C, Blades M, Gunter B, et al. Children’s understanding of television advertising: A qualitative approach. J Mark Commun. 2003;9:59-71.

72. Bowen N. Match Report: Terrific Tigers End 37-year Wait [Internet]. Melbourne (AUST): Australian Football League; 2017 [cited 2018 Oct 19]. Available from: http://www.afl.com.au/match-centre/2017/27/adel-v-rich

73. Australian Food and Grocery Council. Quick Service Restaurant Initiative for Responsible Advertising and Marketing to Children. Canberra (AUST): AFGC; 2018.

74. Australian Food and Grocery Council. Responsible Children’s Marketing Initiative. Canberra (AUST): AFGC; 2018.

75. Free TV Australia. Commercial Television Industry Code of Practice. Sydney (AUST): Free TV; 2018.

76. Australian Association of National Advertisers. Code of Ethics. Sydney (AUST): AANA; 2018.

77. Australian Association of National Advertisers. Code for Advertising and Marketing Communications to Children. Sydney (AUST): AANA; 2017.

78. The ABAC Scheme Limited. ABAC Responsible Alcohol Marketing Code [Internet]. Stirling (AUST): ABAC; 2017 [cited 2018 Fe 26]. Available from: http://www.abac.org.au/wp-content/uploads/2017/07/ABAC_CodeofConduct_2017_web.pdf

79. Outdoor Media Association. OMA Alcohol Advertising Guidelines. Sydney (AUST): OMA; 2016.

80. Australian Association of National Advertisers. Food and Beverages Advertising and Marketing Communications Code. Sydney (AUST): AANA; 2009.

81. O’Brien KS, Carr S, Ferris J, et al. Alcohol advertising in sport and non-sport TV in Australia, during children’s viewing times. PLoS One. 2015;10(8):e0134889.

82. Clarkson J. Time to get tough on unhealthy sponsorships. Health Promot J Austr. 2010;21(3):164-5.

83. Holman CD, Donovan RJ, Corti B, et al. Banning tobacco sponsorship: replacing tobacco with health messages and creating health-promoting environments. Tob Control. 1997;6(2):115-21.

84. Corti B, Donovan RJ, Holman CD, et al. Using sponsorship to promote health messages to children. Health Educ Behav. 1997;24(3):276-86.

85. Western Australian Cricket Association. Commercial Partners [Internet]. Perth (AUST): WACA; 2019 [cited 2019 Jan 7]. Available from: https://www.waca.com.au/waca/commercial-partners

86. West Coast Fever. Sponsors [Internet]. Perth (AUST): 2019 [cited 2019 Jan 7]. Available from: http://www.westcoastfever.com.au/sponsors

Nuss et al. Article


Regular physical activity is a leading factor in promoting good health and preventing chronic disease,1 and

physical inactivity is a major contributor to the global burden of disease.2 In Australia, more than 30% of all adults are insufficiently active;3 81% of children do not meet the physical activity guidelines;4 and physical inactivity is responsible for 2.6% of the total burden of disease and injuries.5 The term ‘insufficiently active’ refers to people who do not meet the Australian Physical Activity Guidelines.6

The Australian Federal Government recently released a sport-based strategy (Sport 2030) aiming to reduce the number of physically inactive Australians by 15% by 2030.7 At a state level, the Victorian Health Promotion Foundation (VicHealth), a statutory authority focused on promoting good health and preventing chronic disease, has a strategic imperative to get 300,000 more Victorians engaging in physical activity by 2023.8 Between 2015 and 2018, VicHealth endeavoured to encourage the engagement of new participants not interested or able to participate in traditional sports9 by investing in two programs – the State Sport Program (SSP) and the Regional Sport Program (RSP) – to facilitate the development of new sport products or scale their existing, flexible, non-traditional social sport products to target insufficiently active members of the community. For the purposes of their program work with sports organisations, and to aid communication with the general public, VicHealth also divided those classified

as insufficiently active into ‘somewhat active’ and ‘inactive’.10

The aims of the SSP and RSP align with previous calls for policy makers to view informal sports as an opportunity to encourage new user groups to engage in sport and physical activity,9 and to respond to shifting physical activity participation trends in Australia. These trends, which are also evident internationally,11,12 have included

stagnant or declining participation in many organised and team sports alongside a growth in informal and lifestyle sport and physical activity participation.13-15 There is a growing demand for opportunities to participate in sport that is social, flexible and non-competitive, fits in with busy lifestyles, and focuses on achieving personal health and social objectives, rather than winning and competition.13 In short, more people

Challenges for sport organisations developing and delivering non-traditional social sport products for insufficiently active populationsKiera Staley,1 Alex Donaldson,1 Erica Randle,1 Matthew Nicholson,1 Paul O’Halloran,1 Rayoni Nelson,2 Matthew Cameron2

1. Centre for Sport and Social Impact, La Trobe University, Victoria2. Victorian Health Promotion Foundation (VicHealth)Correspondence to: Dr Alex Donaldson, Centre for Sport and Social Impact, La Trobe University, Plenty Road, Bundoora, Victoria 3086; e-mail: [email protected]: November 2018; Revision requested: April 2019; Accepted: May 2019 The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,



Abstract

Objectives: To explore the challenges that Victorian sporting organisations experience when developing, delivering or scaling non-traditional social sport products to engage insufficiently active people.

Methods: Online Concept Mapping was used to gather qualitative data and analyse it quantitatively.

Results: A total of 68 participants (27 organisations) brainstormed 158 challenges. The research team synthesised these to 71 unique challenges for participants to sort into groups and rate for importance (0–5) and ease of overcoming (0–5). A nine-cluster solution – Deliverers; Capacity to drive the product; Facilities and partnerships; Product development; Sustainable business model; Marketing to insufficiently active; Attracting the insufficiently active; Clubs and volunteers; and Shifting traditional sport culture – was considered most appropriate. Participants rated the Deliverers challenges as the most important (mean=3.52), and the Marketing to insufficiently active challenges as the easiest to overcome (2.72).

Conclusions: Key ingredients to successfully developing and delivering non-traditional sport opportunities for insufficiently active populations are: recruiting appropriate product deliverers; building the capacity of delivery organisations and systems; and developing products relevant to the delivery context that align with the needs and characteristics of the target population.

Implications for public health: A system-wide response is required to address the challenges associated with sport organisations developing, scaling and delivering innovative social sport products for insufficiently active populations.

Key words: concept mapping, sport organisations, insufficiently active, social sport products, physical activity

SPORT


Staley et al. Article

want sport opportunities that fit in with their lifestyle, rather than having to fit their lifestyle around sport.13

Sport organisations need to provide opportunities that appeal to insufficiently active people, reflect the shift towards more flexible, social offerings, and can leverage the established latent sport delivery infrastructure and systems if they are to achieve government goals of addressing population-level physical inactivity through sport. This paper explores the challenges that a group of sport organisations experienced when they developed and delivered new products, or scaled existing social versions of their sport, for insufficiently active people. It focuses on 21 Victorian State Sporting Associations (SSAs), which are the state governing bodies for sports in the Australian state of Victoria and nine Victorian Regional Sports Assemblies (RSAs), which are organisations located in regional areas of the state that are responsible for supporting community sport and recreation groups within their region. These bodies were funded by VicHealth between 2015-18 through the SSP and RSP, respectively, to develop and deliver – or scale – social sport products to insufficiently active people.

The number and type of products developed or scaled for SSP (range 1–10 products per organisation) and RSP (range 10–38 products per organisation) varied across the 30 funded organisations. They varied in terms of: 1) delivery models (from set session times and season lengths requiring commitment by participants to attend all sessions, to come-and-try days or pay-as-you-go sessions requiring no regular commitment); 2) business models (for example, centralised delivery by paid sport staff; contracting third party deliverers such as personal trainers; paying school or community recreation facility staff to deliver social sport products in their facilities; or using volunteer coaches to deliver social sport products at community club facilities); and 3) different program activities (such as modified games of a traditional sport; skill-based sessions; fitness-based training using core elements of a sport).

For example, RSAs partnered with Netball Victoria to scale their existing Rock Up Netball products in regional locations through three delivery models: a social netball game; a netball-based training session; and a round-robin event day. Participants could just ‘rock-up’ without pre-registering or committing to

playing or training every week. By contrast, Triathlon Victoria developed TRIactive, a six-week program for beginners with an interest, but limited experience, in triathlon. Program participants trained twice a week as a group and aimed to complete a triathlon at the end of the program. For more information about the number of social products developed and scaled up by each SSA, see Supplementary Table 1 and https://www.vichealth.vic.gov.au/programs-and-projects/vichealth-state-sport-program

Research findings support exploring and promoting physical activity participation from a system or ecological perspective.16 However, most physical activity research has focused on individual participants,17,18 and the relatively small amount of institutional- or organisational-level research has been conducted in school, community and healthcare settings,18 with an emphasis on environmental and policy-based interventions.17 To the best of our knowledge, no previous research has been published on the promotion of physical activity at the organisational level from the perspective of sport organisations that develop and deliver social sport products, and have a focus on increasing participants’ physical activity levels.

Although most published physical activity research has explored the frequency, patterns, correlates or predictors of physical activity,19 more recent studies explore the barriers and facilitators to implementing physical activity interventions.20–26 This research was conducted across a range of settings including schools,25 youth-serving organisations,20 and healthcare settings.26 To date, no previous published research has explored the barriers or facilitators to developing and implementing physical activity interventions in community sport settings.

To fill these gaps in the extant literature, and help address the paucity of research examining how informal participation fits with traditional sport development structures and systems,27 this research explored the challenges that sport organisations experienced when developing and delivering social sport products to engage insufficiently active people in regular physical activity. The findings of this study can be used to leverage the considerable infrastructure and resources already invested in sport organisations, to respond to changing trends in physical activity participation, and to tackle the

growing proportion of the population who are insufficiently physically active in Australia.

Methods

As a component of the process evaluation of VicHealth’s investment in the SSP and RSP, we were interested in integrating the applied knowledge of practitioners (i.e. of sports that developed and delivered social sport products) with the scientific knowledge of researchers and policy makers (i.e. VicHealth). Therefore, we employed Concept Mapping (CM), a mixed-method participatory approach to gather qualitative data and analyse it quantitatively.28,29 The key CM steps of preparation, ideas generation (brainstorming), statement structuring (sorting and rating), and concept mapping analysis, are described in detail elsewhere.30 We used the Concept Systems Global MAX™31 web platform to undertake this study.

Sample selection and recruitmentIn mid-March 2018, we invited multiple contact people (total N=70: RSA=32, range=2–7 per organisation; SSA=38, range=1–4 per organisation) from each of the 30 organisations funded through the RSP/SSP to participate in the CM exercise. All participants were identified by VicHealth as integral to the development and delivery of the social sport products within their funded organisation. The emailed invitations were sent to all participants simultaneously and included a hyperlink to the online brainstorming. Several reminder emails were sent to all potential participants before the ideas generation step closed after 14 days.

Before undertaking their first CM activity, participants provided online consent (implied by self-registering to participate in the study), and were asked to describe: their gender; the type of sport organisation they worked for; how long they had worked for the organisation; their position within the organisation; and how long they had been working on developing and delivering social sport products funded through the RSP/SSP. All background questions were categorical with multiple choice responses.

Data collectionThe focus prompt used to brainstorm ideas in this study was: “Based on your experiences of the RSP/SSP, what challenges are there to designing, developing and delivering a


Sport Challenges engaging insufficiently active people

successful program to engage inactive and somewhat active people in sport or physical activity?” The two rating instructions used were: “On a scale from 0 (least important) to 5 (most important), how important is overcoming this challenge for program success?” and “On a scale from 0 (hardest) to 5 (easiest), how easy is this challenge for your organisation to overcome?”

We asked participants to brainstorm as many single-thought statements as they could in response to the focus prompt (see above). As is usual practice in CM studies, participants could review the statements other participants made, and access the online platform multiple times.

After the brainstorming had been completed, the authors (KS and AD) conducted multiple rounds of synthesising and editing the brainstormed statements to: delete statements unrelated to the focus prompt; split compound statements; identify statements that represented the same idea, and select the statement that best captured the essence of the idea; and edit statements to reflect an agreed meaning. This iterative process involved all members of the research team and continued until there was consensus that the final statement list contained a manageable (i.e. not so many statements that participants would be unwilling to sort and rate them all) set of unique (i.e. each idea was represented once), clear and pertinent ideas. We cross-referenced the final and original sets of statements to ensure all relevant brainstormed ideas were represented in the final set of statements.

We invited all RSA and SSA contacts (N=70) to participate in the statement structuring, even if they had not participated in the brainstorming. Multiple reminder emails were sent to anyone who had not responded or completed the sorting and rating tasks over 14 days in early May 2018.

During the statement structuring process, each participant sorted the randomised synthesised statements into groups that made sense to them. They were instructed to group statements according to similarity in meaning, and to name each group based on its theme or contents. Participants could create single-statement groups if they thought a statement was unrelated to all other statements. They were asked to put every statement somewhere, and to avoid creating ‘miscellaneous’ or ‘other’ groups. They were also informed that 5 to 15 groups usually work well to organise the number

of statements they were asked to sort. Participants were also instructed to rate each challenge on ‘importance’ and ‘ease of overcoming’, using the full six-point scale (0–5), relative to the other challenges in the list.

Data analysisDuring data analysis, we created a square symmetric similarity matrix from the sorted data, before applying two-dimensional non-metric multidimensional scaling to locate each statement as a separate point on an X–Y ‘point map’. We then used hierarchical cluster analysis to partition the point map into groups of statements creating a ‘cluster map’. A detailed description of the multidimensional scaling, including the stress index calculation, and hierarchical cluster analysis used in the Concept Systems Global MAX™31 web platform, is available from Kane and Trochim (pp. 87–100).28 We also calculated mean importance and ease of overcoming ratings for each statement, and used them to generate a ’go-zone’ graph, in which we plotted each statement’s mean ratings on a graph divided into four quadrants using the overall mean of each rating as the axes.

To select the most appropriate number of clusters, the research team followed Kane and Trochim’s recommended process,28(pp101-103) examining the cluster maps for a 6-cluster solution through to a 12-cluster solution and paying particular attention to which clusters of statements were split as the number of clusters increased. This negotiated process was used to identify the cluster level that the research team believed retained the most useful detail between clusters, while merging those clusters that seemed to logically belong together. After agreeing on the most appropriate cluster level, statements that subjectively seemed to belong in an adjacent cluster were identified and reassigned to the more appropriate neighbouring cluster.32

Ethics approval for this study was given by the Human Research Ethics Committee of La Trobe University (Application Number: E15-081 Modification).

Results

Sixty-eight individual participants from 28 of the 30 funded organisations contributed CM data: 57 in the ideas generation, 55 in the statement sorting, 60 in the importance

rating, and 57 in the ease of overcoming rating. Forty-three participants contributed data in all phases, while three contributed to the ideas generation only.

Thirty-eight participants represented 19 SSAs (mean 2.0 participants per organisation, range 1–4, mode 2) while 30 participants represented nine RSAs (mean 3.3, range 2–6, mode 2 and 4). Just over half (53%) of participants were male, and more than three-quarters described their position as a program coordinator (60%) or an executive officer (19%). Half (50%) of the participants had been employed with their current organisation for three years or longer, and nearly three-quarters (72%) of participants had worked on the RSP/SSP program for 12 months or longer. Full details of the participants are available in Supplementary Table 2.

The participants brainstormed 158 challenges in response to the project focus prompt. The research team synthesised and edited these to 71 unique challenges for participants to sort and rate (Table 1). Fifty-five participants sorted the 71 challenges into groups (mean=7.65 groups; mode=7 groups (11 participants); range 4–12 groups).

The mean importance rating for all challenges was 3.34 out of 5 (Table1). Challenges in Cluster 1 (Deliverers) were rated the most important (3.52), while those in Cluster 9 (Shifting traditional sport culture) were rated the least important (2.83). The mean ease of overcoming rating for all challenges was 2.40. Challenges in Cluster 6 (Marketing to insufficiently active) were rated the easiest (2.72), and those in Cluster 5 (Sustainable business model) the hardest (1.98) to overcome.

The research team agreed that a 9-cluster solution: Deliverers (10 challenges); Capacity to drive the product (6 challenges); Facilities and partnerships (5 challenges); Product development (12 challenges); Sustainable business model (10 challenges); Marketing to insufficiently active (5 challenges); Attracting the insufficiently active (9 challenges); Clubs and volunteers (12 challenges); and Shifting traditional sport culture (2 challenges) retained the most useful detail while merging those clusters that seemed to logically belong together (see Figure 1). The distances between the individual points on the cluster map (Figure 1) represent the degree of similarity between challenges (i.e. the challenges grouped together by


Table 1: Statements generated during the concept mapping brainstorming process including the cluster in which each statement fits, mean importance and ease of overcoming ratings and go-zone graph quadrants for each statement.

Mean rating Go -zone quadrantb

Importancea Ease of overcominga

All statements

Within cluster

Cluster 1: Deliverers 3.52c 2.61c

5 Finding a deliverer who can engage with your target market. 4.29 2.64 1 119 Finding the right deliverers with the capacity (time, skill, space). 4.28 2.24 2 225 Existing providers/delivers are stuck in their ways and won’t adapt to change. 3.03 2.51 3 429 Educate existing providers/deliverers on the formalities of the product. 3.18 3.12 3 335 Understanding the need to get product deliverers involved in resource and product development. 3.13 2.83 3 343 Getting our providers to run the product in the designed way. 3.17 2.92 3 346 Finding an appropriately skilled deliverer that can engage the target market at a cost that suits. 3.92 2.25 2 249 Convincing deliverers to take a risk and do something differently (‘break the rules’/challenge existing structures across all levels). 3.22 2.49 3 459 Engaging deliverers that understand the barriers to participation. 3.88 2.81 1 171 Developing models that don’t require trained accredited facilitators/coaches. 3.12 2.31 4 4Cluster 2: Capacity to drive the product 3.49c 2.42c

2 Identifying clubs with the capacity to ensure product sustainability. 3.89 2.90 1 18 Community sport organisations not understanding the role of the SSA/RSA in the implementation of the product. 2.67 3.00 3 352e Challenges around staff turnover and losing momentum because product development takes time; research, development, testing,

recruitment, retesting, sustainability.3.25 2.10 4 4

54 The recruitment of participants into products is highly dependent on the quality of partnering ‘gatekeepers’ (agencies and organisations) who introduce their members into products.

3.45 2.51 1 3

56 Finding local drivers to ensure products are sustained. 4.02 2.12 2 267 Ensuring that clubs persist with a product and do not become discouraged with a slow start or low initial interest. 3.72 1.95 2 2Cluster 3: Facilities and partnerships 3.43c 2.33c

13 Managing expectations of partners. 3.02 2.92 3 318 Developing sustainable supports around the delivery of the product (i.e. Councils, local business, other sports, schools). 3.89 2.00 2 233 Ensuring products continue as RSA/SSA involvement is withdrawn. 4.14 1.54 2 263 Facility access was a huge barrier as the traditional model of our sport takes priority. 2.92 2.49 3 368 Access to adequate venues (e.g. with lights). 3.20 2.75 3 3Cluster 4: Product development 3.40c 2.62c

3 Making sure the product is different enough from your usual offerings whilst not losing what the sport is all about. 3.02 3.20 3 37 Creating a product that people with little to no interest in sport/rec, who have sometimes had bad experiences, find interesting,

enticing, and safe.4.11 2.00 2 2

17 Appropriate time to consult with communities and to then implement products around a variety of needs/expectations. 3.67 2.19 2 222 Designing a flexible product that caters for degrees and types of disability, and individual capability and capacity. 3.67 2.63 1 124 Developing products to suit the different regions. When each community/town is different in what is available and the people who

work within it.3.16 2.75 3 3

28 Implementing new and/or adapting existing administration systems for social participation products. 2.67 3.07 3 330 Developing a Social Participation Strategy that includes a social player/team pathway. 2.79 2.62 3 336 Ensuring products that are social are also flexible (time, cost, insurance, membership, scheduling, competitiveness). 3.90 2.44 1 244 Ongoing engagement/feedback with the market when designing/developing the product. 3.73 2.85 1 150 Ensuring the product is simple and easy to understand for someone new to the sport. 3.88 3.19 1 151 Creating something truly original and engaging in an already crowded health and fitness marketplace. The average person is now

so much more adept at ‘self-exercising’ and has the YouTube world as their oyster.3.23 2.12 4 4

53 Having an infrastructure that is based on competition makes it challenging to provide the ongoing participation for those wanting casual and recreational opportunities.

2.98 2.34 4 4

Cluster 5: Sustainable business model 3.40c 1.98c

1d The balance between the need for immediate results and sustainable long term participation (e.g. time to research the needs of the community and design a product that fits that need).

3.82 2.20 2 1

6d Engaging inactive and somewhat active people to participate, while ensuring a sustainable business model. 3.79 1.56 2 214 Insurance is tied to traditional sport products/models, and there is little flexibility with the majority of insurances to allow casual or

flexible participation.2.24 2.88 3 3

15 Building a critical mass for the running of, and social enjoyment of an activity. 3.15 1.88 4 437 The need for different Business/Revenue models (depending on location, facility owner, host club/council/etc.). 3.15 2.53 3 339 Products’ and ‘Programs’ alone are not enough... we need to address systemic issues as a sector (e.g. cost, access, attitudes etc.). 3.47 1.68 2 241 Transitioning inactive or somewhat active people from introductory into ongoing participation (e.g. club products). 3.60 1.44 2 2



Table 1 (cont.): Statements generated during the concept mapping brainstorming process including the cluster in which each statement fits, mean importance and ease of overcoming ratings and go-zone graph quadrants for each statement.

Mean rating Go -zone quadrantb

Importancea Ease of overcominga

All statements

Within cluster

45 Ensuring sufficient numbers at products when offering a flexible attendance policy creates uncertainty and impacts on a successful product.

3.60 1.73 2 2

65d Finding the right cost structure to enable sustainable delivery post a funded pilot/trail, particularly in low SES & isolated towns. 3.58 2.03 2 169d Sustainability - transition from free or low cost to full fees (needs to be considered upfront and communicated during initial

products).3.48 2.05 2 1

Cluster 6: Marketing to insufficiently active 3.34c 2.72c

27 Developing a social marketing strategy that showcases the concept in its new social non-competitive format to entice the participant.

3.18 2.90 3 3

32 Ensuring the value proposition (benefits/sales pitch) is appropriate to the market. 3.67 2.68 1 240 Designing a marketing strategy targeting the inactive cohort (where relevant the strategy supports member clubs). 3.46 2.24 2 260 A lack of marketing expertise within the project team. 2.93 3.14 3 361 Resources (budget allocation) for marketing and promotion to create an ongoing awareness and participant recruitment campaign. 3.47 2.63 1 2Cluster 7: Attracting the insufficiently active 3.21c 2.14c

11 Participant lack of confidence and self-esteem. 3.05 2.42 3 316 At times, the products attracted the ‘sports engaged’ to play more sport. 2.51 2.83 3 320 Hard to get participants to commit to coming on a regular basis - regardless if it was free or low cost. 3.61 1.56 2 223 Consulting with the less active cohort to understand their motivations. 3.89 2.19 2 138 A modified form of a traditional sport is of no interest to this cohort because of their previous personal experiences with sport and

clubs.2.68 2.44 3 3

42 Changing the mindset of the target participant on what ‘sport’ is to include social sport. 3.57 1.92 2 247 Mobilising the inactive/someone active target market ... getting them from ‘contemplating’ physical activity to actually showing up. 3.95 1.25 2 258 As a lot of “Active” players are already engaged in this product, it can be very confronting and competitive for a “non-active” player.

They might attend 1 session but not return the following week.3.02 2.54 3 3

66 Too big a step for participants, from nothing to sport-based activity. 2.62 2.27 4 3Cluster 8: Clubs and volunteers 3.18c 2.42c

4 Club people are not the right people to engage with this cohort. 2.61 2.88 3 39 Limited volunteer base. 3.31 2.22 4 212 Making sure it is not overly time intensive to organise for volunteer clubs. 3.59 2.36 2 221 Lack of club involvement or engagement in modified sports. 3.15 2.25 4 426 Limited expertise, knowledge and skill of volunteers. 2.84 2.71 3 331 Individuals in leadership positions, rather than the whole club/committee agree the club will deliver the product, but don’t support

club members to implement.3.08 2.47 3 3

34 Many clubs/associations struggle to implement their ‘core business’ (eg. field teams and committee roles). Therefore implementing social and modified sport products is not on their agenda.

3.52 1.66 2 2

48 Club volunteers being targeted by multiple organisations with various priorities; when the priority is club admin, compliance, then everything else.

3.17 1.92 4 4

55 Volunteers struggled to see the benefits of social and modified sport to the club. 2.78 2.42 3 362 Engaging clubs/deliverers in the vision. 4.02 2.53 1 164 Getting the clubs across the state to deliver the same product. 2.37 2.80 3 370 Identifying club characteristics required for them to have the capacity to deliver the product successfully. 3.68 2.95 1 1Cluster 9: Shifting traditional sport culture 2.83c 2.43c

10 We’re not geared to deliver to the inactive / somewhat inactive as our membership base is traditionally focused on current participants / active people.

2.59 2.54 3 3

57 Shifting the thinking of the sport community that the social products are not there to recruit people to our traditional formats. 3.10 2.32 4 2For all statements 3.34c 2.40c

Notesa: 0 (least important/hardest to overcome) to 5 (most important/easiest to overcome); ^60 participants rated all 71 statements for importance and 57 rated all 71 statements for ease of overcomingb: Go Zone Quadrants: 1 (Top right = above mean for both importance and ease); 2 (Bottom right = above mean for importance and below mean for ease); 3 (Top left = below mean for importance and above mean for ease); 4 (Bottom left =

below mean for both importance and ease)c: mean importance/ease rating for all the statements in the clusterd: Reassigned from Cluster 1 to Cluster 5e: Reassigned from Cluster 7 to Cluster 6.



had worked for at least 12 months, and 50% for more than two years on the RSP/SSP) who represented a wide variety of sport organisations (n=28). In addition, employing Concept Mapping enabled participants to both identify issues of interest and group them together. This overcomes some of the methodological limitations associated with previous research investigating barriers and facilitators to implementing physical activity interventions, which employed either surveys (in which participants rated their level of agreement with researcher-selected items)23 or semi-structured interviews (in which participants’ responses to open-end questions were coded and grouped by researchers).21

The nine clusters of challenges to developing and delivering social sport products for insufficiently active populations identified in this study span the dimensions encompassed by commonly cited ecological health promotion models.34–36 Cluster 7 highlights the challenges related to individual participants, while Clusters 1 and 4 focus attention on the challenges involved in product development and delivery. Clusters 2,

2732

18

33

3

17

44

51

53

11

2023

42

14

15

41

45

165

8

52

57

12

21

31

6264

35

71

6. Marketing to insufficiently active

7. Attracting the insufficiently active

4. Program development

1. Deliverers

9. Shifting traditional sport culture5. Sustainable

business model

3. Facilities and partnerships

2. Capacity to drive the program

8. Clubs and volunteers

48 9 554

26

70

34

525

1946

29 59

43

49

267

56

54

28

30 24

36

2250

1638

47

61

60

40

39

6

69

13

63

68

37

10

7

58

66

more participants are located closer to each other on the map). For example, challenges #9, #31 and #55 were considered so closely related that nearly all participants grouped them together. By contrast, challenges #61 and #71 were considered so unrelated that almost no one grouped them together. The stress index – a representation of how well the two-dimensional map reflects the square symmetric similarity matrix generated from the sorted data – was 0.24, close to the average stress value across a broad range of CM projects.28 A full list of the challenges within each cluster, including the five challenges that were reassigned to neighbouring clusters to which there was a better conceptual fit, is provided in Table 1.

Figure 2 is a go-zone graph for all 71 challenges. The ‘go-zone’ quadrant of challenges in the top right contains the 12 challenges that were rated above average on both importance and ease of overcoming. The go-zone graph quadrant for each challenge (when all challenges and when challenges within the same cluster only are considered) is provided in Table 1. To aid interpretation of the go-zone graph, see

Table 1 for the details of each challenge, including its mean importance and ease of overcoming ratings.

Discussion

This study is the first published investigation of the challenges faced by sport organisations when they attempt to develop and deliver – or scale – innovative, social sport products to increase physical activity among insufficiently active populations. The findings are internationally relevant, particularly in countries where trends are shifting towards more flexible and social participation in sport and physical activity (e.g. the United Kingdom and the United States),15 and where sport development and delivery systems are similar to the Australian federated, multi-tiered, community sport-based system (e.g. Canada and New Zealand).33

A key strength of this study is that it draws on the reflections of three years of developing and delivering social sport products for insufficiently active populations. The study gathered data from a large number of people (n=68) with considerable experience (75%

Figure 1: A nine-cluster map of challenges to sport organisations developing and delivering social sport products for insufficiently active populations.

Note: Dashed lines indicate clusters before statement reassignment.



3, 5, 6 and 8 identify the challenges associated with organisational capacity and leveraging partnerships to develop, attract participants to, and sustain the delivery of new, social sport products, while Cluster 9 raises broader challenges related to the culture and traditions of sport in the community. In addition, at least three challenges in nearly every cluster (except Cluster 9) are located on the right-hand side of the go-zone graph (see Figure 2), indicating that they were rated above the mean of 3.34 out of 5 for importance in overcoming to ensure product success. These findings support previous calls for multi-strategy ecological approaches to promote physical activity,16,37 and highlight the need for change at all levels of the sports governance system to maximise participation in non-traditional sports in traditional sports settings.27

While acknowledging the need for a system-wide response to the challenges associated with developing and delivering social sport products for insufficiently active populations, recruiting appropriately skilled and experienced product deliverers is clearly a key challenge and an opportunity to influence program success. Not only was the Deliverers cluster rated as the most

important to overcome for program success (mean=3.52/5), but the two challenges rated as the most important to overcome (Statements #5 and #19), were both located in this cluster. In addition, the Deliverers cluster was rated as the third easiest (mean=2.61/5) for the participating organisations to overcome. These findings suggest that sport organisations interested in or tasked with designing, developing and delivering or scaling social products to inactive people should make sure they recruit deliverers with the capacity to deliver social sport products, as well as the ability to engage with – and understand the participation barriers for – the target population. Such organisations should consider producing detailed position statements to facilitate the recruitment of appropriate product deliverers, as well as developing comprehensive orientation/training programs to ensure all recruited deliverers have the knowledge and skill sets required.

The other two relatively important and easy to address clusters of challenges that emerge from this study are those related to Product development (mean=3.40 for importance and 2.62 for ease of addressing) and Marketing to the insufficiently active (mean=3.34 for

importance and 2.72 for ease of overcoming). The challenges within the Product development cluster highlight the need to ensure the social sport products being offered by sports organisations are developed in consultation with potential participants and deliverers, meet the needs of the target population (which vary across geographical locations and sub-populations) and reflect an understanding of the delivery context, and can be delivered using existing systems, resources and infrastructure. This approach is supported by well-cited health promotion planning frameworks.38,39 The challenges within the Marketing to insufficiently active cluster highlight a need to improve the marketing of social sport products through a combination of resourcing and upskilling existing staff, allocating time and resources to developing appropriate marketing strategies, and recruiting staff with specific marketing expertise and an understanding of the target population.

Mapping the nine clusters and 71 challenges that emerged from this study onto the five-domain, 72-construct framework of common barriers to implementing and scaling up physical activity interventions developed by Koorts and colleagues32 reveals that the

Figure 2: Go zone of challenges to sport organisations developing and delivering social sport products for insufficiently active populations.



outcomes of this study are well supported by highly cited implementation science explanatory frameworks. For example, Cluster 1 (Delivers) aligns very strongly with the Koorts and colleagues ‘Implementer characteristics’ domain; Cluster 2 (Capacity to drive the product) and Cluster 3 (Facilities and partnerships) align strongly with their ‘Delivery setting’ domain; and Cluster 4 (Product development) aligns strongly with their ‘Intervention characteristics’ domains.36 In addition, Cluster 5 (Sustainable business model), Cluster 6 (Marketing to insufficiently active), Cluster 7 (Attracting the insufficiently active), Cluster 8 (Clubs and volunteers) and Cluster 9 (Shifting traditional sport culture) represent an amalgam of the constructs contained in Koorts and colleagues ‘Community characteristics’ and ‘Process of implementation’ domains.

The challenges identified in this study also reflect previously identified barriers and facilitators to implementing physical activity interventions across a range of settings, and using a variety of methods.21,22,25 For example, the key factors that experts identified as influencing the implementation of physical activity interventions in youth-serving organisations20 – including: available facilities, equipment, space and staff; competing programs; engaging intervention staff; provider belief, motivation, knowledge and skills about the intervention; and [program] adaptability – all align closely with the more important challenges (i.e. those in Clusters 1 to 4) identified in the current study. In addition, the challenges that emerged in this study, particularly in Cluster 5 (sustainable business model), Cluster 8 (Clubs and volunteers) and more specifically in Cluster 9 (Shifting traditional sport culture) reflect a previously identified need for cultural change within sports organisations to broaden the understanding of how sport is structured and the forms of sport that should be facilitated and prioritised.27

Concept Mapping, similar to other qualitative research methods, has methodological limitations associated with the reliability, validity and generalisability of the findings due to non-random sampling, small sample size and over reliance on the researchers’ skills.40 In this study, the project team used their subjective judgement to synthesise and edit the brainstormed statements, to select the number of clusters that most appropriately reflected the analysed data, and to reassign some statements to

neighbouring clusters. Therefore, despite the project team following standard CM protocol,28 a similar study involving the same participants but conducted by a different project team may produce different results. In addition, as this was a component of the process evaluation of VicHealth’s investment, it was beyond the scope of this study to examine the effectiveness of the social sport products developed through the SSP and RSP. However, an impact evaluation to establish the effectiveness of social sport products in increasing physical activity participation among insufficiently active populations is currently being conducted and will be reported separately.

Conclusion

Understanding the challenges that sport organisations experience when developing and delivering new products or scaling existing flexible, social sport products to engage insufficiently active people is an important first step in supporting these organisations to undertake this task. Using Concept Mapping in the previously unexplored context of sport organisations, this study provides a real-world example of the importance of applying a multi-strategy, ecological approach to developing and delivering physical activity interventions. It also highlights that recruiting appropriate product deliverers, building the capacity of delivery organisations and systems, and developing social sport products relevant to the delivery context, which align with the needs and characteristics of the target population, are key ingredients to the successful development and delivery of sport opportunities for insufficiently active populations.

References1. Australian Institute of Health and Welfare. Australia’s

Health 2008. Catalogue No.: AUS 99. Canberra (AUST): AIHW; 2008.

2. Lee IM, Shiroma EJ, Lobelo F, Puska P, Blair SN, Katzmarzyk PT. Effect of physical inactivity on major non-communicable diseases worldwide: An analysis of burden of disease and life expectancy. Lancet. 2012;380:219–29.

3. Guthold R, Stevens GA, Riley LM, Bull FC. Worldwide trends in insufficient physical activity from 2001 to 2016: A pooled analysis of 358 population-based surveys with 1·9 million participants. Lancet Glob Health. 2018;6(10):Pe1077-e1086.

4. Australian Bureau of Statistics. 4364.0.55.004 - Australian Health Survey: Physical Activity, 2011–12 [Internet]. Canberra (AUST): ABS; 2013 [cited 2018 Sep 27]. Available from: http://www.abs.gov.au/AUSSTATS/[email protected]/Latestproducts/4364.0.55.004 Main%20Features12011-nt&tabname=Summary&prodno=4364.0.55.004&issue=2011-12&num=&view=

5. Australian Institute of Health and Welfare. Impact of Physical Inactivity As a Risk Factor for Chronic Conditions: Australian Burden of Disease Study [Internet]. Australian Burden of Disease Study Series No.: 15. Canberra (AUST): AIHW; 2017 [cited 2018 Sep 27]. Available from: https://www.aihw.gov.au/reports/burden-of-disease/impact-of-physical-inactivity-chronic-conditions/contents/summary

6. Australian Department of Health. Australia’s Physical Activity and Sedentary Behaviour Guidelines and the Australian 24-Hour Movement Guidelines [Internet]. Canberra (AUST): Government of Australia; 2019 [cited 2019 May 30]. Available from: http://www.health.gov.au/internet/main/publishing.nsf/Content/health-pubhlth-strateg-phys-act-guidelines#npa1864

7. Australian Department of Health. Sport 2030: Participation Performance Integrity Industry [Internet]. Canberra (AUST): Government of Australia; 2018 [cited 2018 Sep 27]. Available from: https://www.ausport.gov.au/nationalsportplan/home/second_row_content/have_a_say2/Sport_2030_-_National_Sport_Plan_-_2018.pdf

8. VicHealth. Physical Activity Strategy 2018–23 [Internet]. Melbourne (AUST): Victorian Health Promotion Foundation; 2018 [cited 2018 Sep 27]. Available from: https://www.vichealth.vic.gov.au/media-and-resources/publications/physical-activity-strategy

9. King K, Church A. Lifestyle sports delivery and sustainability: Clubs, communities and user-managers. Int J Sport Policy Politics. 2017;9(1):107–19.

10. VicHealth. Physical Activity, Sport and Walking: VicHealth’s Investment Plan (2014–2018) [Internet]. Melbourne (AUST): Victorian Health Promotion Foundation; 2014 [2019 May 30]. Available from: https://www.vichealth.vic.gov.au/-/media/ResourceCentre/PublicationsandResources/Physical-activity/PhysActivity-InvestmentPlan-2014-18.pdf?la=en&hash=F87588A487798622619F6F80064473C740FD0DCC

11. Borgers J, Pilgaard M, Vanreusel B, Scheerder J. Can we consider changes in sports participation as institutional change? A conceptual framework. Int Rev Sociol Sport. 2018;53:84–100.

12. Laakso L, Telama R, Nupponen H, Rimpelä A, Pere L. Trends in leisure time physical activity among young people in Finland, 1977–2007. Eur Phy Educ Rev. 2008;14:139–55.

13. Hajkowicz S, Cook H, Wilhelmseder L, Boughen N. The Future of Australian Sport: Megatrends Shaping the Sports Sector Over Coming Decades [Internet]. Melbourne (AUST): CSIRO; 2013 [cited 2018 Sep 27]. Available from: https://www.sportanddev.org/sites/default/files/downloads/the_future_of_australian_sport___full_report.pdf

14. Eime RM, Sawyer N, Harvey JT, Casey MM, Westerbeek H, Payne WR. Integrating public health and sport management: Sport participation trends 2001–2010. Sport Manage Rev. 2015;18:207–17.

15. Gilchrist P, WheatonB. The social benefits of informal and lifestyle sports: A research agenda. Int J Sport Policy Polit. 2017;9(1): 1–10.

16. Spence JC, Lee RE. Toward a comprehensive model of physical activity. Psychol Sport Exerc. 2003;4:7–24.

17. Bauman AE, Reis RS, Sallis JF, Wells JC, Loos RJF, Martin BW. Correlates of physical activity: Why are some people physically active and others not? Lancet. 2012;380:258-71.

18. Golden SD, Earp JAL. Social ecological approaches to individuals and their contexts: Twenty years of health education and behavior health promotion interventions. Health Educ Behav. 2012;39:364–72.

19. Milat AJ, Bauman AE, Redman S, Curac N. Public health research outputs from efficacy to dissemination: A bibliometric analysis. BMC Public Health. 2011;11:934.

20. Lau EY, Wandersman AH, Pate RR. Factors influencing implementation of youth physical activity interventions: An expert perspective. Transl J Am Coll Sports Med. 2016;1(7):60–70.

21. Mâsse LC, Naiman D, Naylor P-J. From policy to practice: Implementation of physical activity and food policies inschools. Int J Behav Nutr Phys Act. 2013;10:71.



22. Weatherson KA, Gainforth HL, Jung ME. A theoretical analysis of the barriers and facilitators to the implementation of school-based physical activity policies in Canada: A mixed methods scoping review. Implement Sci. 2017;12:41.

23. Carlson JA, Engelberg JK, Cain KL, et al. Contextual factors related to implementation of classroom physical activity breaks. Transl Behav Med. 2017;7:581–92.

24. van den Berg V, Salimi R, de Groot R, Jolles J, Chinapaw M, Singh A. “It’s a battle … you want to do It, but how will you get it done?”: Teachers’ and principals’ perceptions of implementing additional physical activity in school for academic performance. Int J Environ Res Public Health. 2017;14:1160.

25. Naylor P-J, Nettlefold L, Race D, et al. Implementation of school based physical activity interventions: A systematic review. Prev Med. 2015;72:95–115.

26. Huijg JM, Gebhardt WA, Verheijden MW, et al. Factors influencing primary health care professionals’ physical activity promotion behaviors: A systematic review. Int J Behav Med. 2015;22:32–50.

27. Jeanes R, Spaaij, Penney D, O’Connor J. Managing informal sport participation: tensions and opportunities. Int J Sport Policy Polit. 2019;11(1):79–95.

28. Kane M, Trochim WM. Concept Mapping for Planning and Evaluation. Vol 50. Thousand Oaks (CA): Sage Publications; 2007.

29. van Bon-Martens MJH, van de Goor LAM, Holsappel JC, et al. Concept mapping as a promising method to bring practice into science. Public Health.128:504-14.

30. Trochim WM, McLinden D. Introduction to a special issue on concept mapping. Eval Progam Plann. 2017;60:166–75.

31. Concept Systems Incorporated. Concept Systems Global Max [Internet]. Ithaca (NY): CSI; 2017 [cited 2018 Sep 27]. Available from: https://www.conceptsystems.com/gw/software

32. Mannes M. Using concept mapping for planning the implementation of a social technology. Eval Program Plann. 1989;12:67–74.

33. Stewart B, Nicholson M, Smith A, Westerbeek H. Australian Sport–better by Design? The Evolution of Australian Sport Policy. London (UK): Routledge; 2004.

34. Sallis JF, Cervero RB, Ascher W, Henderson KA, Kraft MK, Kerr J. An ecological approach to creating active living communities. Annu Rev Public Health. 2006;27(1):297–322.

35. McLeroy KR, Bibeau D, Steckler A, Glanz K. An ecological perspective on health promotion programs. Health Educ Q. 1988;15(4):351–77.

36. Koorts H, Eakin E, Estabrooks P, Timperio A, Salmon J, Bauman A. Implementation and scale up of population physical activity interventions for clinical and community settings: The PRACTIS guide. Int J Behav Nutr Phys Act. 2018;15:51.

37. Sallis JF, Owen N, Fisher EB. Ecological models of health behavior. In: Glanz K, Rimer BK, Viswanath K, editors. Health Behavior and Health Education: Theory, Research, and Practice. 4th ed. San Francisco (CA): Jossey-Bass; 2008. p. 465–86.

38. Green LW, Kreuter M. Health Program Planning: An Educational and Ecological Approach. 4th ed. New York (NY): McGraw-Hill; 2005.

39. Bartholomew LK, Parcel GS, Kok G, Gottilieb NH, Fernandez ME. Planning Health Promotion Programs. An Intervention Mapping Approach. 3rd ed. San Francisco (CA): Jossey-Bass; 2011.

40. Burke J, O’Campo P, Peak G, Gielen A, McDonnell K, Trochim W. An introduction to concept mapping as a participatory public health research methodology. Qual Health Res. 2005;15:1392-4.



Supplementary Table 1: Funded organisations and examples of products supported by VicHealth’s State and Regional Sport Programs.

Supplementary Table 2: Characteristics of participants (n= 68).



Traumatic brain injury (TBI) is an important cause of preventable mortality and disability across the

lifespan. By 2030, brain injuries due to traffic accidents and falls are expected to rise to the 7th and 17th major cause of death, respectively.1 TBI ranges from mild to extremely severe injury. Recovery levels are similarly variable depending on the severity of the initial injury, from good outcome and resumption of premorbid lifestyles to profound disability affecting physical, cognitive and/or behavioural function. Although it is a critical public health problem worldwide, the actual incidence of TBI is difficult to establish.

The latest systematic review indicates a pooled annual incidence proportion of 295/100,000 (95% CI, 274-317) for all ages.2 Incidence data, however, still vary widely across countries and among studies,3 with differences in study methodology contributing most notably to this variability. Injury patterns are also changing, showing that TBI incidence is increasing in low-income countries and more injuries are occurring among older people in high-income countries,4 which makes it challenging for estimates to be generalised.

The incidence of TBI in Australia is not well established. From the cross-continental comparison, Australasia (including Australia and New Zealand data) yielded the highest incidence proportion of 415/100,000 population.2 This finding was of particular concern, together with the fact that in Australia only one study5 met inclusion

criteria for meta-analysis, reporting on incidence of TBI in 1988. That study showed much lower rates of 100/100,000 population in a defined New South Wales (NSW) community.5 Other available Australian data confirmed TBI is less common in Australia than in Europe and North America (228-331/100,000) or New Zealand (790-1750/100,000).2 These data included findings from the Australian Institute of Health and Welfare study for the period 2004-05 (107/100,000 population)6 and a Western Australia study for 2003-08 (85.8/100,000 population).7

Differences in case identification and study design are the major limitations to current epidemiological research.8 Evidence strongly suggests population-based studies are the best approach to obtain objective estimates and understand epidemiological patterns of disease.8 Moreover, while injury patterns in Australian sectors have been previously investigated,9-12 little is known about characteristics for the whole population. The need for population studies has become even more imperative in developed countries, including Australia, where TBI patterns have changed over the past decades, with

Epidemiology of hospitalised traumatic brain injury in the state of New South Wales, Australia: a population-based studyIlaria Pozzato,1 Robyn L Tate,1 Ulrike Rosenkoetter,1 Ian D Cameron1

1. John Walsh Centre for Rehabilitation Research, Sydney Medical School - Northern, Kolling Institute of Medical Research, University of Sydney, Sydney, Australia.Correspondence to: Robyn L. Tate, John Walsh Centre for Rehabilitation Research, Sydney Medical School - Northern, University of Sydney, Level 12, Kolling Institute of

Medical Research, Royal North Shore Hospital, St Leonards, NSW, 2065; e-mail: [email protected]: May 2018; Revision requested: July 2018; Accepted: January 2019The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,



Abstract

Objective: To describe the population-based incidence and epidemiological characteristics of hospitalised traumatic brain injury (TBI) in New South Wales (NSW), Australia.

Methods: One-year statewide hospital admission data from the NSW Department of Health were analysed. TBI cases were identified using a combination of TBI-related diagnostic and external cause codes from the International Classification of Diseases (ICD-10th Revision). Sociodemographics, causes, associated factors, severity and medical details of hospitalisation were examined.

Results: There were 6,827 hospitalised TBI cases that met review criteria. Incidence rate was 99.1/100,000 population. Incidence in persons older than 75 years of age and residents in remote areas was three times higher. Aboriginal and Torres Strait Islander peoples were 1.7 times more likely to sustain a TBI than the general population, and risk was greater for all NSW residents from areas that were remote and disadvantaged-socioeconomically. Older adults and those with severe injuries showed prolonged hospitalisation, higher morbidity and mortality.

Conclusions: Overall TBI incidence in NSW is lower than international estimates. Nevertheless, groups with higher incidence rates and/or poor in-hospital outcomes were identified, highlighting directions for prevention and future research.

Implications for public health: There is a need for identifying risk factors/barriers and assessing the impact of recent policies on these population groups.

Key words: brain injury, epidemiology, incidence

GENERAL PUBLIC HEALTH


an ageing population resulting in higher numbers of fall-related injuries.4

Well-established characteristics of TBI include some sociodemographic variables (sex, age at injury, socioeconomic status) and injury-related factors (mechanism and circumstances of injury). Yet other reports have identified risk patterns that warrant further investigation. For example, Jamieson et al.12 found that Aboriginal and Torres Strait Islander peoples are 21 times more likely to incur a TBI due to assault, with other studies indicating a higher risk across all mechanisms of injury.13 Living in rural and remote areas of the country is another factor commonly associated with higher risk of TBI, but these aspects have only been investigated for sectors of the Australian population (e.g. Aboriginal and Torres Strait Islander peoples12 and Australian children).14 Similarly, factors like socioeconomic patterns of TBI in the Australian context and medical details of hospitalisation, such as length of stay (LOS), in-hospital mortality and associated injuries or comorbidities, have been scarcely investigated in previous studies. These factors, together with injury severity, are major determinants of outcome following TBI15 that could inform healthcare planning and economic impacts of hospital-treated TBI.

This study intended to move the research forward in these areas. The purposes were twofold: estimating the incidence of hospitalised TBI in NSW and describing epidemiological characteristics, with detailed analysis of at-risk groups, causes and factors associated with TBI, as well as severity and medical details of hospitalisation.

Method

Design and data sources This is a population-based study of hospital-treated TBI in NSW. NSW is the most populous state of Australia, located on the east coast of the country, covering a geographical area of almost 810,000 square kilometres. Hospital admission data from 166 public and private hospitals in metropolitan, rural and remote areas were obtained from the NSW Department of Health for the 2007 calendar year. The state’s resident population for the study year was 6.8 million16 and this number was used for the calculation of age- and sex-specific incidence rates. Census data from 2006 were used for comparison of socioeconomic characteristics.17 Approval to conduct the study was obtained from the NSW Population and Health Services Research Ethics Committee (HREC/08/CIPHS/56).

Study population and case selectionThe study population comprised the first admissions of residents of any age who presented with a TBI to a NSW hospital during the calendar year 2007. The International Classification of Diseases (ICD-10th revision) diagnostic codes18 from hospital admissions were used to identify potential TBI cases. The ICD-10th revision does not contain a specific rubric for TBI, therefore case identification was based on a combination of TBI-related diagnoses and external cause ICD descriptors.19 A multi-step selection process was adopted. First, admissions were restricted to cases with at least one of the following codes (Figure1): skull fracture, intracranial injury, crushing injury of the head. Second, cases with a reported period of loss of consciousness (LoC) only were matched with external cause of injury (E-Codes) to exclude LoC due to other medical reasons. All ICD descriptors used in this study are provided as supplementary material (Supplementary Table S1). To estimate TBI incidence, only first-time admissions and first-time TBI events during the study year were included in the count. In the analysis of medical details, information from multiple hospital separations relating to the same injury, i.e. inter-hospital and intra-hospital transfers, were incorporated.

Epidemiological characteristics extracted The following personal and injury-related characteristics were extracted directly from the Department of Health database or derived from the available ICD-codes (Supplementary Table S1).• Demographic information included age at

injury, gender, country of birth, postcode and Indigenous status. Participants were stratified based on the following age categories: 0-9, 10-19, 20-39, 40-69, 70+ years. Age-specific incidence was computed for five-year age intervals, with the upper interval being the 75+ years age group to allow comparison with other studies.

• Socioeconomic and geographical distribution were derived from residential postcodes. Socioeconomic status was allocated by mapping postcodes to deciles of the Australian Bureau of Statistics Index of Relative Socio-Economic Disadvantage scores.20 Remoteness was classified according to the Australian Standard Geographical Classification.21 Statistical comparisons involved the following

subgroups: for socioeconomic status, the most disadvantaged (classifications 1-3) versus medium (classifications 4-7) versus least disadvantaged (classifications 8-10);22 for location, metropolitan (classification 1) versus rural (classifications 2-3) versus remote living location (classification 4-5).

• Cause of injury was classified as follows using the ICD external cause (E-Codes): transport accidents, falls, assaults7 and other mechanisms. Cases with more than one E-Code were described as multiple mechanism TBI group and regarded as a separate group.

• Associated factors of TBI risk were ascertained from the diagnosis codes as follows: alcohol consumption, drug use, sports, recreational activities.

• Severity was assessed using ICD-code information on post-traumatic amnesia (PTA) and LoC duration. TBI cases were classified as severe (PTA ≥2 weeks and/or LoC >24 hrs); moderate (PTA 24hrs to 2 weeks and/or LoC 30 min to 24 hrs) and mild (PTA <24hrs and/or LoC <30 min).23

• Medical details of hospitalisation: in-hospital outcomes (LOS, hospital transfers, in-hospital mortality) and morbidity (injuries associated with TBI and comorbidities from ICD codes). Associated injuries were categorised as follows: other mechanical trauma, complications or other injuries.24 Comorbidities were identified according to the Charlson Comorbidity Index25 and grouped into five broad categories: cardiovascular (e.g. ischaemic heart disease, hypertension), neurologic (e.g. stroke, dementia), respiratory (e.g. chronic obstructive pulmonary disease), digestive (e.g. peptic ulcer disease, liver disease) and systemic (e.g. diabetes mellitus, obesity).

Results

Incidence rates, at-risk groups, injury-related characteristics and hospitalisation details were analysed.

Incidence of hospital-treated TBI A total of 10,175 admissions were initially identified using the ICD-codes selection criteria (Supplementary Figure S1). Of these, 2,435 admissions (23.9%) were subsequently excluded as not meeting study inclusion criteria. These consisted of 1,165 (11.4%) cases with no evidence of TBI, 368 (3.6%) non-residents of NSW, 902 (8.8%) non-acute episodes of care. Three-quarters (7,740) were

General Public Health Epidemiology of traumatic brain injury


confirmed TBI cases. A further 913 duplicate cases were excluded from the count: 879 (8.6%) required acute inpatient transfers and 34 (0.33%) sustained a second TBI during the same year, needing multiple admissions. The final study population was thus 6,827 residents of NSW with TBI admitted to a NSW hospital during 2007. The annual incidence of hospitalised TBI in NSW was estimated at 99.1/100,000 population (95%CI 96.8 – 101.5/100,000 population), with a mortality rate of 5.9/100,000 population (95%CI 5.3 – 6.5/100,000 population).

At-risk groups and causes of injury Figure 1 displays age/sex distribution and injury causes of this cohort. Risk of TBI was distinctly higher in the 15-19 years and in the 75+ years age groups, with rates in older people being three times the overall incidence (298/100,000 vs 99.1/100,000; z=9.99, p<0.001). This represents the first standout feature of the study findings. Falls caused most brain injuries in NSW (47.6%), followed by transport accidents (25.9%) and assaults (15.8%). Older people (>70 years) had by far the highest proportion of fall-related injuries (87.3%). Conversely, TBI in teenagers (10-19 years) was more commonly due to motor vehicle crashes (40.5%). Twenty-nine per cent of traffic-related TBI among teenagers were cycling injuries.

Another stand-out feature of the incidence findings was residency. Occurrence of TBI by location of residency, socioeconomic status and indigenous status are described in Table 1. The NSW TBI population was 2.8 times more likely to live in remote areas compared to the NSW general population (1.7% vs 0.6%; χ2=138.12, df=1, p<0.001), which was higher than the relative proportions living in the most socioeconomic disadvantaged areas

(32.3% vs 29.2% respectively, χ2=31.20, df=1, p<.001). In addition, there was an association between remoteness and socioeconomic disadvantage in the TBI sample (χ2=1193.54, df=4, p<0.001), and this reflected the NSW general population where there is also a correspondence between remoteness and socioeconomic disadvantage (χ2=801851.64, df=4, p<0.001) (Supplementary Table S2). Yet, the TBI sample had a higher proportion of both remoteness and disadvantage than the NSW general population (1.4% vs 0.44%, χ2=140.91, df=1, p<0.001), which was three times higher.

Similar disparities were found in other features of the data (Supplementary Table S3). The NSW group identifying as Aboriginal and Torres Strait Islander residents had a much higher proportion of remoteness and socioeconomic disadvantage (4.5%) compared to the NSW general non-Indigenous population (0.35%), a rate that was 12.8 times higher (χ2=52450.36, df=1, p<0.001). Even so, at 9.8% the NSW TBI Aboriginal and Torres Strait Islander peoples had the highest proportion of remoteness and socioeconomic disadvantage of all groups, even being twice as high as the general Aboriginal and Torres Strait Islander population in the state. Overall, risk of sustaining a TBI among Aboriginal and Torres Strait Islander peoples was 1.7 times higher compared to the NSW general population (3.8% vs 2.2%, χ2=81.09, df=1, p<0.001).

Factors associated with TBI risk Personal and environmental factors associated with TBI risk varied with severity (Table 1), age and causes (Table 2). At the time of the injury, 14.3% of the overall TBI cohort was involved in sport and leisure activities, with 40% in the 10-19 years age group. About

one in five people (18.7%) were under the influence of alcohol or recreational drugs, accounting for 30% of adults between 20 and 69 years of age. Mild injuries were more commonly associated (19.3%) with sport/leisure activities (χ2=10.17, df=2, p=0.006), while one in three people in the moderate-to-severe group was under the influence of alcohol or drugs (χ2=16.09, df=2, p<0.001). Substance use was strongly associated with assault-related TBI compared to other causes (41.2%; χ2=490.42, df=5, p<0.001), as well as with injury occurring during the weekend compared to week days (22.2%; χ2=59.73, df=1, p<0.001) (Supplementary Table S4).

Injury severity and medical details of hospitalisationSpecific injury severity data were available for 2,925 (43%) cases (Table 1). A total of 122 TBI cases (4%) were classified as severe, 223 (8%) as moderate and 2,580 (88%) as mild (Table 1). More than half of moderate-to-severe injuries occurred over the age of 40 (χ2=23.52, df=8, p<0.01). By contrast, people younger than 19 years of age had the highest rates (92-94%) of mild TBI. Variation among severity and mechanism of injury was statistically significant (χ2=52.72, df=10, p<0.001). Falls were the leading cause of minor-to-moderate injuries (42-47%), while severe injuries were mostly (42.6%) due to motor vehicle crashes. People living in areas of high socioeconomic disadvantage were more likely to sustain severe injuries compared to the general population (41.2% vs 29.2%, χ2=8.49, df=1, p=0.003).

Table 1 also provides medical details of hospitalisation for the whole sample and severity spectrum. Average LOS was 6.4 days (SD=13.1), with inter-hospital transfers occurring in 9.2% of cases. Forty per cent of

Figure 1: Age/sex-specific incidence rates per 100,000 and proportion of injury cause by age in NSW, 2007 (n=6,827).

Pozzato et al. Article



Table 1: At-risk groups, associated factors and medical details of first-time hospitalised TBI in NSW (n=6,827) and in a subsample with severity data (n=2,925).NSW 2006 Census

Person/yearsHospitalised

TBIDeceased

TBITBI subsample with severity dataa

Total (n= 6,817,182)

Total (n=6,827)

Total (n=365)

Mild (n=2,580)

Moderate (n=223)

Severe (n=122)

Total (n=2,925)

N (%) N (%) N (%) N (%) N (%) N (%) N (%)Sociodemographic M:F (ratio) 0.99:1 2.4:1 1.5:1 2.9:1 2.7:1 2.9:1 2.9:1 Male gender 3,411,349 (50) 4793 (70.2) 218 (59.7) 1,926 (74.7) 163 (73.1) 91 (74.6) 2,180 (74.5)

Age, years (mean-SD) 38.7 (-) 43.1 (27.1) 69.2 (22.8) 38.9 (23.8) 44.3 (23.3) 43.7 (22.3) 39.5 (23.8) 0-9 876,967 (12.9) 536 (7.9) 6 (1.6) 120 (4.7) 6 (2.7) 2 (1.6) 128 (4.4) 10-19 913,136 (13.4) 1,172 (17.2) 15 (4.1) 584 (22.6) 31 (13.9) 19 (15.6) 634 (21.7) 20-39 1,930,938 (28.3) 1,819 (26.6) 28 (7.7) 801 (31) 67 (30) 36 (29.5) 904 (30.9) 40-69 2,438,428 (35.8) 1,695 (24.8) 84 (23) 685 (26.6) 74 (33.2) 43 (35.2) 802 (27.4) 70+ 657,713 (9.6) 1,605 (23.5) 232 (63.6) 390 (15.1) 45 (20.2) 22 (18) 457 (15.6)

Born in Australia 4,703,855 (69) 5,251 (76.9) 243 (71.9) 2,041 (79.1) 158 (70.9) 91 (74.6) 2,290 (80.5)Indigenous status 152,685 (2.2) 260 (3.8) 6 (1.8) 100 (4.0) 7 (3.2) 6 (5.2) 113 (4)Socio-economic status*

Low disadvantaged 2,141,608 (33.1) 1,936 (28.8) 126 (35.5) 747 (29.4) 72 (32.7) 26 (21.8) 845 (29.4) Medium disadvantaged 2,437,292 (37.7) 2,613 (38.9) 118 (33.2) 994 (39.2) 84 (38.2) 44 (37) 1,122 (39) High disadvantaged 1,884,400 (29.2) 2,171 (32.3) 111 (31.3) 798 (31.4) 64 (29.1) 49 (41.2) 911 (31.7)Location of residency**

Metro 4,948,309 (72.6) 4,482 (66.7) 267 (75.2) 1,651 (65) 160 (72.7) 78 (65.6) 1,889 (65.6) Rural 1,831,085 (26.8) 2,124 (31.6) 85 (23.9) 850 (33.5) 54 (24.5) 38 (31.9) 942 (32.7) Remote 37,788 (0.6) 116 (1.7) 3 (0.8) 39 (1.5) 6 (2.7) 3 (2.5) 48 (1.7)Injury cause Transport accident - 1,768 (25.9) 76 (20.8) 765 (29.7) 70 (31.4) 52 (42.6) 887 (30.3) Fall - 3,247 (47.6) 248 (67.9) 1,104 (42.8) 106 (47.5) 38 (31.1) 1,248 (42.7) Assault - 1,080 (15.8) 14 (3.8) 452 (17.5) 37 (16.6) 15 (12.3) 504 (17.2) Other - 485 (7.1) 6 (1.6) 200 (7.8) 7 (3.1) 5 (4.1) 212 (7.2) 2E-codes - 102(1.5) 9 (2.5) 46 (1.6) 3 (1.3) 7 (5.7) 52 (1.8) No E-code - 145 (2.1) 12 (3.3) 17 (0.7) 0 5 (4.1) 22 (0.8)Associated factors Drug/Alcohol use*** - 1,269 (18.6) 34 (9.3) 520 (20.2) 66 (29.6) 36 (29.5) 622 (21.3) Sport/Recreation*** - 976 (14.3) 9 (2.5) 497 (19.3) 27 (12.1) 15 (12.3) 539 (18.4)Medical details LoS, days (mean -SD) - 6.4 (13.1) 6.8 (12.1) 3.5 (7.5) 7.9 (12.7) 28.4 (38.6) 4.91 (12.2) 2 or more facilities - 632 (9.2) 34 (9.3) 195 (7.5) 35 (15.6) 26 (21.4) 256 (8.7) Deceased - 365 (5.3) - 31 (1.2) 17 (7.6) 39 (32) 87 (3)Associated injuries*** - Other mechanical trauma - 2,351 (34.4) 124 (34) 1,038 (40.2) 87 (39) 46 (37.7) 1,171 (40) Complications - 43 (0.4) 4 (1.1) 6 (0.2) 2 (0.9) 6 (4.9) 14 (0.5) Other injuries - 72 (1.1) 8 (2.2) 24 (0.9) 2 (0.9) 5 (4.1) 31 (1.1) 2 or more injuries - 219 (3.2) 24 (6.6) 50 (1.9) 13 (5.8) 22 (18) 85 (2.9) None - 4,142 (60.7) 205 (56.2) 1,462 (56.7) 119 (53.4) 43 (35.2) 1,624 (55.5)Co-morbidities*** - 1,607 (23.5) 248 (67.9) 397 (15.4) 46 (20.6) 57 (46.7) 500 (17.1)Notes:a: Subsample based on Post-Traumatic Amnesia and Loss of Consciousness codes.*Subsample based on available residential postcodes and ABS scores (2006 Census population, n=6463300; TBI sample: n=6720).**Subsample based on available residential postcodes and ABS scores (2006 Census population, n=6817182; TBI sample, n=6722).***Percentages of valid case.The data contains occasional missing data values, which are assumed to be random.

people sustained other mechanical trauma due to the accident, with one-in-two people having some injury or complications in addition to TBI. One-in-four people had pre-existing or simultaneously sustained comorbidities.

Overall, fatalities occurred in 5.3% of the sample. Those who died were generally older

than 70 years of age (63.6%; χ2=372.95, df=4, p<0.001) or sustained the most severe injuries (44.8%; χ2=400.15, df=2, p<0.001). The median age at death was 62.9 years. The majority of fatalities resulted from fall-related TBI (67.9%; χ2=93.61, df=5, p<0.001) compared to other causes.

Not surprisingly, severe TBI, in comparison with the less severe injuries, resulted in longer LOS of about one month (M=28 days; F=301.08, df=2, p<0.001), higher number of inter-hospital transfers (more than one facility in 21.4% of cases), and the highest fatality rates at 32%. A larger number of associated injuries (64.8%; χ2=22.1, df=2, p<0.001)


and co-morbidities (46.7%; χ2=82.83, df=2, p<0.001) was also found.

Medical details of hospitalisation varied significantly by age and cause of injury (Table 2). Compared to the other age groups, older people (70-and-older age group) had the longest LOS (M=10.7 days; F=87.06, df=4, p<0.001), and the highest fatality rates (14.5%; χ2=372.95, df=4, p<0.001). This group also accounted for the largest number (62.8%) of comorbidities (χ2=2070.96, df=4, p<0.001). Of these, cardiovascular and neurological diseases were the most frequent, with 30% of older people affected by multiple disorders. By contrast, comorbidities and mortality rates were very low in those <40 years (1-2%).

Discussion

This is the first study to provide population-based estimates of the incidence of hospitalised TBI in the whole of NSW. We also described specific at-risk groups and prevalent causes of injury. Further analysis included factors associated with TBI risk, injury severity and medical details of TBI hospitalisation. Although the overall TBI incidence in NSW is much lower than

expected from international comparisons, our data clearly reveal that greater TBI rates exist within specific sectors of the NSW population, together with differences in hospitalisation outcomes that were not highlighted by previous studies.

The incidence of 99/100,000 persons per year matched the current best estimates calculated by Tate et al.5 in 1988 in a defined NSW subpopulation, suggesting that the rate of hospitalised TBI in NSW has remained unchanged across a 20-year period, notwithstanding methodological differences between the two studies. The present study used broader inclusion criteria, considering all possible signs of TBI in any diagnosis field, as opposed to intracranial injuries6,26 as principal diagnosis5,6 used by previous studies. The only study adopting a similar approach provided a slightly lower incidence of 86/100,000 for the same period.7 This small difference may be explained by Moorin et al.7 not including cases without a specified external cause of injury, as well as TBI caused by self-harm.

Compared to international studies, NSW had the lowest incidence rate, showing rates two to three times lower than in Europe and the United States, respectively, which are in the

range of 200-300/100,000 population per annum.2 By contrast, there are considerable methodological differences in study design that make it difficult to compare New Zealand estimates with findings of all other countries. Feigin and colleagues27 reported an annual incidence of 790/100,000 population, using multiple sources for case identification that included hospital admissions, as well as community-based assessment and treatment data. Most other epidemiological studies of TBI, including the present study, are based on hospital data only. Another New Zealand study28 reported an unusually high incidence (1,750 /100,000 population). That study employed a longitudinal approach that restricted the age range to 0-25 years.

There were disparities in TBI risk among population groups. Although Tate et al.5 also found a significant increase of TBI incidence in people aged 75+ years in comparison with the preceding three decades (χ2=5.66, df=1, p<0.02), in this study the risk of having a TBI among older people was found to be almost three times higher than the overall incidence in both studies. This shift in observed TBI patterns is consistent with findings from other high income countries,4 where life

Table 2: Associated factors and medical details of first-time hospitalised TBI in NSW (n=6827) by age and mechanism of injury.Hospitalised TBI

(n=6,827)By age groups By injury mechanism

0-9 (n=536)

10-19 (n=1,172)

20-39 (n=1,819)

40-69 (n=1,695)

+70 (n=1,605)

Transport Accident (n=1,768)

Fall (n=3,247)

Assault (n=1,080)

Other (n=485)

Two E-codes (n=102)

No E-code (n=145)

N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%)Associated factors Drug/Alcohol use*** 0 119 (10.2) 563 (31) 495 (29.2) 92 (5.7) 207 (11.7) 527 (16.2) 445 (41.2) 29 (6) 22 (21.6) 39 (26.9)

Sport/Recreation*** 66 (12.3) 480 (41) 267 (14.7) 129 (7.6) 34 (2.1) 348 (19.7) 321 (9.9) 44 (4.1) 220 (45.4) 18 (17.6) 25 (17.2)Medical details LoS, days (mean -SD) 2.9 (7.8) 2.9 (7.1) 4.9 (11.3) 7.4 (16.9) 10.7 (16.2) 7.2 (15.6) 7.1 (11.7) 3.4 (8.1) 2.9 (8.3) 14.1 (33.6) 8.2 (15.8) 2 or more facilities 72 (13.4) 112 (9.6) 142 (7.8) 149 (8.8) 157 (9.7) 152 (8.6) 299 (9.2) 105 (9.7) 31 (6.4) 38 (37.3) 7 (4.8) Deceased 6 (1.1) 15 (1.3) 28 (1.5) 84 (5) 232 (14.5) 76 (4.3) 248 (7.6) 14 (1.3) 6 (1.2) 9 (8.8) 12 (8.3)Associated injuries***

Other trauma 66 (12.3) 424 (36.2) 659 (36.2) 666 (39.3) 536 (33.4) 1,094 (61.9) 888 (27.3) 244 (22.6) 67 (13.8) 40 (39.2) 18 (12.4) Complications 2 (0.4) 2 (0.2) 7 (0.4) 20 (1.2) 12 (0.7) 6 (0.3) 29 (0.9) 4 (0.4) 0 1 (1) 3 (2.1) Other injuries 6 (1.1) 8 (0.7) 19 (1) 29 (1.7) 10 (0.6) 11 (0.9) 27 (0.8) 13 (1.2) 6 (1.2) 6 (5.9) 9 (6.2) 2 or more 5 (0.9) 20 (1.7) 75 (4.1) 75 (4.4) 44 (2.7) 128 (10.3) 55 (1.7) 15 (1.4) 11 (2.3) 7 (6.9) 3 (9.1) None 457 (85.3) 718 (61.3) 1,059 (58.2) 905 (53.4) 602 (62.5) 529 (29.9) 2,248 (69.2) 804 (74.4) 401 (82.7) 48 (47.1) 112 (77.2)Co-morbidities***

Cardiovascular 2 (0.4) 7 (0.6) 43 (2.4) 130 (7.7) 273 (17) 80 (4.5) 322 (9.9) 27 (2.5) 7 (1.4) 6 (5.9) 13 (9) Neurologic 11 (2.1) 15 (1.3) 31 (1.7) 54 (3.2) 172 (10.7) 32 (1.8) 211 (6.5) 20 (1.9) 8 (1.6) 3 (2.9) 9 (6.2) Respiratory 3 (0.6) 6 (0.5) 12 (0.7) 12 (0.7) 12 (0.7) 9 (0.5) 22(0.7) 10 (0.9) 1 (0.2) 1 (1) 2 (1.4) Digestive 0 1 (0.1) 7 (0.4) 18 (1.1) 1 (0.1) 5 (0.3) 16 (0.5) 3 (0.3) 1 (0.2) 2 (2) 0 Systemic 1 (0.2) 6 (0.5) 8 (0.4) 78 (4.6) 88 (5.5) 16 (0.9) 145 (4.5) 10 (0.9) 6 (1.2) 3 (2.9) 1 (0.7) 2 or more 1 (0.2) 3 (0.3) 12 (0.7) 138 (8.1) 462 (28.8) 65 (3.7) 508 (15.6) 16 (1.5) 8 (1.6) 7 (6.9) 12 (8.3) None 518 (96.6) 1,134 (96.8) 1,706 (93.8) 1265 (74.6) 597 (37.2) 1,561 (88.3) 2,023 (62.3) 994 (92) 454 (93.6) 80 (78.4) 108 (74.5)Notes:***Percentages of valid cases.The data contains occasional missing data values, which are assumed to be random.



expectancy is increasing and falls have become the leading cause of sustaining a TBI. Even so, compared to some other studies that reported rates of 361/100,00026 and 380,000/100,000,29 the rate of TBI among the 85+ year group in the NSW population was much higher, reaching a peak of 499/100,000 population.

People living in the most remote areas of NSW represent another group at high TBI risk. While place of residence had previously been found to be associated with a higher risk of sustaining a TBI in specific sectors of the population,12,14 data across all population sectors were not available. This group was shown to have a similarly high risk as the 75+ age group, being nearly three times higher compared to the general population. Lower socioeconomic status was also found to be associated with occurrence and severity of TBI in this study.

In addition, the occurrence of TBI in NSW residents identifying as Aboriginal and Torres Strait Islander was 1.7 times the rate of the general population. This finding, supporting known ethnicity-related disparities in TBI risk,30 allows for a more accurate estimate of the magnitude of risk among Aboriginal and Torres Strait Islander peoples. There are only two studies providing data in this regard but these are limited to small, selected Australian cohorts, such as assault-related TBI12 and incarcerated young people.31

The interrelation among Indigenous status, remoteness and socioeconomic disadvantage has not been previously reported. All three groups separately have a higher risk of TBI than the general population. Although there were associations between remoteness and socioeconomic disadvantage that reflect the general population distribution, TBI risk was higher among all NSW residents living in areas that are both remote and highly disadvantaged. Among these, Aboriginal and Torres Strait Islander people had the highest proportion of remoteness and socioeconomic disadvantage, being twice as likely to sustain a TBI compared to the general Aboriginal and Torres Strait Islander population.

Our analysis of factors associated with TBI risk demonstrated that overall, at 19%, there was a lower incidence of people under the influence of substances at the time of their injury compared to previous estimates of 30-50%.32,33 Previous research has pointed to alcohol as an important factor associated with all TBI causes,34 and Wagner et al.35 linked the use of drugs and alcohol more specifically to violence-related TBIs. In our study, substance

use, especially weekend consumption, was also greatly associated with risk of sustaining a brain injury from assault, and the occurrence of assault-related TBI (15.8%) nearly doubled compared to Tate et al.5 By contrast, substance use in traffic-related TBI was not as high as previously documented,36 being in line with our overall finding of decreased traffic-related TBI rates.5

The exploration of medical details highlighted that not only were brain injuries in older people common, but they had relatively long LOS and high in-hospital mortality (14.5%). While complex care needs and difficulties in discharge planning may have contributed to prolonged LOS in older patients, greater mortality rates confirm this group having a much worse prognosis than younger people, which is likely due to physiological differences37 that impact on recovery from TBI. Several factors other than injury severity may be linked to the risk of dying after a TBI in the older population, such as a higher number of comorbidities.38 In population-based studies, the use of the Charlson Comorbidity Index does not allow distinction between pre-existing and co-occurring comorbidities, yet these are both believed to have a negative impact on recovery from injury.39 Three in four older people had other diseases in addition to TBI. Cardiovascular diseases were the most common, at about 17%, reinforcing previous hypotheses that cardiovascular morbidity26,40 may further complicate recovery from TBI.

Another group with less favorable in-hospital outcome includes people with severe TBI. Although data were only available for 43% of the sample, severity rates were in line with previous studies (mild, 88%; moderate 8%; severe 4%)27. Severe TBI compared to less severe injuries had the longest LOS and the highest fatality rates and proportion of inter-hospital transfers. Beside injury severity, differences between these groups may be partially explained by the high number of complications and systemic injuries associated with severe TBI.

The inclusion of data relating only to hospitalised TBI cases may be a limitation of this study. Death before hospital admission and mild injury for which hospital care was not sought were not included. Furthermore, ICD codes are known to not be specific for identification of hospitalised mild TBI. A strength of this study was the analysis of medical details of TBI hospitalisation, which are rarely available from epidemiological studies, although crucial to inform resource

planning and secondary prevention measures. This is an overview incidence study, drawing upon retrospective administrative data. Had specific data been available for a control population, we would have been able to conduct regression analyses to identify potential predictors/moderators of TBI. Nevertheless, there was a high chance that potentially important predictors/moderators were not contained in this data set, thus weakening any conclusions drawn from such an analysis. This will be a fruitful avenue for future research.

Implications

Public health strategies should be employed to reduce TBI incidence, increase awareness of and meet the healthcare demands for TBI-related services among identified high-risk groups. Effort is needed to establish whether modifiable risk factors exist in these population sectors. For instance, TBI risk among older people may be reduced by targeting anticoagulation,26,40 matching the success of decreased fall-related hip fracture rates in this age group.41

Consideration should be given to prevention efforts and making TBI-related services available in remote and socioeconomically disadvantaged areas. The high representation of Aboriginal and Torres Strait Islander peoples in these areas requires culturally aware health service provision to reduce barriers to health care access and ensure optimal care beyond the issue of medical service availability. In addition, the impact of recent health policy initiatives (e.g. the ‘Close the Gap: National Indigenous Reform Agreement’ in 2012 and the National Aboriginal and Torres Strait Islander Health Plan 2013–2023) on health outcomes in this group should be assessed.

Education about substance use may contribute to reducing assault-related TBI and injury vulnerability during weekends. In fact, there are indirect indications that traffic-related TBI rates and associated substance use may have fallen as a result of targeted prevention campaigns by the Australian government.

Further research is needed to elicit the human and economic impact of in-hospital mortality and prolonged hospitalisation of older people and people with severe TBI. It may be valuable to explore the influence on these population groups of the latest health and disability policy reforms. Among these, there is the introduction of the NSW



Lifetime Care and Support Scheme in 2006 (first participant entered in January 2007) to provide care services to all severe traffic-related TBI, together with the transition to the National Disability Insurance Scheme in 2016 to support recovery of all Australians with permanent and substantial disability below the age of 65. By contrast, hospital outcomes and costs of people over 65 years of age may benefit from significant changes in home and community care services as a result of the Australian aged care reforms starting from July 2013.

Conclusion

We have comprehensively investigated the incidence and epidemiological characteristics of TBI hospitalisations in NSW. These data demonstrate that TBI is much less common than in the rest of the world, but substantial differences exist in the incidence and in-hospital outcomes among population groups, calling for public health preventive strategy and actions. High TBI rates in older people, remote and socioeconomically disadvantaged residents, and Aboriginal and Torres Strait Islander peoples make these groups a public health priority. Further, TBI had a substantially negative impact on hospitalisation course in older people and people with severe injuries.

These findings highlight various challenges and areas for future research. Rigorous multivariate analysis is needed to examine specific contributors to injury vulnerability and the complex interrelationship among at-risk groups and the whole population. Prospective and community-based research is also warranted for more comprehensive estimates of TBI incidence, in particular, mild injuries, as well as clarifying disability and cost implications to the disparity in incidence and outcome between population groups.

References1. World Health Organization. Injuries and violence: the

facts 2014. Geneva (CHO):WHO; 2014 [cited 2019 Feb 14]. Available from: http://www.who.int/iris/handle/10665/149798

2. Nguyen R, Fiest KM, McChesney J, et al. The international incidence of traumatic brain injury: A systematic review and meta-analysis. Can J Neurol Sci. 2016;43(6):774-85.

3. Feigin VL, Barker-Collo S, Krishnamurthi R, et al. Epidemiology of ischaemic stroke and traumatic brain injury. Best Pract Res Clin Anaesthesiol. 2010;24(4):485-94.

4. Roozenbeek B, Maas AI, Menon DK. Changing patterns in the epidemiology of traumatic brain injury. Nat Rev Neurol. 2013;9(4):231-6.

5. Tate RL, McDonald S, Lulham JM. Incidence of hospital-treated traumatic brain injury in an Australian community. Aust N Z J Public Health. 1998;22(4):419-23.

6. Helps Y, Henley G, Harrison J. Hospital separations due to traumatic brain injury, Australia 2004–05. Injury research and statistics series number 45. (Cat no. INJCAT 116) Adelaide (AUST): Australian Institute of Health and Welfare; 2008.

7. Moorin R, Miller TR, Hendrie D. Population-based incidence and 5-year survival for hospital-admitted traumatic brain and spinal cord injury, Western Australia, 2003-2008. J Neurol. 2014;261(9):1726-34.

8. Barker-Collo SL, Feigin VL. Capturing the spectrum: Suggested standards for conducting population-based traumatic brain injury incidence studies. Neuroepidemiology. 2009;32(1):1-3.

9. Harradine PG, Winstanley JB, Tate R, et al. Severe traumatic brain injury in New South Wales: Comparable outcomes for rural and urban residents. Med J Aust. 2004;181(3):130-5.

10. Crowe L, Babl F, Anderson V, et al. The epidemiology of paediatric head injuries: Data from a referral centre in Victoria, Australia. J Paediatr Child Health. 2009;45(6):346-50.

11. Chang VC, Ruseckaite R, Collie A, et al. Examining the epidemiology of work-related traumatic brain injury through a sex/gender lens: Analysis of workers’ compensation claims in Victoria, Australia. Occup Environ Med. 2014;71(10):695-703.

12. Jamieson LM, Harrison JE, Berry JG. Hospitalisation for head injury due to assault among Indigenous and non-Indigenous Australians, July 1999-June 2005. Med J Aust. 2008;188(10):576-9.

13. Ballestas T, Xiao J, McEvoy S, Somerford P. The Epidemiology of Injury in Western Australia, 2000-2008. Perth (AUST): Western Australian Department of Health; 2011.

14. Berry JG, Jamieson LM, Harrison JE. Head and traumatic brain injuries among Australian children, July 2000–June 2006. Inj Prev. 2010;16:198-202.

15. Mollayeva T, Xiong C, Hanafy S, et al. Comorbidity and outcomes in traumatic brain injury: Protocol for a systematic review on functional status and risk of death. BMJ Open. 2017;7(10):e018626.

16. Australian Bureau of Statistics. 32010 - Population by Age and Sex, Australian States and Territories, June 2007.Canberra (AUST): ABS; 2007.

17. Australian Bureau of Statistics. Population Characteristics, Aboriginal and Torres Strait Islander Australians 2006.Canberra (AUST): ABS; 2006.

18. World Health Organization. ICD-10: International statistical classification of diseases and related health problems: tenth revision, 2nd ed. Geneva (CHO): WHO; 2004 [cited 2019 Feb 14]. Available from: http://www.who.int/iris/handle/10665/42980

19. Leibson CL, Brown AW, Ransom JE, et al. Incidence of traumatic brain injury across the full disease spectrum: A population-based medical record review study. Epidemiology. 2011;22(6):836-44.

20. Australian Bureau of Statistics. Census of Population and Housing: Socio-economic Indexes for Areas. Canberra (AUST): ABS; 2006.

21. Australian Bureau of Statistics. 1216.0. - Australian Standard Geographical Classification (ASGC). Canberra (AUST): ABS; 2006.

22. Bakhshaei M, Georgiou T, McAndrew M. Language of Instruction and Ethnic Disparities in School Success. McGill J Educ. 2016;51(2):689-713.

23. Jennett B. Assessment of the severity of head injury. J Neurol Neurosurg Psychiatry. 1976;39(7):647-55.

24. Stephenson S, Henley G, Harrison JE, et al. Diagnosis based injury severity scaling: Investigation of a method using Australian and New Zealand hospitalisations. Inj Prev. 2004;10(6):379-83.

25. Quan H, Li B, Couris CM, et al. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol. 2011;173(6):676-82.

26. Harvey LA, Close JC. Traumatic brain injury in older adults: Characteristics, causes and consequences. Injury. 2012;43(11):1821-6.

27. Feigin VL, Theadom A, Barker-Collo S, et al. Incidence of traumatic brain injury in New Zealand: A population-based study. Lancet Neurol. 2013;12(1):53-64.

28. McKinlay A, Grace R, Horwood L, et al. Prevalence of traumatic brain injury among children, adolescents and young adults: Prospective evidence from a birth cohort. Brain Inj. 2008;22(2):175-81.

29. Chan V, Zagorski B, Parsons D, et al. Older adults with acquired brain injury: A population based study. BMC Geriatr. 2013;13(1):97.

30. Lakhani A, Townsend C, Bishara J. Traumatic brain injury amongst indigenous people: A systematic review. Brain Inj. 2017;31(13-14):1718-30.

31. Moore E, Indig D, Haysom L. Traumatic brain injury, mental health, substance use, and offending among incarcerated young people. J Head Trauma Rehabil. 2014;29(3):239-47.

32. Opreanu RC, Kuhn D, Basson MD. The influence of alcohol on mortality in traumatic brain injury. J Am Coll Surg. 2010;210(6):997-1007.

33. Shandro JR, Rivara FP, Wang J, et al. Alcohol and risk of mortality in patients with traumatic brain injury. J Trauma Acute Care Surg. 2009;66(6):1584-90.

34. Finfer SR, Cohen J. Severe traumatic brain injury. Resuscitation. 2001;48(1):77-90.

35. Wagner AK, Sasser HC, Hammond FM, et al. Intentional traumatic brain injury: Epidemiology, risk factors, and associations with injury severity and mortality. J Trauma Acute Care Surg. 2000;49(3):404-10.

36. Kraus JF, Morgenstern H, Fife D, et al. Blood alcohol tests, prevalence of involvement, and outcomes following brain injury. Am J Public Health. 1989;79(3):294-9.

37. Collaborators MCT, Perel P, Arango M, et al. Predicting outcome after traumatic brain injury: Practical prognostic models based on large cohort of international patients. BMJ. 2008;336(7641):425.

38. Susman M, DiRusso SM, Sullivan T, et al. Traumatic brain injury in the elderly: Increased mortality and worse functional outcome at discharge despite lower injury severity. J Trauma Acute Care Surg. 2002;53(2):219-24.

39. Kumar RG, Juengst SB, Wang Z, et al. Epidemiology of comorbid conditions among adults 50 years and older with traumatic brain injury. J Head Trauma Rehabil. 2018;33(1):15-24.

40. Erlebach R, Pagnamenta A, Klinzing S, et al. Age-related outcome of patients after traumatic brain injury: A single-center observation. Minerva Anestesiol. 2017;83(11):1169-77.

41. Pasco JA, Brennan SL, Henry MJ, et al. Changes in hip fracture rates in southeastern Australia spanning the period 1994–2007. J Bone Miner Res. 2011;26(7):1648-54.



Supplementary Table 1: International Classification of Diseases, 10th Revision diagnosis codes used in selection of traumatic brain injury events and description of injury-related characteristics from NSW Department of Health hospitalisation data.

Supplementary Table 2: Associations between remoteness and socio-economic disadvantage in the NSW general population – 2006 Census (n= 6,463,300) and TBI sample (n=6,720).

Supplementary Table 3: Associations between remoteness and socio-economic disadvantage in the NSW Aboriginal and Torres Strait Islander population – 2006 Census (n= 135,299) and TBI Aboriginal and Torres Strait Islander subsample (n=255).

Supplementary Table 4: Substance use by seasonality in first-time hospitalised TBI in NSW (n=6,827).

Supplementary Figure 1: Flowchart of study population from NSW hospitalisations during calendar year 2007.



Rickettsial diseases are caused by the obligate intracellular bacteria in the order Rickettsiales that can be

transmitted to humans via the bites of fleas, lice, ticks or mites.1,2 Australia has several endemic rickettsial diseases: flea-borne murine typhus (R. typhi) and cat flea typhus (R. felis); mite-borne scrub typhus (Orientia tsutsugamushi); and the tick-borne spotted fever group (SFG), which includes Queensland tick typhus (R. australis), Flinders Island spotted fever (FISF; R. honei), and Australian spotted fever (R. honei subsp. marmionii).1,3,4

Epidemic typhus (R. prowazekii) occurred in Australia in the 18th and 19th centuries but is no longer endemic.1,5

Tasmania (population 524,700)6 is an island state of Australia located 240 kilometres south of the mainland, consisting of the main island of Tasmania and several small, sparsely populated surrounding islands. FISF, the rickettsial infection typically associated with Tasmania, was first described in 1991 on Flinders Island, an island in Bass Strait off the north-east tip of Tasmania, by the island’s sole general practitioner (GP), Robert Stewart, who had identified 26 cases of a spotted-fever-like illness over 17 years.7 The causative agent has been identified as R. honei,1,8 and the reptile associated tick Bothriocroton hydrosauri (Southern Reptile Tick, formerly Aponomma hydrosauri), which commonly feeds from native reptiles such as blue-tongued lizards and snakes, has been confirmed as its host (Figure 1).9,10 Cases have not only been observed on Flinders Island, but also on Schouten Island, Tasmania, and in south-eastern South Australia, where B. hydrosauri are also endemic.11 Additionally, two cases of

locally acquired R. honei infection have been described in Western Australia,12 while cases have also been described internationally.13 Symptoms can include sudden onset fever and chills, myalgia, transient arthralgia and a maculopapular rash.7 Treatment is usually with oral doxycycline;2 however, azithromycin and rifampicin have also been used successfully.

Rickettsial infection is not nationally notifiable in Australia under the National Notifiable Disease Surveillance System (NNDSS), but surveillance currently occurs in three jurisdictions. Epidemic typhus caused by R. prowazekii is notifiable in New South Wales,

while rickettsial infection is notifiable in Western Australia and Tasmania. In Tasmania, cases are notifiable under the Public Health Act 1997.14 Positive rickettsial serology results (antibody titre ≥1:128 to a rickettsial group antigen) occurring in Tasmanian laboratories are forwarded to the Communicable Disease Prevention Unit (CDPU) and investigated. Standardised data collection on cases of rickettsiosis notified to CDPU was introduced at the beginning of 2012.

The existing literature on rickettsial infection in Tasmania describes cases of FISF on Flinders Island, one case on Schouten Island11

and a case of rickettsiosis, assumed to be

New and old hotspots for rickettsial spotted fever acquired in Tasmania, 2012–2017Gabriela Willis,1,2 Kerryn Lodo,1 Alistair McGregor,3 Faline Howes,1 Stephanie Williams,2 Mark Veitch1

1. Communicable Disease Prevention Unit, Public Health Services, Department of Health, Tasmania2. National Centre for Epidemiology and Population Health, Australian National University, Australian Capital Territory3. Department of Microbiology and Infectious Diseases, The Royal Hobart Hospital, Tasmania Correspondence to: Dr Gabriela Willis, Public Health Services, Department of Health, GPO Box 125, Hobart, Tasmania 7001; e-mail: [email protected]: February 2019; Revision requested: March 2019; Accepted: May 2019 The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,



Abstract

Objective: To describe the epidemiology and clinical characteristics of Tasmania-acquired rickettsial disease notified to the Department of Health in Tasmania from 2012 to 2017 inclusive.

Methods: Data on rickettsiosis cases acquired and notified in Tasmania between 1 January 2012 and 31 December 2017 were analysed descriptively.

Results: Eighteen cases of rickettsial infection notified in Tasmania 2012–17 and likely acquired in the state met one of three case definitions: 12 confirmed (67%), four probable (22%), and two possible (11%). The mean number of cases per year was 3.0 (population rate 0.6 per 100,000 population/year); 60% of cases occurred in November and December. Cases were more commonly older males. Fever, lethargy, and rash were commonly reported symptoms. Thirteen cases were likely acquired on Flinders Island, three around Great Oyster Bay and two in the Midlands.

Conclusions: This study extends our knowledge of the epidemiology of rickettsial disease in Tasmania. This is the first account including confirmed cases acquired in the Midlands of Tasmania.

Implications for public health: Increased knowledge and awareness of epidemiology of rickettsial infection in Tasmania is essential for timely diagnosis and appropriate treatment. These findings bear wider relevance outside Tasmania because visitors may also be at risk.

Key words: Rickettsia, infectious diseases, epidemiology, surveillance, Flinders Island spotted fever



Willis et al. Article

R. honei, acquired on the Freycinet Peninsula or Maria Island.15 However, the geographical distribution and other epidemiological characteristics of rickettsial disease in Tasmania outside of Flinders Island are not well understood. A cluster of four rickettsial infection cases was identified on Flinders Island in December 2017, prompting the authors to review the surveillance data. The aim of this study was to review all rickettsiosis notifications in Tasmania over the six years from 2012 to 2017 inclusive and report the clinical and epidemiological characteristics of these cases.

Methods

Study populationConfirmed, probable, and possible rickettsiosis cases notified to CDPU between 1 January 2012, when the standardised case report form was introduced, and 31 December 2017, were included. Cases where travel history was unknown or that were likely acquired outside of Tasmania, based on overseas or interstate travel during the exposure period, were excluded. Confirmed cases required laboratory definitive evidence, including detection (culture or nucleic acid test) of Rickettsia species in a clinical specimen, or seroconversion or a fourfold or greater rise in serum antibody titre to rickettsial group antigen between acute and convalescent phase sera specimens. Probable cases required clinical evidence and a single elevated antibody titre of equal to or greater than 1:256 to a rickettsial group antigen. Possible cases required clinical evidence and a single antibody titre of 1:128 to a rickettsial group antigen. Clinical evidence was defined as a clinically compatible illness (fever and

at least one of headache, myalgia, rash, or eschar).

Data collectionRickettsial infections were notified to CDPU by Tasmanian laboratories and followed up within five days of notification. Further details about the illness, contact details, and permission to contact the patient were obtained from the testing clinician. A questionnaire was administered via telephone to the patient by a public health nurse or doctor, collecting information on demographics, clinical details of the illness, history of rickettsial infection and travel history (overseas, interstate, and intrastate). If only one serology test had been performed, a repeat serology two weeks later was recommended to confirm the diagnosis. Case details including demographic, hospitalisation and basic laboratory data were entered into the Tasmanian Notifiable Disease Database (TNDD).

For this retrospective descriptive analysis of notified cases, data were extracted from the TNDD into Excel (Microsoft, Version 16, 2016). Original case reports were reviewed, and additional data not recorded in the TNDD, including clinical details, tick bite exposure, previous rickettsial disease, travel history, and detailed laboratory data, were added to the dataset.

Laboratory investigationsAll cases had immunoglobulin G (IgG)-specific antibody titres against Rickettsia antigen performed by indirect immunofluorescence assay (IFA). Testing varied depending on which pathology provider was used. The majority were tested for spotted fever group (SFG), typhus group (TG) and scrub typhus

group (STG). SFG antigen included R. australis (Queensland Tick Typhus), R. honei (FISF), and R. africae (African Tick Bite Fever); TG antigen included R. typhi (murine typhus); and STG included Orientia tsutsugamushi. Seven cases had one or both samples tested at the Australian Rickettsial Reference Laboratory in Geelong, and had titres against specific antigens, rather than groups. These included R. australis (Queensland Tick Typhus), R. honei (FISF), R conorii (Mediterranean Spotted Fever), R. africae (African Tick Bite Fever) R. rickettsii (Rocky Mountain Spotted Fever) and R. felis (cat flea typhus/Flea-borne Spotted Fever) in the SFG; R. typhi (murine typhus) and R. prowazekii (Epidemic typhus) in the TG; and Orientia tsutsugamushi and O. chuto in the STG. IgG titres ≥1:128 were considered positive.

Data analysisDescriptive data analysis was performed exploring patterns over time, age and sex distribution, geographic distribution, and clinical and laboratory features of confirmed, probable, and possible cases. Population rates were calculated using the Australian Bureau of Statistics mid-year Tasmanian population estimate for each year.3 Data were analysed in Stata/IC (StataCorp LLC, Texas USA, Version 15.0). Likely place of acquisition was assessed based on residential location and report of overseas, interstate, and intrastate travel in the 10 days prior to onset of illness. Likely place of acquisition was mapped using ArcMap (ESRI, Version 10.6, 2018).

EthicsData were collected under the provisions of the Public Health Act of Tasmania (1997), and the data analysis and report with the approval of the Australian National University Ethics Committee (Protocol 2017/909).

Results

There were 47 rickettsiosis cases in the TNDD notified between 1 January 2012 and 31 December 2017. Of these, 19 did not meet a case definition and five were repeat notifications. Three cases likely acquired outside of Tasmania were excluded: one acquired in South Africa, one acquired in Western Australia, and one acquired in the Northern Territory or Queensland. Additionally, two cases where travel history was unknown were excluded. Eighteen cases were included in the analysis.

Figure 1: Blotched blue-tongue lizard (Tiliqua nigrolutea) with Southern Reptile Ticks (Bothriocroton hydrosauri), Flinders Island, Tasmania.

Photo courtesy of Robert Stewart.


General Public Health Rickettsial spotted fever in Tasmania, 2012–2017

Laboratory investigationsFourteen cases (78%) had two blood samples taken for IFA and four (22%) had one. The median time between onset of illness and the first serology sample was four days (range 0–42 days) and the median time between first and second serology samples was 14.5 days (range 4–43 days).

Twelve (67%) cases demonstrated seroconversion or a fourfold or greater rise in antibody titre to either SFG or multiple Rickettsia species and were classified as confirmed cases. An additional four cases (22%) had single high titres and were classified as probable cases (cases 2, 3, 10 and 11). Two cases (11%) had low titres (1:128) and were classified as possible cases (cases 4 & 18; Table 1).

In the six cases who had IFA against individual species rather than groups, there was evidence of significant cross-reactivity, with all cases having positive titres of the same dilution against several species, including R. honei (Table 1).

Incidence and patterns over timeA mean of 3.0 cases per year were notified between 2012 and 2017 (range 2 to 4; Figure 2). The mean annual rate of notification was 0.6 per 100,000 population/year (range 0.4 to 0.8 per 100,000 population/year).

The date of onset rather than date of notification is shown in Figure 2 (one case notified in January 2012 had onset in December 2011). There was a distinct seasonal pattern, with 11 of the 17 cases with onset during the years 2012 to 2017 (where a full calendar year was included) occurring in the last quarter of the year (mostly in November and December), four with onset in the first quarter and only two in either the second or third quarter.

An increase in notifications was noted in the late 2017, with five cases with onset in the last quarter of 2017 (Figure 2). Three confirmed and one probable case were part of an identified cluster on Flinders Island, with dates of onset between 4 and 27 November 2017.

Demographic characteristicsTwelve cases (67%) were male and six (33%) were female. The median age was 60 years (range 35–77 years). Indigenous status was unknown for seven cases but, among the remaining 11, none reported being Aboriginal and/or Torres Strait Islander.

Table 1: Summary of serology results and cases classification, by case.Case number

Year & quarter (Q) of onset

Number of blood samples

IgG titre against Rickettsia antigen by IFA Case classification

1 Q4, 2011 2 Four-fold increase to SFG (1:512 and 1:8192) Confirmed2 Q1, 2012 1 Single positive titre to SFG (1:4096) and TG (1:128) Probable3 Q1, 2012 1 Single positive titre to SFG (1:1024) Probable4 Q3, 2013 1 Single positive titres to multiple species: R. australis, R. honei, R. conorii,

R. africae, R. rickettsii (1:128 to all) Possible

5 Q4, 2013 2 Seroconversion to multiple species: R. australis, R. honei, R. conorii, R. africae, R. rickettsii (1:256 to all)

Confirmed

6 Q4, 2013 2 Four-fold increase to SFG (1:512 and 1:8192) Confirmed7 Q1, 2014 2 Seroconversion to SFG (negative and 1:2048) Confirmed8 Q1, 2014 2 Seroconversion to SFG (negative and 1:2048)

Positive titres to STG (1:256 and 1:128)

Confirmed

9 Q2, 2014 2 Seroconversion to multiple species: R. australis (≥1:1024), R. honei (≥1:1024), R. conorii (1:256), R. africae (1:256), R. rickettsii (1:256), R. felis (1:128), R. prowazekii (1:256), R. typhi (1:128)

Confirmed

10 Q4, 2014 2 Positive titre to SFG on two samples with no seroconversion (1:2048 and 1:2048)

Probable

11 Q4, 2015 2 Positive titre to SFG on two samples with no seroconversion (1:16384 and 1:32768)

Positive titre to TG with no seroconversion (1: 256 and 1:256)

Probable

12 Q4, 2015 2 Seroconversion to SFG (negative and 1:1024) Confirmed13 Q4, 2016 2 Seroconversion to multiple species: R. australis (≥1:1024), R. honei

(≥1:1024), R. conorii (≥1:1024), R. africae (≥1:1024), R. rickettsii (≥1:1024), R. felis (≥1:1024)

Confirmed

14 Q4, 2016 2 Seroconversion to SFG (negative to 1:4096)

Positive titres to STG (1:256 and 1:128)

Confirmed

15 Q4, 2017 2 Seroconversion to SFG (negative to 1:8192) Confirmed16 Q4, 2017 2 Seroconversion to SFG (negative to 1:1024) Confirmed17 Q4, 2017 2 Seroconversion to multiple species: R. australis (≥1:1024), R. honei

(≥1:1024), R. conorii (≥1:1024), R. africae (≥1:1024), R. rickettsii (≥1:1024), R. felis (1:256), R. prowazekii (≥1:1024), R. typhi (≥1:1024)

Confirmed

18 Q4, 2017 1 Single positive titres to multiple species: R. australis, R. honei, R. rickettsii, R. felis, R. prowazekii, R. typhi (1:128 to all)

Possible

Thirteen of the cases provided their occupation, with five retired (28%), and four working on farms (31%). Other reported occupations included a bulldozer driver, a welder, a retailer and a teacher.

Clinical characteristicsAll but one case (95%) reported fever. Other commonly reported symptoms were lethargy (89%), rash (83%), myalgia and/or arthralgia (72%), and headache (72%), see Table 2. Unsolicited reported symptoms included chills and/or sweats (n=3), itchiness (n=1), shortness of breath (n=1), anorexia (n=1), nausea (n=1), sore throat (n=1), dry mouth (n=1) and photophobia (n=1).

A description was given for seven of the 15 persons with a rash and included ‘widespread’ or ‘full body’ (n=3), ‘full body sparing face and hands’ (n=1), ‘purpuric rash lower legs’ (n=1), and ‘mainly on trunk with patches on legs’ (n=1). One confirmed case reported being red and swollen at the location of a suspected tick bite. A description was given for four of the

13 cases with myalgia and/or arthralgia and included ‘aches whole body’, ‘ankle pain and swelling’, ‘all joints’, and ‘arthralgia of knees, shoulders, and fingers’.

Six cases (33%) reported a tick bite, while three (17%) reported a bite of unknown origin. Hospitalisation status was known for eight cases, with five reporting hospitalisation and three reporting no hospitalisation. No cases died of their disease.

Place of acquisitionThirteen cases were likely acquired on Flinders Island (11 residents and 2 visitors),

Table 2: Clinical presentation of cases (n=18).

SymptomNumber of cases

reporting symptomn %

Fever 17 95Lethargy 16 89Rash 15 83Myalgia/arthralgia 13 72Headache 13 72Other 7 39


to mild illness or other factors; clinical presentation is similar to many other illnesses, particularly viral infections; and – perhaps most importantly – to diagnose rickettsial infection, the clinician must first consider it in their differential diagnosis and order the appropriate laboratory tests. Making the diagnostic connection may be further hindered by the patient not reporting a tick bite. In this sample, only 50% reported a bite of any kind, and only 30% a tick bite.

Tasmania is a popular tourist destination, with 1.26 million visitors in the year ending December 2017,16 and an estimated 1,500 adults holidaying on Flinders Island per year.17{Tourism Tasmania, 2009 #172} A high proportion of visitors undertake outdoor activities that may put them at risk of tick bites.18 Although it is likely that some visitors to Tasmania will develop rickettsial disease, they may seek medical advice after leaving the island. Treating clinicians outside of Tasmania may not be aware of the risk and not perform the appropriate tests. Additionally, as rickettsial infection is not nationally notifiable, many infections acquired by visitors are unlikely to be captured in the surveillance systems of other jurisdictions.

Despite being often described as a mild illness,2 FISF can be severe. Recently, the death of a middle-aged woman due to acute infection of R. honei has been described in Queensland,19 and another severe case was reported in New South Wales that required intensive care (Dr Stephen R. Graves, personal communication). In our surveillance data, half of cases with hospitalisation status reported were hospitalised. This proportion may overestimate true hospitalisation rates, due to the likelihood that this sample is biased towards severe disease. Although much is still unknown about the causes of severe disease and complications, treatment with doxycycline may reduce the risk.2 Furthermore, prompt treatment is likely to reduce burden of disease as duration of illness without treatment is approximately 19 days but can be as long as six weeks.7

This work sheds new light on the geographical distribution of rickettsial infection in Tasmania. Previously cases of FISF in Tasmania have been described only on Flinders Island,7 and Schouten Island off the Freycinet Peninsula on the east coast of Tasmania,11 with one case of rickettsial infection, assumed to be R. honei, acquired on the Freycinet Peninsula or Maria Island,15

Figure 3: Likely place of acquisition of rickettsial infection cases, Tasmania, 2012-2017.

Figure 2: Epidemic curve of notified rickettsial infection cases, Tasmania, 2012-2017.

with nine being confirmed, two probable, and two possible. Three cases were likely acquired around Great Oyster Bay: one in Swansea (probable), one in Dolphin Sands (probable), and one in Coles Bay (confirmed). Two further cases were likely acquired in the Midlands: one in Ross (confirmed) and one in Tunnack (confirmed), see Figure 3. Two of the cases acquired around Great Oyster Bay had onset of illness in the first quarter of 2012 and the cases in the Midlands had onset of illness in the first and second quarter of 2014.

Discussion

Between 2012 and 2017 inclusive, a total of 18 confirmed, probable and possible rickettsial infection cases were notified in Tasmania that were likely acquired in the state, at a rate of 0.6 per 100,000 population/year. This number may significantly underestimate the true incidence of cases occurring in the community, due to the limitations inherent with passive surveillance data. Some cases may not have sought medical advice due



an island further down the east coast (Figure 3). As the invertebrate host Bothriocroton hydrosauri is distributed throughout Tasmania,20 it has been postulated that FISF is likely to be distributed more widely, but this is the first published evidence of cases likely acquired outside of these locations. Of note, this is the first description of cases likely acquired in other locations around Great Oyster Bay and of confirmed cases in the Midlands. A significant limitation of the data, however, is that there may be testing bias, with clinicians in areas with known previous cases more likely to consider rickettsiosis and test for it. Our impression is that there is high awareness of rickettsial infection among medical practitioners and the community on Flinders Island, but less awareness elsewhere in Tasmania.

The seasonal pattern is consistent with previous literature and is not unexpected, as summer is when reptiles are most active and humans are likely to spend more time outdoors, increasing the risk of tick bites. During the cluster of cases on Flinders Island in November 2017, there were anecdotal reports of residents seeing snakes and lizards with ‘big ticks’ on them, demonstrating the close proximity of humans and Southern Reptile Ticks. Similarly, the preponderance of cases among older males has been previously described,7,11 and is perhaps explained by a higher chance of them working or engaging in leisure activities outdoors.

The clinical picture among these cases is consistent with previous reports of FISF and particularly with Stewart’s findings in his case series from Flinders Island in 1991.7 Fever, headache, myalgia and/or arthralgia, and rash were prominent features. As in our study, Stewart also reported a varying distribution of the rash and both blanching and purpuric lesions. Similarly, rickettsial infections occurring in Queensland have shown a wide variety of dermatological manifestations, including maculopapular, macular, vesicular, and purpuric rashes.21 In our cases, there were no specific reports of cough or eschar, although one person described redness and swelling at the site of a suspected tick bite. This contrasts with Stewart’s case series, in which 46% had a cough and 46% had a skin lesion other than the rash. However, we did not specifically solicit these symptoms and signs, and our clinical data are limited by the collection of surveillance data via questionnaire over the telephone with the patient, rather than direct

clinical history and examination. Lethargy was reported in 90% of cases, although duration could not be ascertained. Chronic fatigue has been previously associated with rickettsial infection.22,23

We assume that the 18 infections described in this paper represent FISF due to R. honei infection, based on known epidemiology of rickettsioses in Australia. However, given the significant cross-reactivity seen with rickettsial serology infection it is possible that another rickettsial species could be responsible for some of these cases. Ixodes tasmani, another tick common in Tasmania, has been associated with spotted fever elsewhere in Australia.1,24 Izzard et al.24 found that 55% of I. tasmani ticks collected from Tasmanian Devils in 2005/06 were polymerase chain reaction (PCR) positive for a Rickettsia species, subsequently named Candidatus Rickettsia tasmanensis. This tick is known to bite humans, but further research is needed to understand whether this new rickettsial species is a potential human pathogen.3

The only preventative measure against rickettsial infection is to minimise exposure to ticks and thus reduce the risk of tick bites. Current advice is to wear long sleeves and long pants, use tick repellent on skin and clothes and sleep on a raised camp bed when camping. In an attempt to raise awareness, CDPU has developed a factsheet on FISF and has liaised with Environmental Heath Officers and Parks and Wildlife Service Tasmania to distribute information to both residents and visitors.25

Based on this review, CDPU has changed the surveillance case definition for rickettsial infection in Tasmania to include those cases with a single high titre and clinical evidence as probable cases. The aim of this change is to have a more sensitive case definition that improves Tasmanian rickettsial infection surveillance.

Conclusion

This review of rickettsiosis notifications in Tasmania between 2012 and 2017 extends knowledge of the epidemiology of rickettsial infection and FISF within Tasmania over recent years. In particular, it shows that rickettsial infection can be acquired in locations around Great Oyster Bay and the Midlands area. It is possible that the distribution is wider still, given

that the Southern Reptile Tick is distributed throughout Tasmania. These findings have wider relevance outside Tasmania as its many visitors, many of whom undertake outdoor activities, may also be at risk. Increased awareness of potential infection by clinicians is essential to accurately diagnose and appropriately treat rickettsial infections, thus reducing the burden of the disease. Although these data add to our understanding of the epidemiology of rickettsial infection in Tasmania, our knowledge in this complex area remains incomplete and further work is needed, particularly with regard to other potential human rickettsial pathogens.

Acknowledgements

The authors would like to acknowledge the Clinical Nurse Consultants in CDPU including Angela Russell, Kate Turner, Nicola Mulcahy, Juanita Mayne and Bethany Reszke for their expert contribution to collection of data and the public health management of notified cases. We would also like to thank all the patients involved for their time and input and the GPs on Flinders Island, in particular Dr Alexander John, for their assistance and input during the identified cluster of cases in 2017. In addition, thank you to Dr Stephen R. Graves, Medical Director of the Australian Rickettsial Reference Laboratory, for reviewing an earlier version of this manuscript. As an expert in the field of rickettsial microbiology and illness, his review and comments were invaluable. Dr Rob Stewart of Flinders Island also kindly reviewed a late draft and gave his valuable historical perspective, in addition to providing the photo presented in Figure 1.

References 1. Graves S, Stenos J. Rickettsioses in Australia. Ann N Y

Acad Sci. 2009;1166:151-5. 2. Heymann D. Control of Communicable Diseases Manual.

20th ed. Washington (DC): American Public Health Association; 2015.

3. Graves SR, Stenos J. Tick-borne infectious diseases in Australia. Med J Aust. 2017;206(7):320-4.

4. Williams M, Izzard L, Graves SR, Stenos J, Kelly JJ. First probable Australian cases of human infection with Rickettsia felis (cat-flea typhus). Med J Aust. 2011;194(1):41-3.

5. Cumpston JHL. Health and Disease in Australia: A History. Canberra (AUST): AGPS; 1989.

6. Australian Bureau of Statistics. 3101.0. - Australian Demographic Statistics [Internet]. Canberra (AUST): ABS; 2018 [cited 2018 Aug 29]. Available from: http://abs.gov.au/ausstats/[email protected]/0/D56C4A3E41586764CA2581A70015893E?Opendocument

7. Stewart RS. Flinders Island spotted fever: A newly recognised endemic focus of tick typhus in Bass Strait. Part 1. Clinical and epidemiological features. Med J Aust. 1991;154(2):94-9.

General Public Health Rickettsial spotted fever in Tasmania, 2012–2017


8. Graves SR, Stewart L, Stenos J, Stewart RS, Schmidt E, Hudson S, et al. Spotted fever group rickettsial infection in south-eastern Australia: Isolation of rickettsiae. Comp Immunol Microbiol Infect Dis. 1993;16(3):223-33.

9. Stenos J, Graves S, Popov VL, Walker DH. Aponomma hydrosauri, the reptile-associated tick reservoir of Rickettsia honei on Flinders Island, Australia. Am J Trop Med Hyg. 2003;69(3):314-17.

10. Whitworth T, Popov V, Han V, Bouyer D, Stenos J, Graves S, et al. Ultrastructural and genetic evidence of a reptilian tick, Aponomma hydrosauri, as a host of Rickettsia honei in Australia: Possible transovarial transmission. Ann N Y Acad Sci. 2003;990:67-74.

11. Unsworth NB, Stenos J, McGregor AR, Dyer JR, Graves SR. Not only ‘Flinders Island’ spotted fever. Pathology. 2005;37(3):242-5.

12. Raby E, Pearn T, Marangou A, Merritt A, Murray R, Dyer J, et al. New Foci of Spotted Fever Group Rickettsiae Including Rickettsia honei in Western Australia. Trop Med Infect Dis. 2016;1(1):5.

13. Graves S, Stenos J. Rickettsia honei: A spotted fever group Rickettsia on three continents. Ann N Y Acad Sci. 2003;990:62-6.

14. Australian Department of Health. Australian National Notifiable Diseases Case Definitions - Appendix B: Australian State and Territory Notifiable Diseases [Internet]. Canberra (AUST):Government of Australia; 2016 [cited 2018 Aug 29]. Available from: www.health.gov.au/internet/main/publishing.nsf/Content/cda-surveil-nndss-casedefs-statedis.htm

15. Chin RH, Jennens ID. Rickettsial spotted fever in Tasmania. Med J Aust. 1995;162(12):669.

16. Tourism Tasmania. Tasmanian Tourism Snapshot - Year Ending December 2017 [Internet]. Hobart (AUST): State Government of Tasmania; 2017 [cited 2018 Aug 29]. Available from: https://www.tourismtasmania.com.au/__data/assets/pdf_file/0010/62992/2017-Q4-Tasmanian-Tourism-Snapshot-YE-December-2017.pdf

17. Tourism Tasmania. Flinders Island Vistors Survey Report. Year Ending June 2009 [Internet]. Hobart (AUST): State Government of Tasmania; 2009. [cited 2018 Aug 29]. Available from: https://www.tourismtasmania.com.au/__data/assets/pdf_file/0018/54450/Flindersislandreport09.pdf

18. Tourism Tasmania. Tasmanian Visitor Survey- TVS Analyser [Internet]. Hobart (AUST): State Government of Tasmania; 2018 [cited 2018 Aug 29]. Available from: http://www.tourismtasmania.com.au/research/tvs

19. Graham RMA, Donohue S, McMahon J, Jennison AV. Detection of spotted fever group rickettsia dna by deep sequencing. Emerg Infect Dis. 2017;23(11):1911-13.

20. Barker SC, Walker AR. Ticks of Australia. The species that infest domestic animals and humans. Zootaxa. 2014;3816(1):1-144.

21. Stewart A, Hajkowicz K. Heterogeneity in skin manifestations of spotted fever group rickettsial infection in Australia. Australas J Dermatol. 2018;59(4):349-51.

22. Unsworth N, Graves S, Nguyen C, Kemp G, Graham J, Stenos J. Markers of exposure to spotted fever rickettsiae in patients with chronic illness, including fatigue, in two Australian populations. QJM. 2008;101(4):269-74.

23. Watts MR, Benn RA, Hudson BJ, Graves S. A case of prolonged fatigue following an acute rickettsial infection. QJM. 2008;101(7):591-3.

24. Izzard L, Graves S, Cox E, Fenwick S, Unsworth N, Stenos J. Novel rickettsia in ticks, Tasmania, Australia. Emerg Infect Dis. 2009;15(10):1654-6.

25. Communicable Disease Prevention Unit. Flinders Island Spotted Fever- Factsheet [Internet]. Hobart (AUST): Tasmanian State Government Department of Health; 2018 [cited 2018 Aug 29]. Available from: https://www.dhhs.tas.gov.au/publichealth/communicable_diseases_prevention_unit/infectious_diseases/flinders_island_spotted_fever



In 2015 the Sendai Framework for Disaster Risk Reduction (SFDRR) was endorsed by the United Nations General Assembly.

Sendai is a non-binding agreement that recognises the State has the primary role to reduce disaster risk but that responsibility should be shared with other stakeholders including local government and the private sector. It aims to substantially reduce disaster risk and loss of life, livelihoods and health in the economic, physical, social, cultural and environmental assets of persons, businesses, communities and countries.1

The SFDRR has evolved beyond its predecessor, the Hyogo Framework for Action 2005-2015, to embrace human health and wellbeing, and in encompassing science and technology (for example, there were three references to ‘technology’ in the Hyogo Framework for Action and 19 in the SFDRR).2 This includes connecting policy development and implementation with evidence and facilitating the transformation and transfer of research into practice. Three components of the SFDRR – health, economic development and climate change – demonstrate how public health is situated within Sendai, and how the boundaries between public health and environmental health are increasingly less distinct.3 Public health as a discipline has accordingly expanded beyond responding to specific events: collaboration, capacity building and research need to be widespread

and diverse to enable bottom-up innovation to meet top-down goals and ideals.4-7

Beneath Sendai’s overarching principles, fire science explores an expanding spectrum of fire-related social, economic, physical and agricultural sciences. This knowledge contributes to the successful and dynamic management of increasingly complex fire problems that affect human populations in a changing climate. This study contributes to that knowledge base. Implementing

the SFDRR has clear benefits including improved preparedness, and discerning ways to translate risk mitigation and reduction strategies into standard, practical applications to curb human suffering.8-10

Objective

The aim of this study is to contribute to the goals of the SFDRR by improving protection of human life and wellbeing in bushfire and

Public health and natural hazards: new policies and preparedness initiatives developed from an Australian bushfire case studyRachel Westcott,1,2 Kevin Ronan,2,3 Hilary Bambrick,4 Melanie Taylor2,5

1. Translational Health Research Institute, School of Medicine, Western Sydney University, Sydney, New South Wales 2. Bushfire and Natural Hazards Cooperative Research Centre, Melbourne, Victoria3. School of Health, Medical and Applied Sciences, Central Queensland University, Rockhampton, Queensland4. School of Public Health and Social Work, Queensland University of Technology, Brisbane, Queensland5. Department of Psychology, Macquarie University, Sydney, New South Wales Correspondence to: Dr Rachel Westcott, Translational Health Research Institute, School of Medicine, Western Sydney University, Locked Bag 1797, Penrith, New South Wales 2751;

email: [email protected]: October 2018; Revision requested: February 2019; Accepted: March 2019The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium,



Abstract

Objective: Public preparedness for natural hazard events is low. With worsening severe weather events due to climate change, public health policy and practices must evolve to more effectively engage communities. This study’s findings identify and suggest new strategic public health policies to shift the practice of all-hazards preparedness into routine, everyday life.

Methods: Semi-structured interviews, focus groups and Thematic Analysis were used to investigate the interactions between participant groups: emergency responders and animal owners.

Results: Three policies designed to improve human safety and well-being are proposed and discussed. These are (i) a new system of workplace leave, (ii) an innovative regime of financial incentives for fire-ready properties, and (iii) review of the use of firebreaks on farms and rural blocks.

Conclusion: Policies proposed in this research aim to proactively narrow the awareness-preparedness gap and build adaptive capacity to minimise risk to human health in all-hazards contexts. Further research could evaluate the efficacy of trialled public policy.

Implications for public health: These new policies seek to contribute to establishing and maintaining a culture of preparedness as a routine aspect of everyday life, and thus promote and protect public health in the short, medium and long terms.

Key words: bushfire, preparedness, public health policy, emergency responders, animal owners



Westcott et al. Article

other emergencies; to promote ‘fire-fitness’ (an original term coined by the lead author, described in detail on p3 of reference no. 11). This is achieved by establishing preparedness behaviours as routine – thereby reducing the awareness-preparedness gap (i.e. the mismatch between awareness and preventative action). From an economic perspective alone, the public cost of natural hazards in Australia is expected to “triple to US$17.7 billion by 2050”19 (equivalent to $AU24.7 billion). Adaptive capacity building is required in communities around the globe.

At present, household levels of fire-fitness11 in Australia and elsewhere are low, with fire-safe routines often assigned a lower priority than other competing complexities of everyday life.12-14 Practising considered, timely and safe action – to be fire-fit15-18 – within and outside the fire season is a present-day imperative.

The awareness-preparedness gap is narrowing disproportionately slowly compared with the magnitude of public resources assigned to help people attain readiness.12,13,20-22 Making safe, potentially life-saving fire preparedness behaviours a routine element of daily life is one of a suite of lifestyle changes people need to adopt due to the escalating influence of climate change on natural hazards.23 This paper proposes three areas of public health policy that aim to actively cultivate sustainable patterns of routine behaviours to better enable protection of lives and property, fortify psychological and physical preparedness and facilitate resilient and effective responses.

The emergencies literature currently lacks evidence from animal owners as a diverse whole. This study, with a combination of participant groups designed specifically to cross demographic boundaries, records and analyses some of the experiences, expectations and needs of communities who have ‘lived through’ bushfire emergencies, and expect to face this hazard again.

Emergencies can occur when people, property, the environment and other assets (including animals) intersect adversely with hazards.4 Animals may be considered ‘dependent others’ and their welfare is frequently linked to human physical and psychological health.24-26 For livestock farmers, an economic relationship does not exclude emotional attachment and both are considerations at the responder-owner interface.20,24,27

The presence of animals adds complexity to owners’ preparedness and planning.20,27 Incidents involving animals have been identified as a reason why people risk their own welfare and safety.28-31 There is also an increasing understanding of the link between effective animal management in an emergency and the saving of human life, and a growing awareness of the longer-term adverse human health implications of losing animals in an emergency incident.24

Responding safely and appropriately to a fire emergency is a realistically attainable goal – though frequently thwarted by the magnitude of the awareness-preparedness gap.20,27 To overcome this requires fire-fitness to be elevated to ‘business as usual’ status – as routine as buying groceries or fuelling a car. While the basic human urge to save a dependent other at the risk of personal safety may never be overcome, learned coping appraisals and adaptive responses, in combination with proactive preparedness routines as part of everyday living, could facilitate pre-hazard behaviours that overall reduce risk-taking while achieving a more effective response with less trauma and anxiety. Therefore, this paper’s research question is: what preparedness initiatives can be learned from the emergency responder-animal owner interface in a bushfire at-risk community that can be usefully applied to generate new public health policy?

Method

Research participants were firefighters, police officers, rescue officers of the State Emergency Service (SES), farmers with farm fire units and animal owners (a diverse group owning from one pet to thousands of livestock). Study participants resided in a bushfire at-risk regional area in South Australia – ‘the driest state in the driest continent’,32 chosen for its fire history.33

A situationalist orientation, i.e. the needs of the study govern a philosophical paradigm, indicated a pragmatic approach within a critical realist ontology and contextualist, experiential epistemology.27,34-36 Active recruitment by the researcher was assisted by leaders in the responder groups. Local businesses with an agriculture or animal health focus were invited to participate. Local media helped raise awareness of the project, and information flyers were placed in public places such as the local Council offices, public library and some retail outlets.

Interested potential participants contacted the researcher and some invited others to join from within their own networks. Prior to taking part, all participants received information sheets covering ethics approval, privacy and contact details, and signed a consent form.

Data were collected from 67 participants via 12 semi-structured individual interviews and seven focus groups (n=55), each between 45 and 90 minutes duration. Gender distribution was 46.3% female (n=31) and 53.7% male (n=36). All participants were aged between 18 and 70 years. Approximately two-thirds of participants had experienced fire on their properties.

The interview guide was flexible according to group context and composition.37 Major topics were: (1) hazard severity and likelihood; (2) fire-related animal issues; (3) information gathering, communication and trust; (4) uncertainty and confidence; (5) mitigation and self-efficacy; and (6) special circumstances and adaptive solutions.

Thematic Analysis (TA) was chosen because it is a flexible qualitative method independent of theory.35,38 Extraction of experiential material from the data was inductive and contextualist: analysis moved from descriptive to interpretative and explored latent meanings. Data-driven coding yielded codes that were grouped into ‘like’ clusters and then organised into 10 themes. Data was managed using the Computer Assisted Qualitative Data Analysis Software (CAQDAS) system, NVivo 11, on a spreadsheet and a thematic map and table.11

Ethics approval for this research was granted by the Human Research Ethics Committee of Western Sydney University, approval number H11118. Names assigned to data extracts are pseudonyms.

Results and discussion

Of the 10 actively identified themes in the study, this paper examined data from the seminal ‘preparedness’ theme and from the ‘farmers’ theme. In summary, the themes were (1) animal owners and farmers; (2) Preparedness, fire-fitness; (3) Complexity of the social microclimate; (4) Trust; (5) Information gathering; (6) Responders; (7) Adaptive safe responses; (8) Maladaptive, unsafe responses; (9) The “tree-changers”; and (10) Recovery – and are discussed elsewhere in the published literature (tabulated themes


General Public Health Public health preparedness in natural hazards

are described in concise detail on p 217 of Reference No. 11 ).11,15,27

A serious fire affecting people, their livelihoods and microclimates is a complex non-routine social problem that falls within the remit of the SFDRR.39 Discerning how people and emergency managers and responders can better equip communities to protect themselves, and the things they hold dear, is an urgent requirement given the increasingly severe weather conditions that indicate a ‘new reality’.40 Effectively addressing this requires prioritising innovative preparedness initiatives.40-45

To achieve and maintain fire-fitness, it is necessary to understand and establish prerequisite conditions that precede and predispose towards successful preparedness messaging and action outcomes. This foundation, built on medium- to long-term strategies, will help develop a culture of preparedness and is required before a substantial shift in the implementation of preparedness practice is generally evident.

Preparedness – Be fire-fit: weekly is worth it!The preparedness theme ‘Be fire-fit: weekly is worth it!’ was prominent in the data and is the subject theme of this paper because the implied corollary of being prepared (fire-fit), and of frequency (weekly), is a net benefit (is worth it). This theme is pivotal to addressing the awareness-preparedness gap and achieving fluency between knowledge and action among people at all levels – by linking science with policy and evidence with implementation.1,4,27

The academic literature exploring preparedness considers psychological and physical capability and suggests reasons why people do, or do not, prepare.22,46-50 One contributing factor is the dilemma of competing superimposed tasks.21 For example, the concurrent desire to save family and home and property can result in action inertia. The present study argues that by promoting preparedness as ‘business as usual’ both outcomes are achievable and could result in a healthier outcome experience with less physical and/or psychological trauma. In turn, safely protecting property, including animals, contributes to building confidence, resilience and well-being, as espoused in the SFDRR.6,51,52

There are some limitations: affording equipment is a potentially limiting factor as

not everyone will have the disposable income or the resources to reach their ideal level of preparedness in one fire season, particularly in rural areas where income may be sporadic. However, a bushfire plan can dynamically map a strategy to attain the desired level of preparedness over a specified time frame. This leads to adaptive capability and confidence – self-efficacy and response-efficacy.53,54 The challenge remains how to engage with those who elect not to prepare their homes, their properties and their social microclimate (such as family or workplace group). Complete consensus is unlikely and some people will remain unconvinced – the problem then is how to help people on adjacent land49 and people who live on the outer peri-urban fringe because they prefer to live with less social interaction. Local knowledge in such instances can literally mean the difference between life and death.55,56

Bushfire prevention and preparedness is promoted in Australia and internationally as everyone’s responsibility.57-61 This is not intended to preference the actions of an individual over the involvement of community and collaboration between people. Both are important and mutually inclusive. Both benefit from shared communication and from the synergy achieved by collaboration among a group of people with a common goal.

Social connectedness and community engagement can reduce the negative outcomes of natural hazard emergencies.12,46,62,63 Akama and Ivanka30 discuss the need to understand and promote the real meaning of ‘community’; the creation of sub-groups bonded by a common goal, and how self-empowerment can catalyse behaviour change.12,51,64,65 Individuals, community groups, local, state and federal governments, workplaces and policy development can all contribute to and promote this change, increasing the status of a culture of preparedness and fire-fitness – to make investment in resilience “gainful”.66

However, self-responsibility is the pre-requisite building block for a strong community effort. Jayne explained:

Protection is about self-help as much as it is about relying on the services that you’ve got … being bush fire ready isn’t easy and simple and quick and cheap. It’s not that hard … if you just want to pack and go. But being bush fire ready is no different to any other problem or complexity that people have in their life. So, and I’m really upfront with people, I will

say forget it. If you think you’re going to do [everything] by tomorrow – no. So get rid of those unrealistic expectations … you can’t do 20 or 30 or 40 jobs when you can smell the smoke, you can only probably do one or two.

The danger of ‘waiting to smell the smoke’ is described by Penman et al.49 – late decisions are made under duress, with potentially fatal consequences.

The critical challenge is to first defuse the sometimes overwhelming nature of the preparedness task, and to facilitate the transition from knowledge and understanding to intention and action. Breaking tasks down into manageable steps and writing a ‘bushfire action plan’ to reduce the need for strategic thinking when an emergency situation arises is one way to achieve the former, and is already actively encouraged by fire authorities as part of ongoing multi-media public outreach. However, an environment conducive to achieving effective action must necessarily occur before preparedness can be substantially realised – before advertising and use of messages intended to motivate the target audience with fear or ‘shock tactics’, the effect of which can be short-lived.67,68 This requires a cultural, paradigm shift, which itself can be created incrementally3 via a foundation that preferences and facilitates routine, effective preparedness activities. Thus, hurdles such as lack of time, or the dangerous maladaptive responses of ‘action inertia’21 or acting impulsively without even a brief dynamic risk assessment can be overcome.

Shaping policy – cultivating a culture of preparednessFuture natural hazards are likely to increase in severity and frequency due to climate change.45,69-71 For this reason, a greater knowledge-base is urgently needed to shape policy for disaster preparedness and response48 (emphasis added).

Proactively promoting preparedness and the capability to effectively manage risk needs strategic awareness and a concurrent problem-solving approach.12,30,64,72 At the local level, participants identified several common barriers to preparedness – including lack of time, resources, knowledge or information – as well as the problem of how to act appropriately on days that are declared catastrophic (or the equivalent jurisdictional nomenclature).73-76 Although this terminology


can be shocking for people unfamiliar with Australian bushfire weather conditions, it realistically represents weather conditions that favour the ignition of potentially uncontrollable fires threatening public health and safety. As well, disaster literacy of vulnerable demographics and populations needs to be addressed. Programs aimed at broad acceptance and application need to be piloted and evaluated with respect to everyone in the community. The quality of public health emergency messaging must be rapid, accurate and useful.77,78 Establishing a normalised culture of fire-fitness could arguably greatly assist in improving disaster literacy. Further commitment is needed from government to actively demonstrate a proactive approach to building a culture of preparedness from new, evidence-based initiatives by trialling and evaluating innovative strategies, as discussed below.

New policy – Catastrophic Day Leave (CDL) This research proposes instigating workplace agreements to help narrow the awareness-preparedness gap. On days of high fire danger, people are faced with the dilemma of how to manage required tasks even if they have a well-written bushfire survival plan. Catastrophic Day Leave (CDL) could effectively assist to alleviate the dilemma. The concept of CDL is an analytic construct – where the analysis shifts to a more constructionist and critically interrogative style.34,35

Employer-employee negotiations could ‘trade’ other workplace leave for a certain number of CDL days, or work an extra hour a day for eight or nine days a fortnight to accrue CDL days. Wilkinson et al.31 report varied and, at times, problematic employee experiences with employers when requesting leave of absence during the 2013 ‘Red October’ bushfires in New South Wales, Australia. A formal contractual arrangement for CDL with employers could obviate this difficulty and promote shared responsibility with mutual workplace benefits. Initiating CDL as a new form of workplace leave would have the dual effect of elevating a culture of bushfire preparedness to ‘business as usual’ status, thus raising active awareness of the need to prepare well in the wider community and enabling employees to act safely in a timely manner. For these reasons it is important to name this proposed leave according to the purpose for which it is intended: generic

‘personal leave’, which may be made available to employees for many different reasons, does not satisfy this requirement. This type of initiative is representative of new policy, which will be necessary to manage the impact of climate-change induced, worsening natural hazards.41,48,79

Potentially, colleagues, workmates and neighbours encouraged by a CDL policy would be prompted to actively instigate shared plans and arrangements within their community networks. This could help build stronger productive relationships between communities and responders, and help communities better equip themselves to confront barriers to preparedness, and dismantle them step by step.30,80,81 Subsequently, improved communication and safer decision-making between all parties supplements knowledge bases formulated cooperatively and collaboratively across government agencies, fire authorities, research findings and community members, depicting the synergistic interface of science and policy in the SFDRR.

Financial incentives and rewarding best practiceFinancial inducement or reward can help achieve a societal shift towards establishing a culture of preparedness by implementing a system of rebates or discounts on insurance premiums, local government charges or other taxes,82 and by actively rewarding ‘best practice’. Sandy, in the business focus group, unhesitatingly commented, “People respond very well to financial incentive. There needs to be an incentive for groups to actually come together and discuss things.”

An example is the French CatNat scheme (Catastrophes Naturelles), a public/private scheme based on the principle of national solidarity: everyone pays for the benefit of the common good.82,83 In France, household policies cover ‘insurable’ risk, and the CatNat scheme, created by law in 1982, is designed for events considered uninsurable, such as natural disasters. It is based on paragraph 12 of the preamble of the Constitution of 27 October 1946, which states: “The Nation declares all French citizens to be equal and united in solidarity when faced with loss resulting from natural disasters”.84

Residents who are well prepared and fully insured need to be recognised and acknowledged for their contribution to public health. A scheme that rewards excellence in

preparedness and property management could be linked to an existing Local Government inspectorate authorised to issue fines for the reverse. Additional workload and costs would foreseeably be offset by savings given the high cost of recovery after an emergency event.19,85

A financial incentive for new residents needing to increase their bushfire knowledge for their own and their community’s health and safety could be achieved by offering discounts linked to their attendance at non-compulsory community fire-safety information sessions. They could be encouraged to do so via an invitation accompanying their first Local Government rates notice, offering all attendees a meaningful discount to be applied to the second year’s fees. To qualify, participation in a given number of fire information seminars would be required, which could be spread over a 12-month period to give maximum opportunity for people to attend. Senior firefighter Shane recalled an observation he often makes to newcomers to the community regarding shared responsibility, “I point out there are three fire trucks sitting in that shed and six hundred homes over that hill”. Costs could be met by savings against recovery.19,85

In the longer term, public awards and recognition such as ‘Bushfire Best-Prepared Towns’, could attract additional funding from government or corporate sources and boost the local tourist economy due to increased publicity, or if preferentially considered as a holiday destination. Proactively promoting a culture of bushfire safety in this way builds community pride as well as strong relationships with emergency services.

Value-adding to properties at point-of-sale by making bushfire compliance a desirable, marketable commodity is another financial incentive. This could be achieved by adding a notation on advertising material identifying ‘bushfire-safer properties’ compliant with current relevant Standards,86 and encourage others to similarly ‘value-add’. This strategy would need to be aligned with a formal system of acknowledging eligible properties. Qualifying properties could be given the option of displaying a gateway notice, or participate in community ‘fire-ready’ open days, similar to the familiar ‘open gardens’, to showcase and educate others to do likewise.

Overarching jurisdictional in-principle support for strategies involving financial incentives is needed, but the success of



a scheme could well depend upon local knowledge and respectful local community consultation as strategies to build ‘fire-fitness’ may be best managed on a locally bespoke basis.20 Whole-of-jurisdiction plans may not be functional if applied state-wide beyond the parameters of local conditions. This has been a failing in previous attempts to successfully apply discounts on insurance schemes, for example, and has been asserted as a reason not to pursue financial incentives. Desirable choices can be positively influenced by the magnitude of reward89,90 and proactive, locally appropriate successful applications could motivate others in the area. This research asserts the need for such an initiative to be trialled and evaluated.82,83,87,88

Farming practices, fuel loads and firebreaksMost farmer participants agreed that modern farming techniques could influence fire behaviour. Practices such as no-till cropping, greater crop productivity, density of crop per hectare, improved plant structure and reduced farm firebreaks have the potential to significantly compound the complexities of a fire. How these issues are managed is likely to influence preparedness strategies and tactics.

Sheep and wheat farmer Paul noted:

I think with our modern farming and agricultural techniques we’re achieving crop yields that are way and above what we’ve ever been able to do in the past … in 30 years, I’ve seen cereal yields double, and what that means, of course, is that there’s double the amount of crop residue over the summer period after the crops have been taken off, and the proportion of arable land going into crops has increased also.

Firebreaks have fallen out of favour, seemingly because of potential economic losses associated with decreased crop areas. Paul added:

There are fewer fire breaks across the landscape. Once upon a time, farmers were quite diligent about preparing firebreaks – they give you something to burn back to. This could be made mandatory with a council by-law, so everyone has to do it. A little bit of loss could mean that a lot of people are safer. It would be better to see more fire breaks across the landscape. I think that’s something that we could consider … a by-law type of arrangement for strategic fire breaks.

Farmers who choose to implement effective fire breaks, whether or not required by regulation, could offset potential economic loss to some degree by being rewarded for on-farm best-practice preparedness, again motivating fire-fitness practice.

Conclusion and implications for public health

The outcomes of this study are intended to be transformative in that the new, public health preparedness initiatives proposed here aim to be practical and realistic. They seek to motivate the translation of knowledge into effective, adaptive actions attainable by all residents of bushfire at-risk communities. Making good preparedness behaviour and practices routine – and thereby narrowing the awareness-preparedness gap – requires all stakeholders to undertake a proactive reassessment of how to ‘do’ preparedness and become fire-fit. As evidenced by participants in this study, such a reassessment would help reduce hazard-related human mortality and morbidity and the associated negative social, economic and environmental impacts of natural hazard emergencies. The premise on which existing public resources are founded needs a proactive re-evaluation to help establish a culture of preparedness as ‘business as usual’ in society generally. Until that culture becomes established, this research suggests that more of the same messaging, however professional and sophisticated, will not significantly, nor sufficiently, narrow the awareness-preparedness gap without the help of additional, supplementary strategies. New social and workplace policies that are practical and achievable such as Catastrophic Day Leave, financial incentives such as rewards for ‘best practice’ and reductions in municipal fees, and mandatory fire breaks on farms have the potential to cultivate a more desirable culture of routine preparedness.

The implementation, evolution and efficacy of such applied preparedness initiatives and policy, broadly translatable across many societal groups, will need to be evaluated by future research – following preliminary scoping studies to assess acceptance within a community. Given the probability of increased fire weather and fire severity, and as documented in the SFDRR, the challenge of fortifying community wellbeing in a bushfire emergency requires a dynamic, problem-solving paradigm melded from science,

government and the at-risk communities themselves. Based on current findings, this paper concludes that implementing new practical and achievable policies that work across social and workplace contexts are steps toward achieving this goal.

References1. United Nations Office for Disaster Risk Reduction.

Sendai Framework for Disaster Risk Reduction. Geneva (CHE): UNISDR; 2015.

2. Trogrlić RŠ, Cumiskey L, Triyanti A, Duncan MJ, Eltinay N, Hogeboom RJ, et al. Science and technology networks: A helping hand to boost implementation of the sendai framework for disaster risk reduction 2015–2030? Int J Disaster Risk Sci. 2017;8(1):100-5.

3. Bambrick H, Moncada S. From social reform to social transformation: Human ecological systems and adaptation to a more hostile climate. In: Health of People, Places and Planet: Reflections Based on Tony McMichael’s Four Decades of Contribution to epidemiological Understanding. Canberra (AUST): ANU Press; 2015. p. 353.

4. Aitsi-Selmi A, Egawa S, Sasaki H, Wannous C, Murray V. The Sendai Framework for Disaster Risk Reduction: Renewing the global commitment to people’s resilience, health, and well-being. Int J Disaster Risk Sci. 2015;6(2):164-76.

5. Aitsi-Selmi A, Murray V. The Sendai Framework: Disaster Risk Reduction through a health lens. Bull World Health Organ. 2015;93(6):362-.

6. Aitsi-Selmi A, Murray V. Protecting the health and well-being of populations from disasters: health and health care in The Sendai Framework for Disaster Risk Reduction 2015-2030. Prehosp Disaster Med. 2016;31(1):74-8.

7. Blanchard K, Aitsi-Selmi A, Murray V. The Sendai Framework on Disaster Risk Reduction: From science and technology to societal resilience. Int J Disaster Resil Built Environ. 2015;6(2): https://doi.org/10.1108/IJDRBE-04-2015-0019

8. McClean D. Global Platform Stresses Action on Sendai Cancun. Geneva (CHE): United Nations Office for Disaster Risk Reduction; 2017.

9. Raine J. Sendai Framework for Disaster Risk Reduction 2015-2030 - Australia’s Approach. Canberra (AUST): Government of Australia, Attorney-General’s Department; 2017.

10. Attorney General’s Department. Emergency Management Australia. Canberra (AUST): Government of Australia; 2017.

11. Westcott R. Advancing Public Health in the Context of Natural Hazards: Normalising Preparedness within a Framework of Adapted Protection Motivation Theory. Sydney (AUST): Western Sydney University; 2018.

12. Paton D. Disaster resilient communities: Developing and testing an all-hazards theory. J Integr Disaster Risk Manag. 2013;3(1):1-17.

13. Ronan KR, Johnston DM. Promoting Community Resilience in Disasters: The Role for Schools, Youth, and Families. New York (NY): Springer; 2005.

14. Paton D, Johnston D. Disasters and communities: Vulnerability, resilience and preparedness. Disaster Prev Manag. 2001;10(4):270-7.

15. Westcott R, Mitigating action inertia and the bushfire awareness-action gap: Findings from a South Australian Case Study. Proceedings of the Australian and New Zealand Disaster and Emergency Management Conference; 2017 May 22-23; Gold Coast, Queensland. Nerang, Queensland: Australian and New Zealand Mental Health Association; 2017.

16. Westcott R. PreventionWeb DRR Voices [Internet]. Geneva (CHE): United Nations Office for Disaster Risk Reduction; 2017 [cited 2017 Oct 11]. Available from: http://www.preventionweb.net/experts/oped/view/54363

17. Westcott R. Narrowing the awareness-action gap: Cultivating fire-fitness as a social norm through public policy initiatives. Aust J Emerg Manag. 2017;31(4):37-41.

General Public Health Public health preparedness in natural hazards


18. Westcott R, Ronan K, Bambrick H, Taylor M. Bushfire Preparedness: How to Become ‘Fire-fit’ Without Really Noticing. Adelaide (AUST): Bushfire and Natural Hazards Cooperative Research Centre; 2017.

19. McClean D. Australia Integrates Climate and Disaster Risk. Geneva (CHE): United Nations Office for Disaster Risk Reduction; 2017.

20. Westcott R, Ronan K, Bambrick H, Taylor M. Expanding protection motivation theory: Investigating an application to animal owners and emergency responders in bushfire emergencies. BMC Psychol. 2017;5(1):13.

21. McNeill IM, Dunlop PD, Skinner TC, Morrison DL. Predicting delay in residents’ decisions on defending v. evacuating through antecedents of decision avoidance. Int J Wildland Fire. 2015;24:153-61.

22. Grothmann T, Reusswig F. People at risk of flooding: Why some residents take precautionaryacton while others do not. Nat Hazards. 2006;38:101-20.

23. Banwell C, Dixon J, Bambrick H, Edwards F, Kjellstrom T. The sociocultural context of climate change adaptation in Australia, with a focus on hot weather. In: Butler CD, Dixon J, Capon AG, editors. Health of People, Places and Planet: Reflections Based on Tony McMichael’s Four Decades of Contribution to Epidemiological Understanding. Canberra (AUST): Australian National University; 2015. p. 379.

24. Chur-Hansen A. Grief and bereavement issues and the loss of a companion animal: People living with a companion animal, owners of livestock and animal support workers. Clin Psychol. 2010;14(1):14-21.

25. Cordaro M. Pet loss and disenfranchised grief: Implications for mental health counseling practice. J Ment Health Couns. 2012;34(4):283-94.

26. Heath SE, Linnabary RD. Challenges of Managing Animals in Disasters in the U.S. Animals. 2015;5(2): 173-92.

27. Westcott R, Ronan K, Bambrick H, Taylor M. Don’t just do something … stand there! emergency responders’ peri-incident perceptions of animal owners in bushfire. Front Vet Sci. 2017;4:34.

28. Thompson K. Save me, save my dog: Increasing natural disaster preparedness and survival by addressing human-animal relationships. Aust J Commun. 2013;40(1):123-36.

29. Hall MJ, Ng A, Ursano RJ, Holloway H, Fullerton C, Casper J. Psychological impact of the animal-human bond in disaster preparedness and response. J Psychiatr Pract. 2004;10(6):368-74.

30. Akama Y, Ivanka T. What community?: Facilitating awareness of ‘community’ through playful triggers. Proceedings of the 11th Biennial Participatory Design Conference; 2010 Nov 29 - Dec 3; Sydney, Australia. New York: ACM; 2010. p. 11-20.

31. Wilkinson C, Eriksen C, Penman T. Into the firing line: Civilian ingress during the 2013 “Red October” bushfires, Australia. Nat Hazards. 2016;80(1):521-38.

32. Department of Environment Water and Natural Resources. Managing Natural Resources. Adelaide (AUST): State Government of South Australia; 2016.

33. South Australian Country Fire Service. Bushfire History. Adelaide (AUST): SA CFS; 2016.

34. Savin-Baden M, Major CH. Qualitative Research. Abingdon (UK): Routledge; 2013.

35. Braun V, Clarke V. Successful Qualitative Research. London (UK): Sage; 2013.

36. Cornish F, Gillespie A. A pragmatist approach to the problem of knowledge in health psychology. J Health Psychol. 2009;14(6):800-9.

37. Taylor GW, Ussher JM. Making sense of S&M: A discourse analytic account. Sexualities. 2001(4):293-314.

38. Braun V, Clarke V. Using thematic ananlysis in psychology. Qual Res Psychol. 2006;3(2):77-101.

39. Drabek TE. Social Dimensions of Disaster. 2nd ed. Emmitsburg (MD): Emergency Management Institute Federal Emergency Management Agency; 2004.

40. Golgowski N. Wildfires are the ‘new reality’ for California, Gov. Brown warns. The Huffington Post Australia; 2017;Dec 21.

41. Gibbs L, Bryant R, Harms L, Forbes D, Block K, Gallagher HC, et al. Beyond Bushfires: Community Resilience and Recovery Final Report. Melbourne (AUST): University of Melbourne, 2016 November.

42. New Scientist. Wild weather: extreme is the new normal. New Scientist. 2013;Jan 19:2900.

43. Beynon H. Extreme Weather the “New Normal”. New York (NY): United Nations Office for Disaster Risk Reduction; 2016.

44. Lewis S. Time to Adapt: 2015’s Record-breaking Temperatures will be Normal by 2030. Sydney (AUST): Australian Broadcasting Corporation; 2016.

45. Steffen W, Hughes L, Alexander D, Rice M. Cranking up the Intensity: Climate Change and Extreme Weather Events. Potts Point (AUST): Climate Council of Australia; 2017.

46. Kanakis K, McShane CJ. Preparing for disaster: Preparedness in a flood cand cyclone prone community. Aust J Emerg Manag. 2016;31(2):18-24.

47. Eriksen C, Prior T. Defining the importance of mental preparedness for risk communication and residents well-prepared for wildfire. Int J Disaster Risk Reduct. 2013;6:87-97.

48. Gibbs L, Waters E, Bryant RA, Pattison P, Lusher D, Harms L, et al. Beyond bushfires: Community, resilience and recovery - a longitudinal mixed method study of the medium to long term impacts of bushfires on mental health and social connectedness. BMC Public Health. 2013;13:1036.

49. Penman TD, Eriksen C, Blanchi R, Chladil M, Gill AM, Haynes K, et al. Defining adequate means of residents to prepare property for protection from wildfire. Int J Disaster Risk Reduct. 2013;6:67-77.

50. Prior T, Eriksen C. Wildfire preparedness, community cohesion and social–ecological systems. Glob Environ Change. 2013;23(6):1575-86.

51. Warrandyte Community Association. Be ready Warrandyte. Melbourne (AUST): State Government of Victoria; 2013.

52. Australian Academy of Science. Theo Murphy High Flyers Think Tanks. Canberra (AUST): AAS; 2014.

53. Grothmann T, Patt A. Adaptive capacity and human cognition: The process of individual adaptation to climate change. Glob Environ Change. 2005;15:199-213.

54. Lwin MO, Saw S-M. Protecting children from myopia: A PMT perspective for improving health marketing communications. J Health Commun. 2007;12(3):251-68.

55. Blaikie P, Cannon T, Davis I, Wisner B. At Risk: Natural Hazards, People’s Vulnerability and Disasters. London (UK): Routledge; 2014.

56. Elsworth G, Gilbert J, Rhodes A, Goodman H. Community safety programs for bushfire: what do they achieve, and how? Aust J Emerg Manag. 2009;24(2):17-25.

57. South Australian Country Fire Service. Preparing Yourself for a Bushfire. Adelaide (AUST): State Government of South Australia; 2017.

58. Smalley JCE. Protecting Life and Property from Wildfire. Quincy (MA): National Fire Protection Association; 2005.

59. Ministry of Forests Lands and Natural Resource Operations. FACTSHEET: Wildfire Prevention is Everyone’s Responsibility. Vancouver (CAN): Government of British Columbia; 2016.

60. Country Fire Authority. Plan and Prepare. Melbourne (AUST): State Government of Victoria; 2016.

61. Adelaide Hills Council. Fire Prevention in the Adelaide Hills: It’s Everyone’s Responsibility. Adelaide (AUST): State Government of South Australia; 2016.

62. Paton D, McClure J. Preparing for Disaster: Building Household and Community Capacity. Springfield (Ill): Charles C. Thomas; 2013.

63. Gordon R. Community impact of disaster and community recovery. InPsych. 2009;31(2):12-13.

64. Akama Y. Passing on, handing over, letting go – the passage of embodied design methods for disaster preparedness. Proceedings of the Service Design and Innovation Conference; 25 June; Lancaster University: Linköping University Electronic Press; 2014. p. 173-83.

65. Akama Y. Birds of a Feather. Melbourne (AUST): RMIT University; 2012.

66. United Nations Office for Disaster Risk Reduction. Leaders’ Forum for Disaster Risk Reduction. The Cancun High-Level Communiqué - Ensuring the Resilience of Infrastructure and Housing. Cancun (MEX): UNISDR; 2017 May 26.

67. Tanner JF, Day E, Crask MR. Protection motivation theory: An extension of fear appeals theory in communication. J Bus Res. 1989;19(4):267-76.

68. So J. A further extension of the extended parallel process model (e-eppm): implications of cognitive appraisal theory of emotion and dispositional coping style. Health Commun. 2013;28(1):72-83.

69. Liu Y, Stanturf J, Goodrick S. Trends in global wildfire potential in a changing climate. For Ecol Manage. 2010;259(4):685-97.

70. Intergovernmental Panel on Climate Change. Fifth Assessment Report. Geneva (CHE): IPCC; 2014.

71. Kitching R. IPCC: Australia and New Zealand Face Greater Fire and Flood Risk, Damage to Coral Reefs. The Conversation. 2014; March 31.

72. Eriksen C, Gill N, Bradstock R. Trial by fire: Natural hazards, mixed-methods and cultural research. Aust Geogr. 2011;42(1):19-40.

73. South Australian Country Fire Service. More about Fire Danger Ratings. Adelaide (AUST): State Government of South Australia; 2017.

74. Department of Fire and Emergency Services. Fire Danger Ratings. Perth (AUST): State Government of Western Australia; 2018.

75. Natural Resources Canada. Canadian Wildland Fire Information System. Montreal (CAN): Government of Canada; 2017.

76. United States Forest Service. Wildland Fire Assessment System. Washington (DC): US Department of Agriculture; 2017.

77. Brown LM, Haun JN, Peterson L. A proposed disaster literacy model. Disaster Med Public Health Prep. 2014;8(3):267-75.

78. Savoia E, Lin L, Viswanath K. Communications in public health emergency preparedness: A systematic review of the literature. Biosecur Bioterrore. 2013;11(3):170-84.

79. Council of Australian Governments. National Strategy for Disaster Resilience: Building Our Nation’s Resilience to Disasters. Canberra (AUST): Government of Australia Department of the Prime Minister and Cabinet; 2011.

80. Paton D. Preparing for natural hazards: The role of community trust. Disaster Prev Manag. 2007;16(3): 370-9.

81. Boura J. Community fireguard: Creating partnerships with the community to minimise the impact of bushfire. Aust J Emerg Manag. 1998;13(3):59-64.

82. Poussin JK, Botzen WJW, Aerts JCJH. Factors of influence on flood damage mitigation behaviour by households. Environ Sci Policy. 2014;40:69-77.

83. Poussin JK, Botzen WJW, Aerts JCJH. Stimulating flood mitigation through insurance: an assessment of the French CatNat system. Environ Hazards. 2013;12(3-4):258-77.

84. Caisse Centrale de Réassurance (CCR). Natural Disasters Compensation Scheme. Paris (FRA): CCR; 2016.

85. Attorney-General’s Department. Productivity Commission Inquiry into Natural Disaster Funding Arrangements. Canberra (AUST ): Australian Government Attorney General’s Department; 2014. p. 12-22.

86. Standards Australia. Australian Standard® Construction of buildings in bushfire prone areas. Sydney: Council of Standards Australia; 2009.

87. Van den Bergh R, Faure M. Compulsory insurance of loss to property caused by natural disasters: Competition or solidarity? World Compet. 2006;29(1):25-54.

88. Penman TD, Eriksen CE, Horsey B, Bradstock RA. How much does it cost residents to prepare their property for wildfire? Int J Disaster Risk Reduct. 2016;16 Suppl C:88-98.

89. Kubanek J, Snyder LH, Abrams RA. Reward and punishment act as distinct factors in guiding behavior. Cognition. 2015;139:154-67.

90. Gajda K, Sülzenbrück S, Heuer H. Financial incentives enhance adaptation to a sensorimotor transformation. Exp Brain Res. 2016;234(10):2859-68.



The authors have stated they have no conflict of interest.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is

properly cited, the use is non-commercial and no modifications or adaptations are made.

doi: 10.1111/1753-6405.12907

Observed vaping and smoking in outdoor public places: piloting objective data collection for policies on outdoor vapingGeorge Thomson,1 Johanna Nee-Nee,1 Kirsty Sutherland,1 Rebecca Holland,1 Miriam Wilson,1 Nick Wilson1

1. Department of Public Health, University of Otago, Wellington, New Zealand

The impacts of e-cigarettes, also known as vaping products, are highly debated1-3 and vaping in public places raises further issues.4 As we found no published observations of the point prevalence of vaping in outdoor places, we aimed to pilot: i) the assessment of the point prevalence of vaping, and the relative prevalence of smoking and vaping outside hospitality venues; and ii) the assessment of the relative prevalence of smoking and vaping among those walking in downtown pavement areas. The context is a high-income country (New Zealand) where e-cigarette use has been officially encouraged since 20175 and which has the second-highest level of vaping in a 14-country study (at 7.8% for daily use of nicotine vaping products among smokers and recent ex-smokers6). There is no legislation in New Zealand requiring smokefree or vapefree areas outside hospitality venues.

Methods

Observations were made of patrons, smokers, vapers and children sitting in outside areas of 56 hospitality venues in central Wellington City, New Zealand. These were taken between 3.30pm and 9.00pm on weekdays and 12.00pm and 9.00pm on weekends from 16 to 27 May 2018. The longer weekend observation times were to increase the efficiency of data collection (since more people were at hospitality venues on weekends) in a time-constrained study. Ten-minute observations were also made of the number of active smokers and active vapers passing within a five-metre radius in three defined pavement areas. Detailed methods are available online (see Supplementary File). Wellington has some voluntary smokefree outdoor areas, but these are not situated where the observations were made.7 Data

were collected on smartphones by filling in a standardised online form. Weather data was recorded at 5.00pm daily. Ethics approval (D18/121) was obtained from the University of Otago.

Results

There was rain on 5/10 days of field work with an average temperature of 14°C (range 10.5–16.9°C) and average wind speed 27kmph (5–45kmph). Totals of 7,977 adult patrons, 214 child patrons (8,191 patrons in total), and 114 active vapers were observed during 2,422 venue observations. Active vapers were 6.12 times more likely to be observed at venues without children present (2,355 venue observations, 113 active vapers), compared to venues with children present (67 venue observations, three active vapers; 95%CI: 1.9 to 19.2).

A ratio of 10 active smokers to one active vaper was observed at the venues (1,113/116), with a point prevalence of vaping at 1.5% of patrons (116/7,977). During 121 static observations of pedestrians at the three pavement locations, a ratio of 2.9 active smokers to one active vaper was seen (120 vapers, 350 smokers – Table 1). On average, six active vapers and 17 active smokers were observed per hour across these three locations. Observers reported that it was sometimes difficult to distinguish vaping devices from other handheld items and, in the evenings, it was difficult to see the devices compared to cigarettes as they do not light up. However, vaping clouds were easy to identify.

Conclusions

As countries increasingly consider constraining smoking and vaping in outdoor settings,8,9 it is important to base such policies on objective data. Vaping can be done stealthily10 and the devices can be pocketed between puffs, so observers may be less likely to see vaping compared to smoking. Further studies are required to assess vaping in other settings and jurisdictions, and to assess trends in visible vaping over time.

Implications for public health

Having the same rules for vaping and smoking for particular outdoor places provides simplicity in policy communication and implementation. Vaping outdoors raises the issue of the exposure of children and non-smokers to the normalisation of the activity, and the possible contribution from normalisation to the adoption of vaping by these groups. The use of nicotine by non-smoking youth is a matter of public health concern.11-13 Some of the health issues around vaping indoors may apply to outdoor areas where people are close together.4

While our results indicated that in this setting there was less vaping than smoking, the balance is likely to change over time as smoking decreases and vaping increases. Policies on vaping outdoors also need to consider the long-term needs of a future post-tobacco smoking situation, where the consumption of nicotine by other means may be seen as a public health issue.

Table 1: Comparison of active vaping and active smoking by people walking within 5m of observer at three different static observation points recorded in 10 min blocks by location and time of day in Wellington city, May 2018.Setting and time of observation

Static observations

(n)

Active vapers

(n)

Active smokers

(n)

Rate of active vaping observed per

hour

Rate of active smoking

observed per hour

Rate ratio of active smokers

to active vapers

Cuba Street12 – 4pm 17 28 76 9.9 26.8 2.74 – 9pm 25 37 107 8.9 25.7 2.9Waterfront12 – 4pm 16 4 10 1.5 3.8 2.54 – 9pm 24 17 19 4.3 4.8 1.1Courtenay Place12 – 4pm 17 18 66 6.4 23.3 3.74 – 9pm 22 16 72 4.4 19.6 4.5Total12 – 4pm 50 50 152 6 18.2 3.04 – 9pm 71 70 198 5.9 16.7 2.8All times all locations 121 120 350 6.0 17.4 2.9

Letter


PUBLIC HEALTH ASSOCIATION OF AUSTRALIA

Working together for better health outcomes

Join today and be part of it! Public Health Association

A U S T R A L I A

www.phaa.net.au

Letter

References1. Kalkhoran S, Glantz SA. E-cigarettes and smoking

cessation in real-world and clinical settings: A systematic review and meta-analysis. Lancet Respir Med. 2016;4(2):116-28.

2. Warner KE. How to think - not feel - about tobacco harm reduction. Nicotine Tob Res. 2018 Apr 30. doi: 10.1093/ntr/nty084.

3. Hartwell G, Thomas S, Egan M, et al. E-cigarettes and equity: A systematic review of differences in awareness and use between sociodemographic groups. Tob Control. 2017;26(e2):85-91.

4. Wilson N, Hoek J, Thomson G, et al. Should e-cigarette use be included in indoor smoking bans? Bull World Health Organ. 2017;95(7):540-1.

5. New Zealand Ministry of Health. Ministry of Health Position Statement – Vaping Products [Internet]. Wellington (NZ): Government of New Zealand; 2017 [cited 2018 Aug 30] October 11. Available from: https://www.health.govt.nz/our-work/preventative-health-wellness/tobaccocontrol/vaping-smokeless-including-heated-tobacco

6. Gravely S, Driezen P, Ouimet J, et al. Prevalence of awareness, ever-use and current use of nicotine vaping products (NVPs) among adult current smokers and ex-smokers in 14 countries with differing regulations on sales and marketing of NVPs: Cross-sectional findings from the ITC Project. Addiction. 2019 Jan 25. doi: 10.1111/add.14558

7. Wellington City Council. Smokefree Wellington [Internet]. Wellington (NZ): WCC; 2018 [cited 2018 Aug 30]. Available from: https://wellington.govt.nz/your-council/plans-policies-and-bylaws/policies/smokefree-wellington

8. Marynak K, Kenemer B, King BA, et al. State Laws Regarding Indoor Public Use, Retail Sales, and Prices of Electronic Cigarettes - U.S. States, Guam, Puerto Rico, and U.S. Virgin Islands, September 30, 2017. MMWR Morb Mortal Wkly Rep. 2017;66(49):1341-6.

9. Institute for Global Tobacco Control. Country Laws Regulating E-cigarettes: Policy Domains [Internet]. Baltimore (MD): Johns Hopkins Bloomberg School of Public Health; 2018 [cited 2018 Aug 31]. Available from: https://www.globaltobaccocontrol.org/e-cigarette/policy-domains

10. Yingst JM, Lester C, Veldheer S, et al. E-cigarette users commonly stealth vape in places where e-cigarette use is prohibited. Tob Control. 2018 Aug 10. doi: 10.1136/tobaccocontrol-2018-054432

11. Bartter T. Electronic cigarettes: Aggregate harm. Ann Intern Med. 2015;163(1):59-60.

12. Barrington-Trimis JL, Berhane K, Unger JB, et al. The E-cigarette social environment, e-cigarette use, and susceptibility to cigarette smoking. J Adolesc Health. 2016;59(1):75-80.

13. Stanwick R. E-cigarettes: Are we renormalizing public smoking? Reversing five decades of tobacco control and revitalizing nicotine dependency in children and youth in Canada. Paediatr Child Health. 2015;20(2): 101-5.



Supplementary File 1: Further details on methods.

Correspondence to: Dr George Thomson, University of Otago, Wellington, Box 7343, Wellington, New Zealand; e-mail: [email protected]


PUBLIC HEALTH ASSOCIATION OF AUSTRALIA

Working together for better health outcomes

Join today and be part of it! Public Health Association

A U S T R A L I A

www.phaa.net.au

PHAA MEMBER BENEFITS

BENEFITS OF INDIVIDUAL MEMBERSHIP

� Online access to the Australian and New Zealand Journal of Public Health, Australia’s premier public health publication

� The PHAA e-newsletter intouch and other electronic mailings and updates � The right to vote and hold office in PHAA � Opportunity to join up to 17 national Special Interest Groups (SIGs) (fees apply) � Access to State/Territory branch events and professional development opportunities � Reduction in fees to the PHAA annual conference and other various special interest

conferences � Access to PHAA forums and input into developing policies � Access to emailed list of public health job vacancies and opportunities � Networking and mentoring through access to senior public health professionals at branch

meetings, as well as through SIGs and at conferences and seminars � Eligibility to apply for various scholarships and awards � The ability to participate in, benefit from, or suggest and promote public health advocacy

programs

ADDITIONAL BENEFITS OF ORGANISATIONAL MEMBERSHIP (All the above benefits of individual membership apply to the nominated representative)

� Up to two staff members may attend PHAA Annual Conference and special interest conferences, workshops and seminars at the reduced member registration rate

� Discounted rates for advertising or for placing inserts in our current publications intouch and the Australian & New Zealand Journal of Public Health (does not apply to job vacancies and event promotional e-campaigns)

Australian and New Zealand Journal of Public HealthNotes for contributors

Copies of the full Notes, including referencing of electronic sources, and of the ‘Uniform requirements’ are available from the Australian and New Zealand Journal of Public Health website at

http://onlinelibrary.wiley.com/page/journal/17536405/homepage/ForAuthors.html

Submission procedure Manuscripts are subject to assessment by independent referees and editorial

revision. When a submitted manuscript is received, it is given a reference number and an acknowledgment is sent to the corresponding author. The appropriateness of the topic area and treatment is then considered by the Editors, sometimes with advice from members of the Editorial Board. If appropriate, the manuscript is sent for review. A decision, including referees’ comments, is usually sent to the authors within two or three months. The paper may be accepted or rejected, or authors may be invited to revise the paper and resubmit it. In the latter case, the Journal does not undertake to publish the paper as pressure of space may make acceptance and publication impossible.

All manuscripts are to be submitted online via our online submission service at: http://mc.manuscriptcentral.com/anzjph.

All manuscripts accepted for publication will incur a fee of $600 if the first author is a member of the Public Health Association of Australia or Public Health Association of New Zealand, and $1,800 if the first author is not a member of either organisation. The PHAA will invoice authors when the paper is accepted; the invoice must be paid prior to the accepted manuscript being processed and made available online. Invited Editorials and Book Reviews will not be subject to fees.

Papers will be made Open Access once they are published in an issue. Papers published in Early View, prior to publication in an issue, will only be available to PHAA members.

LengthBrief, pithy articles are preferred to long, wordy ones. Substantive findings

from a study are preferred to multiple minor papers. Pilot studies will only be considered if they are novel or add substantially to the existing literature.

Research Articles should be no longer than 5,000 words, including title, abstract, all references. This word count allows for three figures or tables (each equivalent to 500 words [half a Journal page]). Articles that do not include figures or tables have a word limit of 6,500 words. A structured abstract must be included.

Brief Reports should be no longer than 1,500 words and can include one table or figure. References must be included in the overall word count. A structured abstract must be included.

A Commentary should be no more than 2,500 words, including any references. A maximum of one table or figure may be included.

Editorials are limited to 1,500 words, including title and references. A maximum of one table or figure may be included.

Letters to the Editor are limited to 1,000 words, including references. A maximum of one table or figure may be included.

Abstract: The Abstract has a word limit of 200 words and must be structured using the headings Objective, Methods, Results, Conclusions, Implications for public health. An alternative format may be considered, if authors provide good justification. Please make sure you also address the implications for public health, including public health references, within the body of the manuscript.

Presenting your manuscript Authors intending to submit a paper should read the general guidelines of

the International Committee of Medical Journal Editors (ICMJE) ‘Uniform Requirements for Manuscripts Submitted to Biomedical Journals: Writing and Editing for Biomedical Publication’, available on http://www.icmje.org.

ANZJPH does not consider material that has been published at any length elsewhere or that is under consideration for another publication. It is prohibited by copyright law and can result in distortion of systematic

evidence. Particular attention should be paid to simultaneous or previous media reports or reports to funding bodies. Material that is readily accessible by internet search is considered to be a publication.

All authors must be prepared to take full responsibility for the content of a paper. All authors must certify that they satisfy at least the following three requirements: they contributed substantially to the conduct of the study, to both drafting and revision of the paper, and that they approve the final version. We may require each author to sign a statement describing their individual contributions, which will be included in the published paper.

We limit the number of authors for a paper to six in most cases. Justification for a larger number can be provided at submission. Additional contributions short of authorship can be addressed briefly under Acknowledgements. The maximum number of authors for a Letter to the Editor is three.

A separate title page must be uploaded with the manuscript files, listing all authors with a maximum of two affiliations each. Give full contact details for the corresponding author, including postal and e-mail addresses, and telephone number. Provide e-mail addresses for all other authors. Acknowledgements should be listed on the title page.

Tables must be submitted in Word format. Graphics should be in a suitable format (e.g. EPS, 300dpi JPG, or print resolution PDF) to provide clear, high resolution images in print.

ReferencesThe reference system used by the Journal is that recommended by

International Committee of Medical Journal Editors (ICMJE) and commonly known as ‘Vancouver style’. References are numbered in the text in the order in which they are first cited, and listed after the text of the manuscript in that order. An individual reference carries the same number each time it is cited and therefore appears in the list of references just once; ‘ibid’ and ‘op cit’ are not used. Please do not use automatic numbering of references. Examples of how to cite sources of various kinds can be found at http://authorservices.wiley.com/bauthor.

CopyrightAll articles published by ANZJPH are fully open access: freely available

to read, download and share, including to those who do not subscribe to the journal. The author, the author's funding agency, or the author's institution pays a fee to ensure that the article is made open access. Authors retain copyright in their articles, and are permitted to post the final, published PDF of their article on their personal website, and in an institutional repository or other free public server immediately after publication.

Authors have the option of publishing under the following Creative Commons License terms:

• Creative Commons Attribution License (CC BY)• Creative Commons Attribution Non-Commercial-NoDerivs License

(CC BY NC ND).To preview the terms and conditions of these open access agreements, please

visit the Copyright FAQs hosted on Wiley Author Services and visit http://www.wileyopenaccess.com/details/content/12f25db4c87/Copyright--License.html.

Disclaimer The Publisher, Association and Editors cannot be held responsible for errors or any consequences arising from the use of information contained in this journal; the views and opinions expressed do not necessarily reflect those of the Publisher, Association and Editors, neither does the publication of advertisements constitute any endorsement by the Publisher, Association and Editors of the products advertised.

Public Health Association A U S T R A L I A

The Journal of the Public Health Association of Australia Inc.

Editorial

305 The Public Health Association of Australia’s advocacy to prevent suicide Samantha Battams, Fiona Robards

Commentary

307 Listen, understand, collaborate: developing innovative strategies to improve health service utilisation by Aboriginal and Torres Strait Islander men Kootsy Canuto, Stephen Harfield, Gary Wittert, Alex Brown

310 The important role of charity in the welfare system for those who are food insecure Fiona H. McKay, Rebecca Lindberg

Indigenous Health

313 Feasibility and acceptability of opportunistic screening to detect atrial fibrillation in Aboriginal adults Rona Macniven, Josephine Gwynn, Hiroko Fujimoto, Sandy Hamilton, Sandra C. Thompson, Kerry Taylor, Monica Lawrence, Heather Finlayson, Graham Bolton, Norman Dulvari, Daryl C. Wright, Boe Rambaldini, Ben Freedman, Kylie Gwynne

319 Anaemia in early childhood among Aboriginal and Torres Strait Islander children of Far North Queensland: a retrospective cohort study Dympna Leonard, Petra Buttner, Fintan Thompson, Maria Makrides, Robyn McDermott

328 Participant profile and impacts of an Aboriginal healthy lifestyle and weight loss challenge over four years 2012-2015 Anne C. Grunseit, Erika Bohn-Goldbaum, Melanie Crane, Andrew Milat, Aaron Cashmore, Rose Fonua, Angela Gow, Rachael Havrlant, Kate Reid, Kiel Hennessey, Willow Firth, Adrian Bauman

334 Breast screening attendance of Aboriginal and Torres Strait Islander women in the Northern Territory of Australia Kriscia A. Tapia, Gail Garvey, Mark F. McEntee, Mary Rickard, Lorraine Lydiard, Patrick C. Brennan

340 Limited progress in closing the mortality gap for Aboriginal and Torres Strait Islander Australians of the Northern Territory Tom Wilson, Yuejen Zhao, John Condon

Food and Beverage

346 The frequency and magnitude of price-promoted beverages available for sale in Australian supermarkets Christina Zorbas, Beth Gilham, Tara Boelsen-Robinson, Miranda R.C. Blake, Anna Peeters, Adrian J. Cameron, Jason H.Y. Wu, Kathryn Backholer

352 Development of Australia’s front-of-pack interpretative nutrition labelling Health Star Rating system: lessons for public health advocates Michael Moore, Alexandra Jones, Christina M. Pollard, Heather Yeatman

355 The performance and potential of the Australasian Health Star Rating system: a four-year review using the RE-AIM framework Alexandra Jones, Anne Marie Thow, Cliona Ni Mhurchu, Gary Sacks, Bruce Neal

Sport

366 Unhealthy sport sponsorship at the 2017 AFL Grand Final: a case study of its frequency, duration and nature Tegan Nuss, Maree Scully, Melanie Wakefield, Helen Dixon

373 Challenges for sport organisations developing and delivering non-traditional social sport products for insufficiently active populations Kiera Staley, Alex Donaldson, Erica Randle, Matthew Nicholson, Paul O’Halloran, Rayoni Nelson, Matthew Cameron

General Public Health

382 Epidemiology of hospitalised traumatic brain injury in the state of New South Wales, Australia: a population-based study Ilaria Pozzato, Robyn L Tate, Ulrike Rosenkoetter, Ian D Cameron

389 New and old hotspots for rickettsial spotted fever acquired in Tasmania, 2012–2017 Gabriela Willis, Kerryn Lodo, Alistair McGregor, Faline Howes, Stephanie Williams, Mark Veitch

395 Public health and natural hazards: new policies and preparedness initiatives developed from an Australian bushfire case study Rachel Westcott, Kevin Ronan, Hilary Bambrick, Melanie Taylor

Letter

401 Observed vaping and smoking in outdoor public places: piloting objective data collection for policies on outdoor vaping George Thomson, Johanna Nee-Nee, Kirsty Sutherland, Rebecca Holland, Miriam Wilson, Nick Wilson

Australian and New Zealand Journal of Public Health 2… · Professor John Lynch School of Public...

Documents

Transcript of Australian and New Zealand Journal of Public Health 2… · Professor John Lynch School of Public...