With or Without: an Assessment of Brazil’s More Doctors Program on Population Health

Abstract

Renewed commitment to the Universal Health Coverage agenda reinvigorated interest in primary care. Yet, there is limited evidence on the performance of community-level targeting methods used in large-scale primary care programs. Empirical evidence from developing countries on the contribution of primary care on population health over time relies primarily on ecological studies. This dissertation addresses these knowledge gaps by providing a comprehensive assessment of Brazil’s More Doctors Program (MDP). We, first, examine the beneficiary targeting performance of the MDP in the period between 2013-2017 (Chapter II). We find that almost 70% of municipalities with vulnerability designation enrolled in the MDP from 2013 to 2017; whereas 33% of municipalities that joined MDP in this period did not match any of the vulnerability criteria. We show that vulnerable municipalities that did not receive MDP physicians had the highest poverty rate and lowest physician availability in primary care settings. Next, we evaluate the impact of MDP on hospitalizations due to a set of conditions considered to be avoidable with timely and appropriate provision of primary care services. We combine coarsened exact matching and different-in-difference methods to construct valid counterfactual estimates using a municipal-aggregated dataset covering the period from 2009 to 2017. We show that the MDP led to 10.04 (95%CI: -19.16, -0.92) fewer hospitalizations per 100,000 municipal inhabitants for avoidable cardiovascular conditions (Chapter III). The estimated MDP coefficient for avoidable hospitalizations among children and adolescents under 20 years of age was 1.84 (95%CI: -5.34, 1.66), suggesting that the MDP did not reduce the avoidable hospitalizations for this age group (Chapter IV). The program contributed to the narrowing of urban-rural disparities, with an estimated MDP coefficient of -7.43 (95%CI: -13.20, -1.65) for rural municipalities, compared to 1.23 (CI95%: 3.16,5.61) for urban areas. The dissertation findings highlight that targeting methods used in large-scale primary care programs have important implications for the extent to which they are able to allocate program resources to their target populations. They demonstrate the complexities in translating increased investment in primary care. They further underscore that a sizable infusion of resources in primary care in underserved settings can improve population health.