Approximately eleven million children under five years die each year and almost half of these deaths occur in sub-Saharan Africa. More than half of these children will die of diseases (e.g., diarrhea, pneumonia, malaria, and neonatal disorders) that could easily have been prevented or treated if the children had had access to a small set of proven, inexpensive services (Black, Morris, and Bryce 2003; Jones et al. 2003). Maternal mortality remains unacceptably high across much of the developing world. In sub-Saharan Africa a woman is 300 times more likely to die from treatable or preventable complications of pregnancy and childbirth compared to a woman in the developed world. This dramatic picture of child and maternal mortality calls for urgent attention to better understand which programs and policies work and which do not, and how effective policies can be used to change the negative trend. The research project “Fighting child and maternal deaths by experimenting with health delivery systems: Evidence from randomized controlled trials in Nigeria and Uganda” has been funded by Riksbankens Jubileumsfond with the aim to start filling this gap in order to rigorously evaluate two health service delivery programs implemented in Nigeria and Uganda and study its effect on child and maternal health.

The two health service delivery programs implemented within this project has been evaluated using a randomized controlled trial methodology which is a well-regarded quantitative impact evaluation method that enables rigorous evaluations of the impact of the interventions to answer the research and policy questions effectively (Duflo, Glennerster, and Kremer, 2008). Both projects have collected new and unique survey data from households.

The first program evaluated within this research project is an innovative health delivery model (the Living Goods’ model) to improve health services to poor people in Uganda. The aim of the project is to improve access to and adoption of simple, proven health interventions in rural and peri-urban areas, while at the same time creating sustainable livelihoods for community-based health workers. The Living Goods model creates “Avon-like” networks of door-to-door mobile Community Health Promoters (CHPs) who provide basic health education and make a modest income by selling a diverse basket of basic health goods anchored by essential items emphasizing prevention like bed nets and water treatment. Most health products are sold at prices considerably below market, and in some cases provided for free.  The system of Community Health Promoters is expected to be fully financially self-sustaining. The main objective of the evaluation was to assess the impact on child health of the Living Goods innovative model to deliver primary health in poor developing countries.

By the beginning of 2011 the program, implemented as a randomized controlled trial across 214 rural villages spread across Uganda, was fully operational in all treatment villages with one community health promoter (CHP) locally recruited to the program.  To evaluate the program we conducted a household survey of 7,018 households and their 11,563 under-5 year old children. The results after the program had been running for three years show that the intervention resulted in a substantial health impact: under-5 child mortality was reduced approximately by 27%, infant (i.e. under 1 year) mortality by 33% and neonatal (i.e. under 1 month) mortality by 28%. The evidence further shows that households in treatment villages were more significantly likely to use products sold by the CHPs, such as insecticide treated bed nets and oral re-hydration salts for treating diarrhea – consistent with the effects on mortality. We also find the large increases in the treatment relative the control group in terms of behavior for follow-up visits and counseling. Households with a newborn baby were more than 70% more likely to have received a follow-up visit in the first week after birth, and households with a child under-five years old that fell sick with malaria or diarrhea were, respectively, 73% and 62% more likely to have received a follow-up visit. For households with infants that fell sick with malaria or diarrhea the increases were 109% and 105%, respectively. While a growing body of evidence has identified effective interventions that can be delivered by community health workers, a key consideration for the success and sustainability of such programs is how high-quality performance by community workers can be achieved and maintained. This study is the first impact evaluation of a community health delivery intervention based on an incentivized approach. We also estimated the cost-effectiveness of the program and found the estimated cost per averted death under-five during the study period to be $4,237. To the best of our knowledge, this is the first assessment of the cost effectiveness of a community health worker program, based on a rigorous randomized controlled trial approach. This paper is a CEPR working paper and has been submitted for publication to an economics journal.

In the second part of this research project, I have worked on a program in northern Nigeria that separately evaluate three different interventions that focus on reducing maternal and child mortality through health education, building and sustaining support for appropriate health seeking behaviour and practice, improving confidence and accountability of health facilities to the communities they serve and increase utilization of health services. This project was undertaken in selected communities in very high mortality zones in northern Nigeria. The three project interventions were evaluated through a randomized controlled trial. The first intervention was a Voluntary Health Worker Program (VHW) where trained voluntary health workers conduct door to door campaigns on personal hygiene, household hygiene and sanitation, and upon identification of a pregnant woman will take the woman and the family through a Birth Preparedness schedule. The second intervention was the standard package of VHW services with the provision of a birth kit to each pregnant woman.  A birth kit is a small package of sterile materials that can be used either in a clinic or at home to diminish infection risks for both the mother and the child. The third intervention was the standard package of VHW services supplemented with community-wide media activities intended to reinforce community norms and perceptions over appropriate health practices around childbirth and maternal and infant care. We surveyed 7000 women in 96 communities over the period of four years. Preliminary results from the evaluation indicate that the interventions had a significant impact on increasing antenatal and postnatal care utilization, but did not increase the probability of a facility-based delivery, and did not significantly affect maternal or neonatal health outcomes. Further exploration suggests that this may reflect both weak coverage of the intervention and a low quality of care at health facilities. This research paper is currently in a working paper draft and the plan is to submit the paper to a top-field journal by early 2018. In addition, we have published two public health papers from the project and the plan is to write more public health papers using the data we have as well.