Can Community Health Services Offset the Effect of Poverty and Low Maternal Educational Attainment on Childhood Mortality? Evidence from the Navrongo Experiment in Northern Ghana

The Government of Ghana has instituted a National Poverty Reduction Program with an initiative known as the Community-based Health Planning and Services (CHPS) as its core health development strategy. Originally launched as an experimental project of the Navrongo Health Research Centre, CHPS is currently a national initiative for shifting the focus of primary health care from clinics to communities. Four contrasting strategies of community health care were tested. In response to early evidence that community nursing could impact on survival, CHPS was launched to scale up the Navrongo community health service delivery system. The program currently functions in over 1000 villages dispersed in all 172 districts of Ghana. This paper presents a longitudinal demographic survival histories of 17,967 under age five children who were observed over the July 1993 to December 2010 period. Analysis of trends show that placing nurses in CHPS communities was associated with more pronounced mortality declines than was observed in comparison communities. Differentials show that Navrongo study areas exhibit equity relationships that are often observed elsewhere in Africa: Children whose parents are uneducated and relatively poor experience significantly higher mortality than children of the educated and less poor. Time conditional Weibull regression models estimate the impact of CHPS on health equity by estimating the interaction of equity indicators with duration of household exposure to CHPS service operations. Results show that the association of maternal education and poverty with childhood mortality is offset by community exposure to community health nursing services. This finding lends support to policies that promote community health services as a component of poverty reduction programs. Introduction: Evidence across Africa has consistently demonstrated the adverse association between low parental educational attainment and child survival, a relationship that is compounded by the effects of relative poverty (Muhuri 1995; Montgomery et al. 2000; Filmer and Pritchett 2001; Gwatkin 2002a; Bawah and Zuberi 2005; Debpuur et al. 2005; Kahn et al. 2005; Nathan et al. 2005). Scientific research has also revealed the impact of community-based primary health care on effectively reducing childhood morality (Pence, et al. 2007; Haines, et al. 2007; Community Directed Interventions Study Group. 2010; Binka et al. 2007). In recognition of these findings, it has been widely assumed that the adverse child-survival effects of poverty and low parental education may be offset by the provision of convenient and inexpensive community village based services (Wallerstein et al. 2011). Community health service delivery has become a central component of policies aimed at improving health equity, as facilitated by 3 Heilbrunn Department of Population and Family Health, Mailman School of Public Health, Columbia University, 60 Haven Avenue, New York, NY 10032 U.S.A. 4 Navrongo Health Research Centre, Ghana Health Service, GPO Box 114, Navrongo, Upper East Region, Ghana 3 the provision of non-differential health services across social groups (Lake 2012; Marmot 2007; Victora 2003). Nonetheless, an examination of the mitigating impact of community-based health services on the association between poverty and child survival requir8ues direct appraisal. Such evidence seldom exists. Despite increased efforts across sub-Saharan Africa to improve health access through the expansion of primary health care systems, the health conditions of many vulnerable populations, especially children, continue to deteriorate (Timaeus 1997a, 1997b, and 1999; Ahmad et al. 2000; Black et al. 2003; Hill 1993) Although modest improvements have been identified, the region continues to lag behind the rest of the world (United Nations 2012; World Bank 2005). Extrapolations of trends also suggest that the Millennium Development Goals (MDGs) for child survival are not on target, as suggested by the sub-Saharan Africa’s continuously slow reduction rates (Murray et al. 2007; Sahn and Stifel 2003). Furthermore, controversy also persists regarding the importance of focusing policies on the promotion of specific health interventions versus a systems approach which integrates packages of essential services. Achieving the child-survival MDG may require simultaneous attention to a range of prevalent sources of childhood morbidity, such as diarrheal diseases, malaria, acute respiratory infections, and malnutrition. The concurrent delivery of interventions relies heavily on the strengthening of the entire health system, which has been deemed as a prerequisite for the achievement of all health-related MDGs (Travis et al. 2004). Some proponents of healthsystems reform and sector-wide community health services also argue that policies focused on promoting specific health interventions are tantamount to abandoning global social and economic health-equity goals (Magnussen et al. 2004). The child-survival effects of maternal educational attainment have been demonstrated for nearly three decades (Caldwell 1979), but questions persist about the efficacy of community based health care on counteracting their harmful impact (Desai and Alva 1998). This paper examines the empirical basis for a poverty reduction program focused on communitybased primary health care by testing the hypothesis that the presence of such services are able to offset the detrimental effects of poverty and low educational attainment in the impoverished rural setting of northern Ghana. Our research is based on a 1996-2003 quasi-experimental study of the Navrongo Health Research Centre which assessed the health and survival effects of community-based primary health care (Binka et al. 1995; Phillips et al. 2006). The program, referred to as the Navrongo Community Health and Family Planning project, served as a field trial of organizational strategies for the delivery of health care in rural and impoverished settings. Assessment of the direct experimental effects indicated that the placement of nurses directly in communities, equipped with the training and equipment to provide basic primary health care services, could accelerate mortality reduction (Binka et al. 2007) and fertility reduction (Debpuur, et al. 2002). Consequently, these findings lead to the Government of Ghana’s formal