The delivery of basic health products and services remains abysmal in many parts of the world where child mortality is high. This paper shows the results from a large-scale randomized evaluation of a novel approach to health care delivery. In randomly selected villages, a sales agent was locally recruited and incentivized to conduct home visits, educate households on essential health behaviors, provide medical advice and referrals, and sell preventive and curative health products. Results after

more »

... 3 years show substantial health impact: under 5-years child mortality was reduced by 27 percent at an estimated average cost of 68perlif e − yearsaved. The delivery of basic health products and services remains abysmal in many parts of the world where child mortality is high. This paper shows the results from a large-scale randomized evaluation of a novel approach to health care delivery. In randomly selected villages, a sales agent was locally recruited and incentivized to conduct home visits, educate households on essential health behaviors, provide medical advice and referrals, and sell preventive and curative health products. Results after 3 years show substantial health impact: under 5-years child mortality was reduced by 27 percent at an estimated average cost of $68 per life-year saved. (JEL I12, I18, J13, O15, O18) D espite significant reductions in child and infant mortality over the last few decades, about 1 in 13 children in sub-Saharan Africa still die before his or her fifth birthday (WHO 2017). Many, if not most, of these deaths can be avoided through simple preventative care and through simple, low cost treatments delivered at home. This means that an effective response to reduce child deaths is not out of reach. While health outcomes can be tied to a host of factors, both on the demand and supply side, there is limited evidence on effective and scalable solutions to the "Effect of a micro entrepreneur-based community health delivery program on under-five mortality in Uganda: a cluster-randomized controlled trial." The trial was approved by the ethic committee of Fondazione IRCSS (D2291696), by the Harvard IRB (protocol P20141-101), by the Uganda National Council for Science and Technology (UNCST) (SS3195), and by the IRB Office of the Joint Clinical Research Center (JCRC) in Uganda. The trial was registered in the Pan African Clinical Trials Registry (PACTR201308000601715) and in the American Economic Association's registry for randomized controlled trials (AEARCTR-0000530). We gratefully acknowledge two anonymous referees for many valuable comments and suggestions. We appreciate comments on an earlier draft from May Sudhinaraset, Jenny Liu, Dominic Montagu, and Rebecca Weintraub. We thank Aletheia Donald and Charles Ntale for help during different phases of the evaluation, and the IPA-Uganda and its staff, specifically