Neighbourhood deprivation and smoking outcomes in South Africa

Lisa Lau, MPH. Department of Epidemiology, School of Public Health, University of Michigan.!Jamie Tam, MPH. Department of Health Management and Policy, School of Public Health, University of Michigan.!!RESEARCH CONFERENCE ON TOBACCO CONTROL IN AFRICA, 16 JULY 2014CAPE TOWN, SOUTH AFRICA

Background and Justification

• Individual-level predictors for adult smoking behaviours in South Africa have been established through previous research (Strebel et al., 1989; Peer et al., 2009 Vellios & van Walbeek, 2013)

• Smoking disparities exist across the country:

• 12% males and 6.2% of females are cigarette smokers (Tobacco Atlas 4th edition, 2009)

• Among men: BLACK - 32.8%, COLOURED - 52.1%, WHITE - 35.7%, INDIAN - 55.5% (DHS, 2003)

• Among women: BLACK - 5.2%, COLOURED - 41.8%, WHITE - 27.3%, INDIANS - 13.1% (DHS, 2003)

Background and Justification

• Understanding the role of neighbourhood influences on smoking-related outcomes is important for addressing these disparities

• The effectiveness of tobacco control strategies depends on the socioeconomic and environmental context in which they are implemented

• Economic and social deprivation has been shown to be significantly associated with smoking status in other contexts (Kleinschmidt et al., 1995; Shohaimi et al., 2003; Baumann et al., 2007; Blakely et al., 2013)

• This study is the first multi-level analysis of individual and neighbourhood socioeconomic contexts for smoking behaviours in South Africa

Neighbourhoods and Health

Neighbourhoods and Health

Residential segregation by race and SES

position ↑↓

Inequalities in resource

distributions

Neighbourhoods and Health

Residential segregation by race and SES

position ↑↓

Inequalities in resource

distributions

Neighbourhood physical environments - Environmental exposures - Food & recreational resources - Built environment - Aesthetic quality/natural spaces - Services - Quality of housing

Neighbourhoods and Health

Residential segregation by race and SES

position ↑↓

Inequalities in resource

distributions

Neighbourhood physical environments - Environmental exposures - Food & recreational resources - Built environment - Aesthetic quality/natural spaces - Services - Quality of housing

Neighbourhood social environments - Safety/violence - Social connections/cohesion - Local institutions - Norms

Neighbourhoods and Health

Residential segregation by race and SES

position ↑↓

Inequalities in resource

distributions

Neighbourhood physical environments - Environmental exposures - Food & recreational resources - Built environment - Aesthetic quality/natural spaces - Services - Quality of housing

Neighbourhood social environments - Safety/violence - Social connections/cohesion - Local institutions - Norms

Neighbourhoods and Health

Residential segregation by race and SES

position ↑↓

Inequalities in resource

distributions

Neighbourhood physical environments - Environmental exposures - Food & recreational resources - Built environment - Aesthetic quality/natural spaces - Services - Quality of housing

Neighbourhood social environments - Safety/violence - Social connections/cohesion - Local institutions - Norms

Behavioural mediators ↑↓

Stress

Neighbourhoods and Health

Residential segregation by race and SES

position ↑↓

Inequalities in resource

distributions

Neighbourhood physical environments - Environmental exposures - Food & recreational resources - Built environment - Aesthetic quality/natural spaces - Services - Quality of housing

Neighbourhood social environments - Safety/violence - Social connections/cohesion - Local institutions - Norms

Adapted from Diez Roux & Mair (2010)

HealthBehavioural mediators ↑↓

Stress

Neighbourhoods and Health

Residential segregation by race and SES

position ↑↓

Inequalities in resource

distributions

Neighbourhood physical environments - Environmental exposures - Food & recreational resources - Built environment - Aesthetic quality/natural spaces - Services - Quality of housing

Neighbourhood social environments - Safety/violence - Social connections/cohesion - Local institutions - Norms

Adapted from Diez Roux & Mair (2010)

HealthBehavioural mediators ↑↓

Stress How might smoking outcomes vary by neighbourhood

environment?

Neighbourhoods and smoking

Residential segregation by race and SES

position ↑↓

Inequalities in resource

distributions

Policy implementation - Smoke-free policies - Availability of tobacco - Tobacco advertising - Youth access to

tobacco - Local enforcement !

Smoking norms Economic conditions

Smoking status

Behavioural mediators ↑↓

Stress

Neighbourhood deprivation

How does it associate with smoking outcomes?

Image by FreddieA, shared under Creative Commons 3.0

Neighbourhood deprivation

• Conditions of the physical and social environment in local communities that contribute to the experience of deprivation including:

• (1) income and material, (2) employment, (3) education, and (4) living environment deprivation

• These four dimensions of deprivation are used by the South African Index of Multiple Deprivation (SAIMD) at municipality level (Wright and Nobel, 2009)

• Subindices of each domain, and an overall multiple deprivation index, were constructed with the same items as SAIMD, using Principal Component Analysis

Deprivation and smoking?

Social and physical deprivation due to living circumstances

Deprivation and smoking?

Social and physical deprivation due to living circumstances

- Poor enforcement and implementation of tobacco control policies

- Lack of local institutions to support health

- Social norms that encourage smoking

- Life stressors

Deprivation and smoking?

Social and physical deprivation due to living circumstances

- Poor enforcement and implementation of tobacco control policies

- Lack of local institutions to support health

- Social norms that encourage smoking

- Life stressors

Worse smoking outcomes

Deprivation and smoking?

Social and physical deprivation due to living circumstances

- Poor enforcement and implementation of tobacco control policies

- Lack of local institutions to support health

- Social norms that encourage smoking

- Life stressors

Worse smoking outcomes

- Reduced availability of tobacco products

- Insufficient money to purchase tobacco

- Social norms that do not encourage smoking

Deprivation and smoking?

Social and physical deprivation due to living circumstances

- Poor enforcement and implementation of tobacco control policies

- Lack of local institutions to support health

- Social norms that encourage smoking

- Life stressors

Worse smoking outcomes

- Reduced availability of tobacco products

- Insufficient money to purchase tobacco

- Social norms that do not encourage smoking

Better smoking outcomes

Methods

• NIDS 2008 Data: Adult and Household

• Multilevel analysis (Oakes & Kaufman, 2006)

• A neighbourhood is defined as a census enumeration area

• Outcomes of interest:

• Smoking status

• Smoking intensity: cigarettes per day

Methods

• Covariates

• Individual level: age, age squared, gender, race, educational attainment, alcohol consumption

• Household level: household income quintiles, other smokers in household (yes/no), urban/rural

• Neighbourhood level: neighbourhood deprivation constructed from SAIMD Index (Wright and Nobel, 2009)

Preliminary results: smoking status

Model 1: individual-level covariates onlyOdds ratio 95% CI

Age 1.1642*** [1.1408;1.1881]

Age 0.9983*** [0.9980;0.9985]

Male 7.7015*** [6.7026;8.8492]

Race: Coloured 8.2898*** [1.1237;3.6109]

Indian/Asian 2.0144** [1.1237;3.6108]

White 4.5478*** [3.4129;6.0599]

Education: Primary school or less 3.1935*** [2.6407;3.8621]

Some high school 1.9847*** [1.6741;2.3528]

Employed 1.1315 [0.9833;1.3020]

Alcohol 6.7123*** [5.5512;8.1162]

p≤0.05* p≤0.01** p≤0.001***

Model 2: adding household-level covariates to Model 1

• Positive significant associations found with being a smoker and:

• Urban (OR = 1.21**)

• has other smoker(s) in household (OR = 15.92***)

• 2nd (OR = 1.27*), 3rd (OR = 1.45**) and 4th quintiles (OR = 1.48***) [ref: 5th quintile]

• Individual level associations stayed generally the samep≤0.05*; p≤0.01**; p≤0.001***

Models 3A - E: add neighbourhood deprivation to Model 2

Model 3A Model 3B Model 3C Model 3D Model 3EOdds ratio

Multiple deprivation 1.02 — — — —

Living environment deprivation

— 0.99 — — —

Education — — 1.85** — —

Employment — — — 0.61** —

Income and material — — — — 1.03

Models differ in the deprivation indexes used

Models 3A - E: add neighbourhood deprivation to Model 2

Model 3A Model 3B Model 3C Model 3D Model 3EOdds ratio

Multiple deprivation 1.02 — — — —

Living environment deprivation

— 0.99 — — —

Education — — 1.85** — —

Employment — — — 0.61** —

Income and material — — — — 1.03

Models differ in the deprivation indexes used

Models 3A - E: add neighbourhood deprivation to Model 2

Model 3A Model 3B Model 3C Model 3D Model 3EOdds ratio

Multiple deprivation 1.02 — — — —

Living environment deprivation

— 0.99 — — —

Education — — 1.85** — —

Employment — — — 0.61** —

Income and material — — — — 1.03

Models differ in the deprivation indexes used

Results stratified by race

WHITE

p-value

Multiple deprivation 0.978

Living environment 0.876

Education 0.509

Employment 0.880

Income and material 0.191

BLACK

p-value

Multiple deprivation 0.231

Living environment 0.561

Education 0.001***

Employment 0.122

Income and material 0.536

COLOURED

p-value

Multiple deprivation 0.069

Living environment 0.276

Education 0.403

Employment 0.637

Income and material 0.006**

Preliminary results: smoking intensity

Results: smoking intensity

• Preliminary results indicate that none of the deprivation indices were significantly associated with smoking intensity

• Individual factors: age, male, being coloured, white, or Indian/Asian (ref: Black), regular alcohol consumption are significantly, positively associated with smoking intensity

• Household factors: Being in households of poorer quintiles is associated with lower smoking intensity

Conclusions

• Some types of deprivation matter more than others

• Proportion of adults who have had no secondary education matters more than the quality of living environment

• Employment deprivation is negatively associated with being a smoker

Policy implications?

• Continuous surveillance of tobacco use should be a priority

• Addressing disparities in smoking behaviours could help address disparities in burden of ill-health (see Ataguba et al., 2011)

• Characterising adult smokers could improve targeting of smoking cessation and tobacco interventions on local level

Limitations• Although stratified analysis (i.e. by race) is straightforward

with individual level (Model 1), it is harder to interpret (and probably needs to be controlled for) in multilevel analysis

• Difficult to draw conclusions for some subgroups due to small sample sizes (female smokers, Asian/Indian)

• Smoking status and smoking intensity are both self-reported; could be subject to recall bias and social desirability bias.

• Analyses do not account for sample weights

Next Steps and Future Directions

• Policies are not affecting all communities the same way

• future research might investigate how deprivation might prevent successful policy implementation

• Consider longitudinal evidence for the relationship between neighbourhood deprivation and smoking status and intensity

ReferencesAtaguba, J.E., Akazili, J. & McIntyre, D. (2011). Socioeconomic-related health inequality in South Africa: evidence from General Household Surveys. International Journal for Equity in Health 10:48. doi: 10.1186/1475-9276-10-48 !Baumann, M., Spitz, E., Guillemin, F., Ravaud, F., Choquet, M., Falissard, B. Chau, N. & Lorhandicap group. (2007). Associations of social and material deprivation with tobacco, alcohol, and psychotropic drug use, and gender: a population-based study. International Journal of Health Geographics 6:50-61. !Blakely, T., van der Deen, F.S., Woodward, A., Kawachi, I. & Carter, K. (2013). Do changes in income, deprivation, labour force status influence smoking behaviour over the short run? Panel study of 15000 adults. Tobacco Control 0:1-8. doi:10.1136/tobaccocontrol-2012-050944 !Diez Roux, A. V. & Mair, C. (2010). Neighborhoods and health. Annals of the New York Academy of Sciences, 1186(1):125-145. !Kleinschmidt, I., Hills, M. & Elliott, P. (1995). Smoking behaviour can be predicted by neighbourhood deprivation measures. Journal of Epidemiology and Community Health 49(Suppl 2):S72-S77. !Oakes, J.M. & Kaufman, J.S. (2006). Methods in Social Epidemiology. San Francisco: Jossey-Bass. !Peer, N., Bradshaw, D., Laubscher, R. & Steyn, K. (2009). Trends in adult tobacco use from two South African demographic and health surveys conducted in 1998 and 2003. South African Medical Journal 99(10):744-749. !Shohaimi, S., Luben, R., Wareham, N., Day, N., Bingham, S., Welch, A., Oakes, S. & Khaw k-T. (2003). Residential area deprivation predicts smoking habit independently of individual educational level and occupational social class. A cross sectional study in the Norfolk cohort of the European Investigation into Cancer (EPIC-Norfolk). Journal of Epidemiology and Community Health 57:270-276, !Strebel, P., Kuhn, L. & Yach, D. (1989). Determinants of cigarette smoking in the black township population of Cape Town. Journal of Epidemiology and Community Health, 43:209-213. doi:10.1136/jech.43.3.209 !Wright, G. & Nobel, M. (2009). The South African Index of Multiple Deprivation 2007 at municipal level. Pretoria: Department of Social Development. !Vellios, N. & van Walbeeck, C. (2013). Determinants of smoking initiation in South Africa. Masters dissertation. University of Cape Town.