Optimal Allocation Rule for Infectious Disease Prevention Under Partial Interference and Outcome-Oriented Targets With Repeated Cross-Sectional Surveys

Chan Park, Guanhua Chen, Menggang Yu, Hyunseung Kang

Abstract: When developing allocation policies for infectious disease prevention, policymakers often set outcome-oriented targets and measure the policy's progress with repeated cross-sectional surveys. For example, the United Nations' Integrated Global Action Plan for Pneumonia and Diarrhea (GAPPD) recommended increasing water, sanitation, and hygiene (WASH) resources to reduce childhood diarrhea incidence and outlined a specific level of diarrhea incidence to achieve. For policymakers, the goal is to design an allocation rule that optimally allocates WASH resources to meet the outcome targets. The paper develops methods to estimate an optimal allocation rule that achieves these pre-defined outcome-oriented targets using repeated cross-sectional surveys. The estimated allocation rule helps policymakers optimally decide what fraction of units in a region should get the resources based on its characteristics. Critically, our estimated policies account for spillover effects within regions in the form of partial interference, and we characterize the policies' performance in terms of excess risk. We apply our methods to design Senegal's WASH policy to prevent diarrheal diseases using the 2014-2017 Demographic and Health Survey. We show that our policy not only outperforms competing policies but also provides new insights about how Senegal should design WASH policies to achieve its outcome targets.