Community-Based Intervention Packages for Improving...

CPiJa

Efa

I

A

4

ommunity-Based Interventionackages for Improving Perinatal Health

n Developing Countries: A Review of the Evidenceessica Schiffman, BA, Gary L. Darmstadt, MD, MS, Siddharth Agarwal,nd Abdullah H. Baqui, MBBS, DrPH, MPH

The Lancet Neonatal Survival Series categorized neonatal health interventions into 3 servicedelivery modes: “Outreach,” “Family-Community Care,” and “Facility-based Clinical Care.”Family-Community Care services generally have a greater potential impact on neonatal healththan Outreach services, with similar costs. Combining interventions from all 3 service deliverymodes is ideal for achievement of high impact. However, access to clinical care is limited inresource-poor settings with weak health systems. The current trend for those settings is tocombine neonatal interventions into community-based intervention packages (CBIPs), whichcan be integrated into the local health care system. In this article, we searched several largedatabases to identify all published, large-scale, controlled studies that were implemented in arural setting, included a control group, and reported neonatal and/or perinatal mortality asoutcomes. We identified only 9 large-scale studies that fit these criteria. Several conclusionscan be reached. (1) Family-Community Care interventions can have a substantial effect onneonatal and perinatal mortality. (2) Several important common strategies were used across thestudies, including community mobilization, health education, behavior change communicationsessions, care seeking modalities, and home visits during pregnancy and after birth. However,implementation of these interventions varied widely across the studies. (3) There is a need foradditional, large-scale studies to test evidence-based CBIPs in developing countries, particu-larly in Africa, where no large-scale studies were identified. (4) We need to establish consistent,clearly defined terminology and protocols for designing trials and reporting outcomes so thatwe are able to compare results across different settings. (5) There is an urgent need to investin research and program development focusing on neonatal health in urban areas. (6) It iscrucial to integrate CBIPs in rural and urban settings into the already existing health caresystem to facilitate sustainability of the program and for scaling up. It is also important toevaluate the packages and to demonstrate the health impact of large-scale implementation. (7)Finally, there is a need for improving the continuum of care between home and facility-basedcare.Semin Perinatol 34:462-476 © 2010 Published by Elsevier Inc.

KEYWORDS community-based, intervention packages, neonatal health, neonatal mortality, urbanslums

it

at(ashm

very year, an estimated 3.6 million infants die world-wide in the first month of life.1,2 Approximately three-

ourths of these deaths occur within the first week of life,nd the majority are preventable through simple, low-cost

nternational Center for Advancing Neonatal Health, Department of InternationalHealth, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD.

ddress reprint requests to Abdullah H. Baqui, MBBS, DrPH, MPH, Inter-national Center for Advancing Neonatal Health, Department of Interna-tional Health, Johns Hopkins Bloomberg School of Public Health, Balti-

nmore, MD 21205. E-mail: [email protected]

62 0146-0005/10/$-see front matter © 2010 Published by Elsevier Inc.doi:10.1053/j.semperi.2010.09.008

nterventions.3 One-half or more of these deaths occur inhe home.2

The most common causes of neonatal mortality includevariety of infections (36%), prematurity (28%), and in-

rapartum-related neonatal deaths (“birth asphyxia”)23%).2 Maternal health during pregnancy and deliveryre key determinants of neonatal survival. In low-resourceettings, where deliveries are commonly performed in theome, household practices are important determinants ofortality.3 Therefore, to prevent the greatest number of

eonatal deaths, we must focus our attention on prevent-

mailto:[email protected]

id

gtvhbnvmvtgtca6

r

tcdcigCocfcdanrcnba

Community-based interventions 463

ng complications in the perinatal and neonatal periods ineveloping countries.The current trend in developing community-based strate-

ies for improving neonatal health in low-resource settings iso combine antepartum, intrapartum, and postnatal inter-entions into “packages” that can be integrated into theealth care system. In this paper, we define a “package” as aundle of 2 or more interventions and refer to the commu-ity-based intervention packages as community-based inter-ention package (CBIP). We are moving away from imple-enting single interventions into health care systems in a

ertical manner because combining evidence-based interven-ions into packages with a common service delivery mode isenerally more cost-effective than implementing a single in-ervention into the health care system. Even in weak healthare systems, reaching high coverage with these basic pack-ges can decrease neonatal deaths by an estimated 35% to6%.3

The authors of the Lancet Neonatal Survival Series catego-

Figure 1 The Lancet Neonatal Survival Series Universaldelivery modes.4

ized the service delivery modes for neonatal health interven- d

ions into “Outreach,” “Family-Community Care,” and “Fa-ility-based Clinical Care” (Fig. 1).3 The “Outreach” categoryescribes services that meet the needs of a population andan be implemented periodically either in health facilities orn the community. Examples include immunization pro-rams and routine antenatal care.3,4 The “Family-Communityare” category combines family-oriented and community-riented services that are designed to complement the spe-ific community’s social and cultural context.3 An importantocus of Family-Community Care is to empower families andommunities to practice healthy preventive behaviors and toemand services, such as care-seeking for illness. Another ex-mple of Family-Community Care is case management of neo-atal illness by community-based workers (CBWs), includingeferral to health facilities for clinical care.3,5 The “Clinical Care”ategory describes facility-based clinical services, such as neo-atal resuscitation, emergency obstetrical care, care of sick new-orns, and care for very low birth weight infants. These servicesre performed by skilled clinicians and are most commonly

nce-based intervention packages organized by service
Evide
elivered in clinics and hospitals.3

pwtficotw3ctec

ftsftCasuma

MLWRsrg(vtwpppi

wnmtWvAtstna

taRpsino

RWscrec

DWt(2eabd

CFsvake

stNlCcisCBtpmpgF2mb

464 J. Schiffman et al

Family-Community Care services generally have a greaterotential impact on neonatal health than Outreach services,ith similar costs.3 In a review article, Haws et al4 assessed

he impact of intervention packages on neonatal health andound that studies that included Family-Community Carenterventions were more likely to provide a continuum ofare from pregnancy through the postnatal period than thether service delivery modes. CBIPs delivered through bothhe Outreach and Family-Community Care service modesere estimated to decrease neonatal mortality by 18% to7%.3 Although Clinical Care services are generally moreostly than the other service delivery modes, they are integralo substantially decreasing perinatal mortality beyond theffect that Outreach and Family-Community Care servicesan have alone.3

On the basis of the principles for service delivery setorth in the Lancet Neonatal Survival Series,3 the objec-ives of this paper are (1) to systematically identify large-cale, controlled studies that test a CBIP with a primaryocus on Family-Community Care interventions becausehey have a greater potential impact than Outreach andlinical Care interventions, (2) to describe the positivend negative findings of the studies identified, and (3) toummarize the lessons learned. We also briefly describerban models of health care as a comparison to the com-unity-based models that have been implemented in rural

reas.

ethodsiterature Search and Selection of Trialse searched PubMed, the Cochrane Database of Systematic

eviews, and Google Scholar to identify all published, large-cale controlled trials or program evaluations carried out in aural setting that implemented a CBIP and included a controlroup. Only studies that reported neonatal mortality rateNMR) and/or perinatal mortality rate (PMR) as outcomeariables were considered. We did not review trials thatested single interventions. We also did not consider trialsith a primary focus on clinical care for neonatal infections,rematurity, intrapartum-related neonatal deaths (“birth as-hyxia”), stillbirths, or hypothermia unless they were incor-orated into a CBIP; these topics are covered in other articles

n this series.6-9

Combinations of the following key terms (among others)ere used to search the databases: “neonatal mortality,” “neo-atal health,” “perinatal mortality,” “perinatal health,” “com-unity based intervention,” and “birth outcomes.” We ini-

ially received 790 hits when performing a broad search.e narrowed our search to 108 citations that tested inter-

entions for improving neonatal or perinatal health.mong these, 61 were studies that were not controlled

rials or program evaluations that tested CBIPs (7 of thesetudies were controlled studies that tested single interven-ions), 8 were review articles or meta-analyses, 28 wereot journal articles (books, presentations, reports, etc.),
nd 11 were controlled trials or program evaluations that l
ested CBIPs (one was an RCT but the trial was ongoingnd the results are not yet published, another one was anCT that did not report NMR or PMR). Although ourrimary focus was on identifying RCTs, we extended ourearch to include controlled trials that were not random-zed and program evaluations that used a randomized oronrandomized controlled trial design due to the scarcityf RCTs in the literature.

esultse identified 10 large-scale, controlled studies that fit our

earch criteria.3,10,11-14,16-19 Because Bang et al 200517 was aontinuation of the Bang et al 199914 study and the resultseported in the 2005 paper were an updated version, wexcluded Bang et al 1999 from our final list of large-scaleontrolled trials of CBIPs (Table 1).

esign of the Studiese identified 9 studies published to date in which the au-

hors evaluated CBIPs in a rural, developing country settingTable 1). Five of the 9 studies were cluster RCTs,10,11,12,15,16

were nonrandomized controlled trials,17,18 one programvaluation had a controlled, quasi-experimental design,19

nd the final trial had a 2-part design; the first part used aefore-after design and the second part used a cluster RCTesign (Table 1).13

omparing CBIPs by Service Delivery Modeigure 1 groups the evidence-based interventions into theervice delivery modes specified by the Lancet Neonatal Sur-ival Series: Family-Community Care, Outreach Services,nd Facility-based Clinical Care.3,4 Table 2 compares the 9ey studies across the Family-Community Care service deliv-ry mode.

Only one study offered all the Family-Community Caretrategies.17 This study also demonstrated the greatest reduc-ion in NMR compared with the other studies that reportedMR as an outcome (Table 3). Baqui et al 200811 (Bang-

adesh) came close to offering all the Family-Communityare strategies but did not include extra community-basedare of low birth weight neonates. The remainder of the stud-es included some, but not all the Family-Community Caretrategies. Few studies offered any of the Outreach or Clinicalare interventions listed in (Fig. 1) (not shown in Table 2).aqui et al 2008 (India)19 and Darmstadt et al 201015 offeredetanus toxoid immunization, an Outreach intervention, asart of the CBIP. Only 411,15,18,19 of the studies included pro-otion or distribution of folic acid supplementation duringregnancy, however the supplements were not necessarilyiven periconceptually. None of the studies included anyacility-based Clinical Care interventions, except Baqui et al00811 (Bangladesh) provided limited support of govern-ent facilities to provide care for mothers and their new-

orns.Some additional strategies not included in the Lancet’s

ist of universal evidence-based packages were also in-

Table 1 Comparison of 9 Key Large-Scale Controlled Studies in Which the Authors Evaluated CBIPs in Rural Resource-Poor Settings

Authors (Year),Location Study Design Intervention Comparison Group

Positive Findings(Significant P <0.05), *NMR per

1000 Births

NegativeFindings orWeaknesses Conclusions

Lessons Learned/ UniqueAspects

Bang (2005),Gadchiroli, India(a continuation ofthe Bang 1999trial)

Controlled trial, notrandomized

Continued interventions fromprevious trial (Bang,1999). The field trial ofhome-based neonatal carewas started in 1993.

1995-1996: home visitsconsisted only ofobservations of neonatalmorbidity, cases of deathand treatment of neonatalpneumonia.

1996-1997: management ofsick neonates wasintroduced as anintervention.

1997-1998: management ofsick neonates wasimproved and healtheducation was added.

Received usual care providedby local health system.

NMR decreased from 62(1993) to 25 (2003),which is a 70%decrease (95% CI 59-81%); Early NMRdecreased by 64%,late NMR by 80%,SBR by 49%, PMR by56%, and IMR by57%. PNMR did notchange. Allreductions weresignificant at P <0.001 except SBR atP < 0.05.

HBNC had no effecton PNMR.

Control groups notrandomlyselected becauseof a lack offeasibility.

The estimatednumber of deathsattributed to thevariouscomponents ofHBNC was basedon a nestedbefore-aftercomparison in theintervention arm(no control).

HBNC decreased the NMRby 70% and contributedequally to early and lateNMR. SBR and PMRalso decreased byapproximately 50%.Most deaths (93.6%)were prevented due tosickness management.This reduction inneonatal mortality canbe explained by thefollowing components ofHBNC: sepsismanagement (36%),supportive care of LBWneonates (34%),asphyxia management(19%), primaryprevention (7%), andother (4%).

The only study that included allof the Family-CommunityCare components listed inTable 2. This study resulted inthe greatest reduction inNMR, among the studies thatreported NMR as anoutcome. Most of thereduction in mortality wasattributed to home-basedmanagement of sepsis, LBWneonates, and birth asphyxia.

Baqui (2008), Sylhet,Bangladesh

Cluster randomizedcontrolled trial

2 interventions over 30 mo:1. Home-care arm: female

community health workersidentified pregnantwomen, made home visitsduring pregnancy toeducate mothers andassess newborns on days1, 3, 7 after birth, referredsick neonates to hospitalsand provided treatment inhomes.

2. Community-care arm-education and careseeking through groupsessions held by femaleand male communitymobilizers, no home visitsincluded.

Received usual health servicesby government, NGOs, andprivate providers withadditional governmenthealth-systemstrengthening, includingrefresher training forfacility-level healthproviders to treat neonatalinfections and providedsupply of antibiotics.

Last 6 mo of 30-moperiod:

NMR 29.2 (home-carearm), 45.2(community-carearm), 43.5(comparison arm);NMR reduced by 34%in home-care arm vscomparison arm(adjusted RR 0.66,95% CI 0.47-0.93).No reduction in NMRin community-carearm vs comparisonarm (adjusted RR0.95, 95% CI 0.69-1.31).

Community Healthworkers attended<5% of birthsbecause of a highworkload, traveldistances, andnot receivingtimely notificationof births. Eachfemalecommunity healthworkerresponsible forlarge population(18,000).

A home-care strategy thatintegrates preventativeand curative newbornservices reduced NMRby 34% and is aneffective intervention toreduce neonatalmortality in communitieswith weak health caresystems and highneonatal mortality.

Showed that home-care is moresuccessful at decreasingneonatal mortality than apurely community-basedmodel.

Comm

unity-basedinterventions

465

Table 1 Continued



1000 Births



Baqui (2008), UttarPradesh, India*

Controlled, quasi-experimentaldesign

Intervention includedprenatal, intrapartum, andpostnatal components.

Prenatal: home visits weremade by auxiliary-nursemidwives, aganwadiworkers (maternal andchild health promotionworkers), and changeagents to providecounseling on preventivecare, nutrition,preparedness for childbirth, and health care usefor complications;

Intrapartum: families wereencouraged to callauxiliary nurse-midwivesor trained TBAs to attenddelivery;

Postnatal: CBWs visitedfamilies at home afterbirth to provide counselingon breastfeeding,essential newborn care(thermal care, hygiene,and clean cord care),maternal and newborndanger signs, and health-care use; follow-up visitswere made for sick,premature, or LBWnewborns.

Received standard governmenthealth and integrated ChildDevelopment services.

The NMR decreased by34% within the groupof women whoreceived a postnatalvisit within 28 d(NMR 35.7, 95% CI29.2-42.1) comparedwith the group whoeither received novisit (NMR 53.8, 95%CI 48.9-58.8) or onlyan antenatal visit(NMR 54, 95% CI45.9-62.1); womenwho received apostnatal visit within3 d of delivery had a25% reduction inNMR compared withthose who did notreceive a visit.

Adjusted baselineNMRs did notdiffer betweenintervention(NMR 46.4, 95%CI 42-50.8) andcomparisongroups (NMR45.8, 95% CI40.6-51).Neonates whoonly received avisit by a CBWduring theantenatal perioddid not havelower NMR (54,95% 45.9-62.1)compared withthose who did notreceive a visit(53.8, 95% CI48.9-58.8).

INHP with a newborncomponent increasedcoverage of antenataland postnatal healthpromotion visits andimproved someneonatal-care practices.Most of theimprovement was in theantenatal period. Fewerthan 1 in 4 neonatesreceived a home visitduring the first 3 d oflife, which is when halfof all neonatal deathsoccurred.

The program failed toreduce neonatalmortality but thosewomen who received apostnatal visit within thefirst 28 d after birthwere less likely to havetheir newborn die. Mostof the reduction in NMRwas most likely due topromotion of essentialnewborn care practices,including exclusivebreastfeeding, thermalcare, clean cord care,and referral of sicknewborns.

Simulation analysesshowed that 90%program coverage withan antenatal visit and apostnatal home visitwithin the first 28 d canreduce NMR by 34%from baseline.

The first study to usemultipurpose governmenthealth workers as part of alarge-scale community-basedneonatal health programevaluation to assess theimpact on neonatal mortality.

466J.Schiffm

anet

al

Table 1 Continued



1000 Births



Bhutta (2008), Hala,Pakistan

Cluster controlledtrial

Over the course of 2 years:LHWs received additionaltraining in home-basednewborn care, visitedpregnant women twiceduring pregnancy, within24 h after birth, and days1,3,7,14, and 28 afterdelivery; LHWs wereencouraged to link up withDais (traditional birthattendants); Voluntarycommunity healthcommittees were createdto work with and supportLHWs—they organizedgroup education sessionswhere LHWs used videosand visual aids to promoteknowledge and behaviorabout maternal andnewborn care.

Received services fromgovernment’s alreadyexisting LHW program, noattempt made to link LHWswith TBAs, all rural healthcenter clinical staff wereoffered training in newborncare.

Reductions frombaseline NMR from57.3 to 41.3(decreased by 28%);reduction in SBR from65.9 to 43.1 per 1000births (decreased by34.6%).

LHWs not providedwith injectableantibiotics.

Because this was apilot study, therewere a smallnumber ofclusters.

Lack of statisticalanalysis due tolimitations of pilotstudy.

Low retention ratesof health-carestaff.

Decreases in NMR andSBR indicate communityhealth workers (LHWsand Dais) can beeffective atimplementing acommunity outreachintervention forimproving neonatalhealth. A larger RCT iscurrently under way.

Focused on strengtheningalready existing primary caremodel centered on LHWs.

Carlo (2010), 6countries

First part (EssentialNewborn CareCourse): before-after design

Second part(NeonatalResuscitationProgram):clusterrandomizedcontrolled trial

Used train-the-trainer modelwith local instructors whotrained birth attendantsfrom rural communitiesusing the WHO essentialnewborn care (ENC)course which focuses onroutine neonatal care,resuscitation,thermoregulation,breastfeeding, skin-to-skincare, care of the smallbaby, and commonillnesses. After completionof the essential newborncare course, the clustersrandomly assigned to theNeonatal ResuscitationProgram received a 3 dcourse.

Before-after design: thecontrol group was a cohortof women who did notreceive training in theessential newborn carecourse (before the coursewas taught).

Decrease in SBR (RR0.69, 95% CI 0.54-0.88, P � 0.003).

No significantreduction inneonatal deathfrom all causes infirst 7 d afterbirth (RR 0.99,95% CI 0.81-1.22) or in thePMR; nosignificantreduction in ratesof neonatal deathin births in whichattendants weretrained in theNeonatalResuscitationProgram.

The NMR in the 7 d afterbirth and PMR did notdecrease after essentialnewborn care training ofcommunity-based birthattendants, however theSBR was significantlydecreased. Subsequenttraining in the NeonatalResuscitation Programdid not significantlydecrease the NMR,PMR, or SBR.

Strengths include multicountrydesign, large sample sizes,use of local trainers to trainbirth attendants. The apparenteffect of ENC on SBR mayhave been due to amisclassification error, whichis a potential weakness of thestudy.

Comm


467

Table 1 Continued



1000 Births



Darmstadt (2010),Mirzapur,Bangladesh


CHWs identified pregnanciesthrough surveillance,made 2 antenatal homevisits, attended thedelivery when possible,made 3 postnatal visits ondays 2,5,8 to promotepreventative newborn carepractices, referred sickneonates to localhospitals, and cared forneonates in the home iffamilies refused to bereferred (no injectableantibiotics were used).

Usual health services providedby government, NGOs, andprivate providers.

No evidence that theintervention reducedNMR.

Adjusted mortalityhazard ratio inintervention armat baseline was1.02 (95% CI0.80, 1.30) vs atend point was0.87 (95% CI0.68, 1.12)–results were notsignificant.

The intervention did notadequately address therisk factors for mortalityin this population.

Baseline NMR was lower thanthat of other low-resourcecommunities and thereforethe causes of neonatalmortality were different. 60%of deaths from birth asphyxiaand prematurity—neonatalsepsis contributed to a lesserextent. CHWs lacked thetools and skills needed totreat asphyxia andprematurity, attended < 20%of home deliveries, and hadonly 18%-33% postnatal carecoverage. Therefore, the leveland cause-structure in thelocal population must beconsidered when developingthe intervention.

Kumar (2008),Shivgarh, India


Two intervention groups over16 mo:

1. Intervention group A:received preventativepackage of interventionsfor ENC, including birthpreparedness, cleandelivery and cord care,thermal care (skin-to-skin),breastfeeding promotion,and danger signrecognition.

2. Intervention group B:received the same ENCpackage as group A plususe of a liquid crystalhypothermia indicator(ThermoSpot).

In both intervention groups,CHWs delivered packagesthrough group meetingsand 2 antenatal and 2postnatal house visits.

Received usual governmentaland NGO services.

NMR was reduced by54% in interventiongroup A comparedwith control (NMR0.46, 95% CI 0.35-0.60 and 52% inintervention group Bcompared withcontrol (NMR 0.48,95% CI 0.35-0.66).

This community-basedintervention thatprimarily focused onbehavioral modificationwithin the socioculturalcontext led to a greaterthan 50% reduction inneonatal mortality. Inaddition, addingThermoSpot to theintervention did not leadto a significant reductionin neonatal mortality.

Intervention was developedbased on research thatactively involved participationof community membersthroughout the researchperiod.

The study shows that it ispossible to change behaviorthat produces positiveoutcomes and is scalable.

468J.Schiffm

anet

al

Table 1 Continued



1000 Births



Jokhio (2005), Sind,Pakistan


TBAs were trained and givendisposable delivery kits,referred women foremergency obstetricalcare, and visited womenat least 3 times duringpregnancy (3, 6, 9 mo);LHWs supported the TBAsand recorded data; Teamsof obstetricians heldoutreach clinics forantenatal care.

TBAs did not receive trainingand were not given deliverykits; LHWs enrolled andfollowed up with allpregnant women and madethe normal monthly homevisits; No outreach clinicswere organized.

Cluster adjusted OR forperinatal death0.70 (95% CI 0.59,0.82) and neonataldeath 0.71 (95% CI0.62, 0.83); PMR wasreduced by 30% andNMR by 29%.

Cluster adjusted ORfor maternalmortality0.74 (95% CI0.45, 1.23)-notstatisticallysignificant; Thestudy was notpowered todetect meaningfulchange inmaternalmortality.

Small number ofclusters of largesize is alimitation of thestudy design.

Training TBAs andintegrating them into analready existing primaryhealth care system waseffective in decreasingperinatal mortality.

The use of disposable, safe-delivery kits by TBAs was aunique component of thepackage.

Goal of improving the existingprimary health care system tomake it sustainable and atlow cost was also unique.

Manandhar (2004),Makwanpur, Nepal


Trained female facilitatorsfrom the local communitysupported groups throughan action-learning cycleusing participatorylearning rather thaninstruction; They identifiedlocal perinatal problemsand developed preventionstrategies.

The authors only wanted totest the effect of thewomen’s group interventionso they provided healthservices strengtheningactivities in bothintervention and controlgroups, includingimprovements in equipmentfor resuscitation andphototherapy, ENC trainingat all levels of thehealthcare system, andbasic newborn care kitsdistributed to CBWs.

NMR 26.2 in theintervention groupcompared with 36.9 incontrol group(adjusted OR 0.70,95% CI 0.53, 0.94)-reduced by 30%;Maternal mortalitywas also reduced(adjusted OR 0.22,95% CI 0.05, 0.90)-reduced by 78%.

Differences inliteracy andpoverty indicatorsbetweenintervention andcontrol groups(less poverty ininterventiongroup). SBRswere similar inintervention andcontrol groups.

An intervention in ruralNepal involvingwomen’s groups led bya local female facilitatorreduced neonatalmortality by 30%.

95% of groups remained activeat end of trial without anyfinancial incentives; This mayindicate that the interventionis scalable.

Interactive, participatorylearning strategies usingpicture card games, stretcherschemes, and interactivelearning during home visitswere unique.

CBW, community-based worker; CI, confidence interval; ENMR, early neonatal mortality rate; IMR, infant mortality rate; INHP, Integrated Nutrition and Health Programme; LHW, lady health worker; NGO, nongovernmentalorganization; HBNC, home-based neonatal care; LBW, low birth weight; NMR, neonatal mortality rate; PMR, perinatal mortality rate; PNMR, postneonatal mortality rate; SBR, stillbirth rate; TBA, traditional birth attendants;WHO, World Health Organization; ENC, Essential Newborn Care.

*The Baqui et al 2008 (India) study is not a trial but a program evaluation. The other 8 studies in this table are trials.

Comm


469

ccsptTb

CITP

hKNSTsdhb2

i

TS

F

A*

*

* t explic

T

BBBBCDKJM

*

†


luded in Table 2. All nine studies had training sessions forommunity health worker (CHWs), and in all but twotudies CHWs made antepartum, intrapartum, and/orostnatal home visits. Seven Clean delivery kits were dis-ributed for use in the intrapartum period in six studies.wo studies used injectable antibiotics as part of home-ased management of neonatal illnesses.

omparing Studies bympact on NMR and PMRable 3 compares the percentage reduction in NMR andMR across all studies. It was found that Bang et al 2005

able 2 Comparison of 9 Key Controlled Studies Using theervice Delivery Mode3,4

Evidence-based Service-delivery Strategies

Bang (1999& 2005)

Gadchiroli,India

Baqui (2008)Sylhet,

Bangladesh

Baqui (2008Uttar

Pradesh,India

amily community careCommunity mobilization &

engagementX X X

Behavior changecommunications*

X X X

Promotion and practice ofclean delivery

X X X

Referral of complicationsto health facility

X X X

Extra community-basedcare of LBW babies

X

Community-based casemanagement ofpneumonia

X X

Home visits byCommunity-BasedWorkers**

X X X

Additional strategies***Use of injectable

antibioticsX X

Distribution of cleandelivery kit

X X

bbreviation: LBW� low birth weight.Antenatal and postnatal domicillary behavior change communication

care, clean cord care, care seeking, and demand for quality clin*Community-Based Workers include Community Health Workers (C

Community Mobilizers and other community workers.**These items were added by the authors of this paper and are no

packages.3

able 3 Comparison of Percent Reduction in NMR and PMR

ang (2005), Gadchiroli, Indiaaqui (2008), Sylhet, Bangladeshaqui (2008), Uttar Pradesh, India*hutta (2008), Hala, Pakistanarlo (2010), six countriesarmstadt (2010), Mirzapur, Bangladeshumar (2008), Shivgarh, India†okhio (2005), Sind, Pakistananandhar (2004), Makwanpur, Nepal

Percent decrease in NMR is reported for neonates who received aonly an antenatal visit was received.

The first number represents the % decrease in the intervention grou
the % decrease in the intervention group that received essential newb
ad the greatest reduction in NMR (70%) and PMR (56%).umar et al 200810 had the second greatest decrease inMR (54% in the intervention group without Thermo-pot) and PMR (41% in the intervention group withouthermoSpot). Three studies did not have a statisticallyignificant decrease in NMR.15,13,19 Baqui et al 200819 (In-ia) did not show a significant decrease in NMR overall,owever, there was a 34% decrease in NMR among infantsorn to women who had received a postnatal visit within8 days after birth (Table 3).By using a before-after design to test the impact of train-

ng birth attendants in the World Health Organization

et’s Evidence-Based Intervention Packages Organized by

utta08)la,

istan

Carlo(2010)

Sixcountries

Darmstadt(2010)

Mirzapur,Bangladesh

Kumar(2008)

Shivgarh,India

Jokhio(2005)Sind,

Pakistan

Manandhar(2004)

Makwanpur,Nepal

X X X

X X X X X X

X X X X X

X X X X

X

X

X X X X

X X X X

gies promoting one of more of the following: breastfeeding, thermalre., Traditional Birth Attendants (TBAs), Lady Health Workers (LHWs),

itly included as strategies included in the Lancet Series universal

s 9 Key Controlled Studies

(% Reduction) PMR (% Reduction)

70 5634 Not reported34 Not reported28 34.6

ot significant Not significantot significant Not reported

54, 52 41, 3829% 30%30% Not reported

tal home visit within 28 d of birth. NMR remained unchanged when

eceived essential newborn care, and the second number represents

Lanc

) Bh(20Ha

Pak

strateical caHWs)

Acros

NMR

NN

postna

p that r
orn care and ThermoSpot.

(crcstpH(iSlevieboaa“oi

RI(fiLmtAvs

trlfpcoiinnma

t2aCt(t

Atti

CaSsacaaticgcarvc

hiacscmevd

Bwhmabeftmvass

aiactTm


WHO) Essential Newborn Care (ENC) course, which in-luded basic newborn resuscitation, Carlo et al 201013

eported no significant decrease in PMR or NMR from allauses in the first 7 days after birth. They also did not find aignificant reduction in NMR in births in which attendants wererained in the Neonatal Resuscitation Program (NRP), whichrovided more advanced training in neonatal resuscitation.owever, they did find a significant decrease in the stillbirth rate

SBR) after training in ENC (relative risk 0.69, 95% confidencenterval 0.54-0.88, P � 0.003). This observed reduction in theBR presents some methodological and interpretation chal-enges. As the authors note, baseline misclassification of somearly neonatal deaths as a “stillbirth” may explain their obser-ation of an apparent benefit of ENC on the SBR. It is antic-pated that neonatal resuscitation training might increasearly neonatal deaths if previously nonresuscitated “still-irths” are resuscitated but subsequently die, but this was notbserved. Thus, the reduction in stillbirth may be attribut-ble to improved recognition and intervention for asphyxi-ted infants who would have otherwise been consideredstillborn”; however, this remains unclear. The lack of impactf NRP training was possibly because resuscitation was alsoncluded in ENC training.

ole of CBWsn this article, we refer to community-based workersCBWs) as an umbrella term that includes a variety ofrontline workers with various levels of training and skills,ncluding CHWs, traditional birth attendants (TBAs),ady Health Workers (LHWs) in Pakistan, communityobilizers (who provide community mobilization activi-

ies but not home visits), and other community workers.ll 9 studies included CBWs that were trained to carry outarious tasks (Table 2); however, these tasks varied con-iderably across the studies.

CBWs often had multiple roles in the community and inhe home. They were commonly literate women who wereecruited from the community. Baqui et al 200811 (Bang-adesh) and Kumar et al 200810 specifically recruited bothemale and male CBWs because they believed that men alsolayed an important role as agents for improving newbornare. In some studies, CBW functions included identificationf pregnant women and data collection, making antepartum,ntrapartum, and/or postnatal home visits to promote behav-or change, providing home-based case management of neo-atal illnesses, and holding group meetings in the commu-ity to promote healthy behaviors. Some CBWs alsoanaged neonatal illnesses in the home by using injectable

ntibiotics.TBA training was also an important component of some of

he intervention packages. Training of TBAs in the Carlo et al01013 trial was the primary focus of the intervention pack-ge. TBAs first participated in the WHO Essential Newbornare course and some of them subsequently participated in

he Neonatal Rsuscitation Program. In Jokhio et al’s study2005),12 TBAs also made antenatal home visits and were
rained to conduct clean deliveries using clean delivery kits. p
lthough Jokhio et al found a statistically significant reduc-ion in perinatal mortality, other studies have also shown thathe benefits of training TBAs may not be compelling and thatnvestment in providing such training may not be justified.20

ommunity Mobilizationnd Promotion of Care Seekingix studies included community mobilization and all ninetudies included behavior change communication (BCC)ctivities in their intervention packages (Table 2). Theommunity mobilization and BCC activities varied widelycross the studies and are discussed in detail in Kumar etl,21 along with interpretation of the differential impact ofhe approaches taken. In brief, 4 out of 6 studies thatncluded community mobilization resulted in a statisti-ally significant decrease in NMR. Baqui et al 200811 (Ban-ladesh) did not find a significant decrease in NMR in theommunity-care arm compared with the control group,nd Baqui et al 2008 (India) also did not find a significanteduction in NMR in women who received an antenatalisit but NMR did significantly decrease when women re-eived a postnatal visit within 28 days of birth19 (Table 1).

Individual and group health education sessions wereeld in the Bang et al 200517 study with the goal of chang-

ng health behaviors and increasing care seeking. Mothersnd grandmothers were educated about topics involvingare of pregnant women and newborns. In the Bhutta et altudy,18 the LHWs were supported by community healthommittees comprised of volunteers from the local com-unities. The committees held group education sessions

very 3 months during which LHWs used flip charts andideos for health education. Adolescents, women of repro-uctive age, and elderly women attended the sessions.An integral component of the community-care arm in

aqui et al11 (Bangladesh) was community mobilization,hereas in the home-care arm, the main strategy wasome visits by CHWs. In both arms, male and femaleobilizers held group meetings to educate the community

bout birth preparedness and newborn care. Female mo-ilizers held visits every 4 months, whereas male mobiliz-rs held visits every 10 months. In the home-care arm,ewer female community mobilizers were included, andhey held visits every 8 months. In addition, in the com-unity-care arm, “community resource people” who were

olunteers from the villages, identified pregnant womennd encouraged them to attend the community meetings,eek antenatal care, and seek care if their newbornshowed signs of illness.11

Only one study focused on community mobilization asprimary component of its CBIP.16 Manandhar et al16

mplemented a community mobilization strategy by usingn action-learning cycle that was promoted by female fa-ilitators trained in perinatal health issues. These facilita-ors held monthly meetings for women in the community.hey encouraged active, participatory learning during theeetings by using strategies, such as stretcher schemes,
roduction and distribution of clean delivery kits, and

ptnlpw

aiuepwgCtTstmmStcnb

cms2stlcitishtBhigtwcrc

DNthttb

rsit

LOeUpcsaetainv

IPFiat

LHitaiL

timtaoueuramscedJ

tnd


icture card games to educate the women about preven-ion and treatment of common problems during preg-ancy and after delivery in the newborn. In the action-

earning cycle approach, the intervention package was notredetermined, but was decided upon by the individualomen’s groups.16

In Shivgarh, Uttar Pradesh, India,4 behavior change man-gement, which included community mobilization as one ofts components, was the primary intervention strategysed.21 This approach included multilevel engagement, forxample, household intervention through antenatal andostnatal home visitations, monthly participatory meetingsith groups of newborn community stakeholders, androup-level folk song meetings with embedded messages.ommunity mobilization created a supportive environment

o facilitate initiation and maintenance of behavior change.he intervention package was designed after conducting re-earch within the community to identify common behaviorshat placed the newborn at risk for key causes of neonatalorbidity and mortality (primarily infections and hypother-ia), but were amenable to change. CHWs called Saksham

ahayaks were recruited from the local community and werehe primary community mobilizers. Volunteers from theommunity, Saksham Kartas, were also important commu-ity mobilizers who helped to spread and sustain changes inehavior.10

BCC for the promotion of care seeking is an integralomponent of the Family-Community Care deliveryode. Although implemented differently across the trials,

ome form of BCC was included in all the studies (Table). Promotion of care seeking, defined as any behavior orervice that enabled or enhanced the ability of mothers orheir newborns to receive skilled medical care during de-ivery or for obstetrical or neonatal complications,4 wasommonly included. The most common form of care seek-ng was encouraging referrals to local health facilities (7 ofhe 9 studies; see Table 2) or instructing CBWs about thendications for referral. Manandhar et al16 also worked ontrengthening the capacity of the health care system toandle referrals by improving equipment and providingraining at all levels of the healthcare system. Similarly,aqui et al19 (India) used the already-existing governmentealth care system as a platform for implementing a CBIP

n the context of a large-scale nutrition and health pro-ram, implemented by CARE-India. They provided fur-her training for the auxiliary nurse-midwives and angan-adi workers and also recruited additional voluntaryommunity workers called “change agents” to enhance theeach of the program. However, none of the studies in-luded follow-up on referrals.

Although care seeking was an important component of thearmstadt et al15 trial, there was no significant decrease inMR after the intervention package was implemented, even

hough CBWs referred sick neonates to the local hospital andelped facilitate transport to the hospital. In the Kumar et alrial, no significant difference in care seeking was found inhe intervention arm compared with the comparison arm,
ut NMR was substantially reduced.10 Based on the mixed d
esults and lack of data to suggest that promotion of careeeking has an independent association with NMR or PMR, its difficult to determine the impact that it has on the reduc-ion in NMR.

ack of Cost-Effectiveness Datanly 3 of the 9 studies reported cost-effectiveness data. Bang

t al14 reported that the cost of their intervention package wasS$5.30 per neonate served. Baqui et al11 (Bangladesh) re-orted a lower cost per neonatal death averted for the home-are intervention arm (US$2995, including health systemstrengthening costs) compared with that reported by Man-ndhar et al16 (US$4397, including health systems strength-ning). These studies show that their packages are less thanhe cost-effectiveness threshold of US$14,872 per deathverted and are thus highly cost-effective.22 However, it ismportant to emphasize that most of the studies reviewed didot report cost-effectiveness data, and those that did usedarying methods.

ntegration of Interventionackage into Existing Health Care Systemive of the studies in this series used the existing health care

nfrastructure as a platform for their intervention package,nd strengthened aspects of the existing system in an attempto decrease neonatal and perinatal mortality.11,12,16,18,19

Both Jokhio et al12 and Bhutta et al18 used the existingHW Program, introduced by Pakistan’s Ministry ofealth in 1994, as a foundation for strengthening the ex-

sting system. In the government system, LHWs arerained and paid to provide antenatal care, contraceptivedvice, and immunizations. They do not receive trainingn delivering babies and there is minimal effort to linkHWs with TBAs.In the Bhutta et al18 package, LHWs received additional

raining in home-based newborn care and made antepartum,ntrapartum, and postnatal home visits. Unlike in the govern-

ent program, the LHWs were directly linked with TBAs inhe intervention package. The TBAs were also encouraged tottend a voluntary, 3-day training session, but the main focusf the program was on training the LHWs. Jokhio et al12 alsosed the LHW program as a platform for strengthening thexisting system, but their focus was on training the TBAs tose clean delivery kits and to recognize the indications foreferral of women for emergency obstetrical care. The TBAslso made 3 antenatal home visits, whereas LHWs played ainor role by supporting the TBAs and recording data. Both

tudies show that strengthening the existing primary healthare system, without the use of injectable antibiotics or otherxpensive technologies, can decrease NMR and PMR; NMRecreased by 28% in the Bhutta et al trial and by 29% in theokhio et al trial.

Baqui et al19 (India) conducted an independent evalua-ion of a CBIP implemented in the context of a large-scaleutrition and health program, implemented by CARE-In-ia (a nongovernmental organization, NGO) and the In-
ian government through the already existing government

hawccrsirhb

DCTtsiepatdlikdlwthrms

teLunslcvndcsptSc

btF

oachiwipfipssh

stccisaspwnaowbecnpckt

cotwcai

owHcborae

a


ealth care system. They provided further training for theuxiliary nurse-midwives and anganwadi workers whoere already making home visits to promote maternal and

hild health. They also recruited additional voluntaryommunity workers called “change agents” to enhance theeach of the program and to also make home visits. Thistudy did not result in a statistically significant reductionn neonatal mortality; however, the infants of women whoeceived a postnatal visit within 28 days after birth didave a significantly lower risk of neonatal death (reducedy 34%).

iscussiononclusions Based on Current Evidence

he current trend in delivering community-based interven-ions for improving neonatal and perinatal health is to bundleingle interventions into packages; we call these “Commun-ty-Based Intervention Packages” (CBIPs) in this article. How-ver, we identified only 8 large-scale, controlled trials and 1rogram evaluation that evaluated CBIPs and reported NMRnd/or PMR as outcomes. The Baqui et al19 study in India washe only program evaluation identified that used a rigorousesign and measured NMR. The finding of this evaluation (ie,

ack of mortality impact) suggests that although technicalnterventions to substantially reduce newborn deaths arenown, appropriate packaging and implementation of evi-ence-based interventions with high coverage and quality in

arge programs remains a challenge. No published studiesere identified from Africa. Therefore, large-scale studies

esting CBIPs in Africa are urgently needed, particularlyealth systems effectiveness trials. We are also in need ofesearch measuring the impact of successful CBIPs imple-ented at scale and integrated into the existing health care

ystem.Although there are a lack of data, particularly in Africa,

here have been recent advancements in developing anvidence-based approach to bundling interventions. Theives Saved Tool (LiST) is a cohort modeling modalitysed to estimate the impact of scaling up specific maternal,ewborn or child health interventions at the country,tate, or district level (ideally using coverage data at theevel of interest) on neonatal or child mortality. Althoughontinually being updated, 19 community-based inter-entions have been identified as having a statistically sig-ificant reduction in NMR.23,24 The LiST model was vali-ated by comparing actual changes in NMR to predictedhanges in 4 studies in South Asia.10,11,14,18 LiST washown to be a reliable tool for use by policymakers androgram managers to identify and prioritize interventionshat have the greatest impact on neonatal mortality inouth Asia.23 However, the model does not yet enableost-effectiveness modeling.

When comparing the key studies identified in this papery service delivery mode, we see that most of the interven-ions implemented by all 9 studies were, by design, in the
amily-Community Care delivery mode (Table 2). Only p
ne of the studies included Clinical Care interventions,nd only a few used Outreach interventions. These studiesonfirm that Family-Community Care interventions canave a substantial effect on neonatal and perinatal mortal-

ty3; however, a mixture of all the service delivery modes,ith continuity of care between homes and health facilities

s ideal. For example, among the studies in this series, therimary link between care in the home and local healthacilities was through referrals of sick newborns, althoughn some cases community-based activities also helped toromote antenatal care and institutional delivery, as de-cribed by Martines et al25 in this Seminars series. In onetudy, CHWs helped to facilitate transportation to theealth facilities.15

Several important themes were common across thetudies. Many of the trials included community-mobiliza-ion strategies, health education, and behavior changeommunication sessions (in groups and in the home),are-seeking modalities, training of CHWs, and home vis-ts during pregnancy and after birth. Although each of thetudies had slight variations in how these interventionpproaches were organized and implemented, theyeemed to be integral to the success of the CBIPs. The keylayer facilitating these interventions was the frontlineorker (CHW, TBA, community mobilizer, other commu-ity workers) in all the studies. For example, Bhutta et al18

nd Jokhio et al12 focused on training LHWs and TBAs andn strengthening the existing primary health care system,hereas Manandhar et al16 focused on a community-mo-ilization strategy for which they trained facilitators toncourage participatory learning in group sessions in theommunity. Kumar et al10 used a strategy that used ante-atal and postnatal home visits, and encouraged activearticipation in group meetings led by CHWs. A uniqueomponent of this strategy was the focus on a limited set ofey behaviors, and multilevel targeting of the various ac-ors and influencers in newborn care.10

However, there remain several hurdles that must be over-ome to establish a well-developed, evidence-based methodf bundling interventions. First, there is a lack of consistenterminology and language. For example, we must developell-defined and measurable definitions for terms, such as

ommunity mobilization, behavior change communication,nd care seeking to make accurate comparisons across stud-es.

Second, a similar inconsistency exists in the reporting ofutcomes. Many studies screened for inclusion in this reviewere excluded because they did not report NMR or PMR.owever, even among the 9 key studies that met study in-

lusion criteria, not all of them reported both outcomes (Ta-le 3). Some did not report the percentage reduction in NMRr PMR and others did not report 95% CIs. In the future, weecommend that NMR and/or PMR be reported as outcomes,s well as SBR and other indicators of maternal health, when-ver possible.

Because the CBIPs did not include the same componentscross the 9 trials, it was difficult to compare the impact of the
ackages. Not all the interventions included in the packages

wtbrdttmt

AVAstW

uupcaawpsaaq

dtotsdbmiid

tbodnidalcsrhp

nt

seiidap

rlhodofvAgtcheMm

scinctbmtafccgg

RfTeiiecaiL


ere evidence-based and the rationale for including some ofhe interventions was unclear. The interventions may haveeen bundled for reasons such as convenience or fundingequirements.4 Furthermore, the concept of bundling evi-ence-based intervention packages is also relatively new,herefore these studies may not have been designed to takehis into account.4 In addition, none of the studies tested theortality impact of the individual interventions included in

he packages.

Call for Attention to theulnerable Urban Newborn

lthough the focus of this work is on newborn health in ruralettings, there is also an urgent need for investment in CBIPsargeting vulnerable newborns in poor, urban communities.

e were unable to identify any urban trials of CBIPs.More than 22 million births take place annually among the

rban poor in low and middle income countries (calculatedsing a crude birthrate of 22 per 1000 live births and a slumopulation of 1.1 billion).26 In many low- and middle-in-ome countries, on the basis of reanalysis of Demographicnd Health Surveys data, infant and neonatal mortality ratesre considerably higher among the urban poor comparedith national averages.27 Cities in developing countries, inarticular, are growing at their edges. These peri-urbanpaces are home to migrants who work in construction, tradend perform other petty jobs. Migrant housing communitiesre usually transitory, informal, and suffer vulnerabilities re-uiring special attention.Newborn care in slum settings is suboptimal and scarcely

ocumented. Although rural and urban population charac-eristics differ, organization of health services in most devel-ping countries has revolved around rural systems. Hence,here is a need to systematically study the outcomes of urban-pecific newborn interventions and tailor CBIPs and serviceelivery strategies to the specific urban setting. Neonatesorn in poor, urban settings are at high risk of death due toultiple risk factors. One clear example is the sharp disparity

n attendance at delivery by trained medical personnel train-ng in poor compared with nonpoor urban settings in mosteveloping countries.27

Providing a slum-based “mother” support systemhrough women’s groups or through training of slum-ased workers has the potential of (1) addressing the lackf family support, which is a common reason for homeeliveries28; (2) overcoming the notion that childbirth is aatural process, which creates resistance to changing typ-

cal delivery methods using unskilled attendants29; (3) ad-ressing the vulnerabilities of powerless individuals, suchs migrant workers who do not have legal citizenship andive in short-term housing where home deliveries are moreommon28; and (4) encouraging care seeking, promotingafe and clean household practices, and supporting out-each services.30 In addition, encouraging slum-levelealth funds as a community risk pooling measure has the
otential to reduce delays in seeking care by mitigating the a
eed to undertake out-of pocket expenditure on transpor-ation, drugs, and other medical supplies.31

Training of slum-based birth attendants is worthy of con-ideration in light of the substantial number of home deliv-ries in slums in developing countries.32,33 Findings pertain-ng to cord care, clean hands and other birth-related practicesn urban slums from Mumbai,28 Dhaka,34 Nepal,35 and In-ore36 suggest that practices during home deliveries in dis-dvantaged urban settlements is far from satisfactory andredispose the newborn to risks of infection.In past decades, a greater emphasis has been placed on

ural health, thus, urban areas of most developing countriesack a planned primary health care infrastructure. Multipleealth providers administer health services without any co-rdination, leading to poor quality of services, as well asuplication of services in some areas, and undercoverage inthers. The large and growing private health sector, with itsee-for-service structure, makes it unaffordable to the mostulnerable. In addition, the private sector is unregulated.lthough NGOs and charitable organizations help fill in theaps, the services they provide are often fragmented. Overall,here are very few instances when NGOs have provided aomprehensive package of maternal newborn and childealth (MNCH) services. Furthermore, efforts to develop andvaluate innovative approaches to delivering comprehensiveNCH care to the urban poor have not been well docu-ented.Questions that future urban newborn health research

hould seek to answer include the following: (1) How canommunity-based support and action groups be effectiven an often heterogeneous urban slum society to reduceeonatal mortality? (2) What approaches and strategiesan ensure implementation of proven technical interven-ions in the urban context? (3) Could these interventionse augmented in urban areas with greater availability ofedical and nursing personnel? (4) How can we engage

he private sector and civil society organizations that havepresence in urban centers? (5) How can we involve the

ederal, regional, and city governments to influence policyhanges? (6) Could individuals from disadvantaged urbanommunities be motivated to take ownership of local pro-rams? (7) How can the challenging health needs of mi-rant populations be addressed?

ecommendationsor Future Investigationhere is a striking need for additional, large-scale RCTs to testvidence-based CBIPs in developing countries, particularlyn Africa. Several such studies are now underway but furthernvestigation is needed. As a research community, we need tostablish consistent, clearly defined terminology and proto-ols for designing trials and reporting outcomes so that were able to compare results across different settings. The ev-dence-based universal packages originally defined in theancet Neonatal Survival Series,3 elaborated on by Haws et
l6 and continually updated by the Child Health Epidemiol-

odfitttfi

auctp

sfAtovi

ctomntc

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3


gy Reference Group24 can be used as the “gold standard” foresigning and testing packages or may be used as a platformor developing even more comprehensive CBIPs. LiST hasncorporated at least 19 evidence-based interventions showno have a statistically significant impact on neonatal mortali-y,23 but cost-effectiveness modeling is not yet available withhis tool and such data from neonatal and perinatal healtheld trials are scarce.In addition, there is an urgent need to invest in research

nd program development focusing on neonatal health inrban areas. Newborn care in urban slums in developingountries around the world is suboptimal. Furthermore,here are no published studies that provide data on the im-act of CBIPs in the urban context.It is also crucial to integrate the CBIPs in rural and urban

ettings into the already-existing health care system toacilitate sustainability of the program and for scaling up.s we have seen in 2 studies in this report,12,18 it is possible

o strengthen the existing primary health care system with-ut introducing additional high-cost technologies or inter-entions to see a substantial reduction in neonatal mortal-ty.

Finally, there is a need for improving the continuum ofare between home and facility-based care. To date, studieshat have implemented CBIPs in rural settings have focusedn referrals as the primary link between activities in the com-unity and services in health care facilities. However, weeed to develop additional linkages and strategies to simul-aneously move families closer to clinical care, and to makelinical care available more peripherally in communities.37

eferences1. Black RE, Cousens S, Johnson HL, et al: Global, regional, and national

causes of child mortality in 2008: a systematic analysis. Lancet 375:1969-1987, 2010

2. Lawn JE, Cousens S, Zupan J: Four million neonatal deaths: when?Where? Why? Lancet 365:891-900, 2005

3. Darmstadt GL, Bhutta ZA, Cousens S, et al: Evidence-based, cost-effec-tive interventions: how many newborn babies can we save? Lancet365:977-988, 2005

4. Haws RA, Thomas AL, Bhutta ZA, et al: Impact of packaged interven-tions on neonatal health: a review of the evidence. Health Policy Plan22:193-215, 2007

5. Baqui AH, Arifeen SE, Williams EK, et al: Effectiveness of home-basedmanagement of newborn infections by community health workers inrural Bangladesh. Pediatr Infect Dis J 28:304, 2009

6. Yakoob MY, Lawn JE, Darmstadt GL: Stillbirths: epidemiology, evi-dence and priorities for action. Semin Perinatol 34:372-386, 2010

7. Wall SN, Lee ACC, Carlo W: Reducing intrapartum-related neonataldeaths in low- and middle-income countries—what works? SeminPerinatol 34:395-407, 2010

8. Simmons LE, Rubens CE, Darmstadt GL: Preventing preterm birth andneonatal mortality: exploring the epidemiology, causes and interven-tions. Semin Perinatol 34:408-415, 2010

9. Ganatra HA, Zaidi AKM: Neonatal infections in the developing world.Semin Perinatol 34:416-425, 2010

0. Kumar V, Mohanty S, Kumar A, et al: Effect of community-based be-haviour change management on neonatal mortality in Shivgarh, UttarPradesh, India: a cluster-randomised controlled trial. Lancet 372:1151-1162, 2008

1. Baqui AH, El-Arifeen S, Darmstadt GL, et al: Effect of community-based
newborn-care intervention package implemented through two service-
delivery strategies in Sylhet district, Bangladesh: a cluster-randomisedcontrolled trial. Lancet 371:1936-1944, 2008

2. Jokhio AH, Winter HR, Cheng KK: An intervention involving tradi-tional birth attendants and perinatal and maternal mortality in Paki-stan. N Engl J Med 352:2091-2099, 2005

3. Carlo WA, Goudar SS, Jehan I, et al: Newborn-care training and peri-natal mortality in developing countries. N Engl J Med 362:614-623,2010

4. Bang AT, Bang RA, Baitule SB, et al: Effect of home-based neonatal careand management of sepsis on neonatal mortality: field trial in ruralIndia. Lancet 354:1955-1961, 1999

5. Darmstadt GL, Choi Y, Arifeen SE, et al: Evaluation of a cluster-ran-domized controlled trial of a package of community-based maternaland newborn interventions in Mirzapur, Bangladesh. PLoS ONE5:e9696, 2010

6. Manandhar DS, Osrin D, Shrestha BP, et al: Effect of a participatoryintervention with women’s groups on birth outcomes in Nepal: cluster-randomised controlled trial. Lancet 364:970-979, 2004

7. Bang AT, Reddy HM, Deshmukh MD, et al: Neonatal and infant mor-tality in the ten years (1993 to 2003) of the Gadchiroli field trial: effectof home-based neonatal care. J Perinatol Off J Calif Perinatal Associa-tion 25:S92-107, 2005 (suppl 1)

8. Bhutta ZA, Memon ZA, Soofi S, et al: Implementing community-basedperinatal care: results from a pilot study in rural Pakistan. Bull WorldHealth Organization 86:452-459, 2008

9. Baqui A, Williams EK, Rosecrans AM, et al: Impact of an integratednutrition and health programme on neonatal mortality in rural north-ern India. Bull World Health Organization 86:796-804, 2008

0. Smith JB, Coleman NA, Fortney JA, et al: The impact of traditional birthattendant training on delivery complications in Ghana. Health PolicyPlan 15:326-331, 2000

1. Kumar V, Darmstadt GL: Behavior change for newborn survival inresource-poor community settings: bridging the gap between evidenceand impact. Semin Perinatol Semin Perinatol 34:446-461, 2010

2. Sachs JD: Health in the developing world: achieving the MillenniumDevelopment Goals. Bull World Health Organization 82:947-949,2004

3. Friberg IK, Bhutta ZA, Darmstadt GL, et al: Comparing modelled pre-dictions of neonatal mortality impacts using LiST with observed resultsof community-based intervention trials in South Asia. Int J Epidemiol39:i11-i20, 2010 (suppl 1)

4. Walker N, Fischer-Walker C, Bryce J, et al: Standards for CHERGreviews of intervention effects on child survival. Int J Epidemiol 39:i21-i31, 2010 (suppl)

5. Martines J, Bahl R: Why is continuum of care from community to healthfacilities essential to improved perinatal survival? Semin Perinatol 34:477-485, 2010

6. UN-HABITAT: Slum Estimates Data, 2003. Available at: http://ww2.unhabitat.org/programmes/guo/documents/Table4.pdf. Ac-cessed September 20, 2010

7. Agarwal S, Srivastava A, Kumar S: Urban health in developing coun-tries, in Gibbons C, Bali R, Wikramsinghe N (eds): Knowledge Man-agement in Urban Public Health. New York, Springer, 2010

8. Das S, Bapat U, More NS, et al: Prospective study of determinants andcosts of home births in Mumbai slums. BMC Pregnancy Childbirth10:38, 2010

9. Khan Z, Mehnaz S, Khalique N, et al: Poor perinatal care practices inurban slums: possible role of social mobilization networks. IndianJ Community Med 34:102-107, 2009

0. Agarwal S, Satyavada A, Patra P, et al: Strengthening functional com-munity–provider linkages: lessons from the Indore urban health pro-gramme, in Global Public Health 3:308-325, 2008

1. Agarwal SS, Agnihotri A, Ali N, et al: Risk pooling to Meet healthexpenditure among urban poor in Agra, India. 36th Annual Inter-national Conference of Global Health: Washington, DC, 2009

2. Agarwal S, Srivastava A, Kumar S: Knowledge management in Urbanpublic health, in Gibbons C, Bali R, Wikramsinghe N (eds): UrbanHealth in Developing Countries. New York, Springer,
2010
http://ww2.unhabitat.org/programmes/guo/documents/Table4.pdf

http://ww2.unhabitat.org/programmes/guo/documents/Table4.pdf

3

3

3

3

3


3. Aziz N, Akhtar S, Kaleem R: Newborn care practices regarding thermalprotection among slum dwellers in Rachna town, Lahore, Punjab. AnnKing Edward Medical University 16:1 SI, 2010

4. Moran AC, Choudhury N, Khan UZ, et al: Newborn care practicesamong slum dwellers in Dhaka, Bangladesh: a quantitative and quali-tative exploratory study. BMC Pregnancy Childbirth 9:54, 2009

5. Sreeramareddy CT, Joshi HS, Sreekumaran BV, et al: Home delivery
and newborn care practices among urban women in western Nepal:
a questionnaire survey. BMC Pregnancy Childbirth 6:27,2006

6. Agarwal S, Sethi V, Srivastava K, et al: Newborn care practices in urbanslums: evidence from central India. J Neonatalperinat Med 2:277-287,2009

7. Lee ACC, Lawn JE, Cousens S, et al: Linking families and facilities forcare at birth: what works to avert intrapartum-related deaths? Int J
Gynecol Obstet 107:S65-S88, 2009

Community-Based Intervention Packages for Improving...

Documents

Transcript of Community-Based Intervention Packages for Improving...