Reducing child mortality in the last mile: Experimental evidence on community health promoters in Uganda


Published 30.11.18
Photo credit:
Esther Havens/flickr

An innovative community health programme providing financial incentives to health workers led to a 27% reduction in child mortality in rural Uganda

The World Health Organisation estimates that about one child in 13 born on the African continent dies before reaching the age of five. What makes this figure even more dramatic is the fact that most of these deaths are caused by diseases that are cheap and easy to prevent or treat with proven interventions that are available in many parts of the world. How to make basic health and medical solutions available to everyone is one of the biggest development challenges of our times.

In recent years many governments and NGOs have attempted to tackle this challenge by setting up Community Health Worker (CHW) programmes (Singh and Sachs 2013). The details of these programmes vary, but in their most common configuration they rely on volunteers that receive a short training on basic health practices and then become the first point of contact for health-related issues in their communities. The objective is to bridge the gap between the official health system and the local communities, especially in those areas where accessibility of health services is limited, and child mortality is the highest. 

While systematic reviews of existing studies indicate that CHWs can be important in promoting positive health behaviour and in providing basic curative and health services (e.g. Haines et al. 2007, Gilmore and McAuliffe 2013), the findings from reviews of randomised controlled trials of CHW programmes and CHW-led interventions are mixed (e.g. Lewin et al. 2010), with many studies finding no impact on child mortality (e.g. Kirkwood et al. 2013, Boone et al. 2016). A commonly discussed threat to the effectiveness of these programmes stems from the lack of financial incentives for the health workers. Competing opportunities such as paid-work or home production, may indeed lead CHWs to devote less time to volunteer health care work. How to incorporate incentives to motivate CHWs in large-scale programmes, and the impact these programmes will have, are open questions. 

The community health promoter programme in Uganda

In 2007, Living Goods, a US-based NGO active in Uganda, in collaboration with BRAC Uganda, began piloting a new community health delivery model – the Community Health Promoter (CHP) programme – intended to improve maternal, newborn, and child health. In many aspects the programme resembles a standard CHW one. Health workers are selected within each community through a competitive process among female community members aged 18 to 45 who apply for the position and who possess basic writing and mathematics skills. Eligible candidates receive an initial two weeks of training, followed by monthly refreshment sessions. The selected candidates are expected to conduct home visits within their community, educate households on essential health behaviours, provide basic medical advice, and refer the more severe cases to the closest health facility. On top of this – and this is the innovation of the programme – the CHPs make a modest income by selling a diverse basket of basic health goods, ranging from anti-malaria drugs and vitamins to soap and fortified foods. The CHPs purchase these products directly from the NGOs at wholesale price and earn a margin on each product sold, while keeping the price lower than the prevailing market price.

The idea is that the incentives on the health products - coupled with small financial incentives to encourage timely services - would not only move households down the demand curve for these products, but also motivate community health promoters to actively provide basic health services to the community. In sum, the CHPs operate as micro-entrepreneurs within the local health care market, with financial incentives to meet household demand and improve child health. 

The impact

In our study (Björkman Nyqvist et al. 2018) we evaluate whether the CHP programme is successful in reducing child mortality. The programme was randomised across 214 rural villages spread across 10 districts of Uganda. Each treatment village had at least one active community health promoter, while the control villages had none. 

The CHPs began their activity in the treatment villages at the beginning of 2011. Our results after three years show that the programme had a substantial health impact: in treatment villages under-5 child mortality was reduced by approximately 27%, infant (i.e. under 1 year) mortality by 33%, and neonatal (i.e. under 1 month) mortality by 28%, compared to control villages. These effects are supported by changes in health knowledge, increased utilization of preventive and treatment health services, as well as increased maternal, newborn, and child health service coverage. 

Importantly, while we document large increase in some incentivised tasks, like visits of newborns, we also document large increases for tasks that were not directly incentivised, such as follow-up visits of children who were sick in malaria and diarrhoea. While we cannot rule out that these non-incentivised services were provided for profit maximising reasons (i.e. as a way to increase the demand for products sold by the CHPs), the findings are also consistent with the view that the CHPs also had non-pecuniary reasons to serve the community and that the pecuniary incentives did not (fully) crowd out non-pecuniary motivations. 


Our analysis indicates that the average cost per averted death over the study period was approximately $4,000. Taking into account that life expectancy is vastly improved conditional on surviving the first few years of life, the estimated average cost per life-year gained is $68. This figure compares favourably to existing estimates of various community health interventions, which range from $82 to $3,400 per life-year gained (Borghi et al. 2005, McPake et al. 2015). 


To the best of our knowledge, this study is the first impact evaluation of a community health delivery intervention based on an incentivised approach. Unlike previous studies that have primarily focused on the impact of specific interventions that could be delivered effectively in a community setting, the focus here is on how to ensure that community health workers successfully implement a set of interventions proven to be effective if delivered and the impact that may have on child health. 

Although the experimental design does not allow us to pin down the specific contribution of the incentives, the results clearly indicate that a community health worker programme that embeds financial incentives for the health workers can be highly effective in reducing child mortality. 

With the accumulated know-how we have today few would question the benefits of community health care provision. How to best motivate the CHW and ensure that they deliver timely and appropriate services is, however, an open question and motivates the continued search for innovative approaches. This particular CHP programme harnesses the power of franchised direct selling (business-in-a-bag) to provide community health providers with incentives to increase access to low-cost, high-impact health products and basic newborn and child health services. The programme is currently being scaled up and by end of the year it is estimated to reach over 5.3 million individuals in more than 6,700 villages. It should be noted that the impact of the CHP programme was conditional on existing facility based professional health care being available, as referral services are a crucial component to the programme. These findings should encourage government and non-government organisations to continue improving their facility based care, but also points to the importance of integrating the programme into the existing health service provision strategy. 


Björkman Nyqvist, M, A Guariso, J Svensson, D Yanagizawa-Drott(2018), “Reducing child mortality in the last mile: Experimental evidence on community health promoters in Uganda”, American Economic Journal: Applied, forthcoming.

Boone, P, D Elbourne, I Fazzio, S Fernandes, C Frost, C Jayanty, R King, V Mann, G Piaggio, A Dos  Santos, P Walker  (2016), “Effects of community health interventions on under-5 mortality in rural Guinea-Bissau (EPICS): A cluster-randomised controlled trial”, Lancet Global Health 4(5): 328-35.

Borghi, J, B Thapa, D Osrin, S Jan, J Morrison, S Tamang, B Shrestha, A Wade, D Manandhar, A Costello(2005), “Economic assessment of a women’s group intervention to improve birth outcomes in rural Nepal”, Lancet 366(9500): 1882-1884.

Gilmore, B and E Mcauliffe(2013), “Effectiveness of community health workers delivering preventive interventions for maternal and child health in low- and middle-income countries: A systematic review”, BMC Public Health 13: 847.

Haines, A, D Sanders, U Lehmann, A K Rowe, J E Lawn, S Jan, D Walker and Z Bhutta  (2007), “Achieving child survival goals: Potential contribution of community health workers”, Lancet 369: 2121-2131.

Kirkwood, B R, A Manu, Ah Ten Asbroek, S Soremekun, B Weobong, T Gyan, S Danso, S Amenga-Etego, C Tawiah-Agyemang, S Owusu-Agyei, Z Hill(2013), “Effect of the Newhints home-visits intervention on neonatal mortalityrate and care practices in Ghana: A cluster randomised controlled trial”, Lancet 381(9884): 2184-2192.

Lewin, S, S Munabi-Babigumira, C Glenton, K Daniels, X Bosch-Capblanch, B VanWyk, J Odgaard-Jensen,M Johansen, G Aja, M Zwarenstein, I Scheel(2010), “Lay health workers in primary and community health carefor maternal and child health and the management of infectious diseases (Review)”, The Cochrane Library 3.

Mcpake, B, I Edoka, S Witter, K Kielmann, M Taegtmeyer, M Dielman, K Vaughan, E Gama, M Kok, D Datiko, L Otiso, R Ahmed, N Squires, C Suraratdecha, G Cometto (2015), “Cost-effectiveness of community-based practitioner programmes in Ethiopia, Indonesia and Kenya”, Bulletin of the World Health Organisation 93: 631-639.

Singh, P and J D Sachs (2013), “1 million community health workers in sub-Saharan Africa by 2015”, Lancet 382: 363-365.