How do disasters affect health decision-making? Do disaster impacts operate, at least in part, through changes in economic preferences and psychological factors? Can public health programs help shield communities from the impacts of disasters? If so, through what mechanisms do such programs achieve their protective effects? We pursue these questions by studying a major disaster that happened to occur in the midst of a randomized evaluation we are conducting of a community-level public health program in Mozambique. Prior to the disaster, in 2017-2018, this study team had administered a baseline survey of approximately 4,700 households, and had facilitated random assignment of the program, Forca a Comunidade e Criancas (FCC, ?Strengthening Communities and Children?). Out of 76 communities, half were assigned to the treatment group, receiving the FCC program; the remainder serve as the control group. In March 2019, Cyclone Idai, the most destructive cyclone ever recorded in Africa, struck our study areas. The FCC program implements an interrelated set of health, educational, and economic interventions. Households are connected to community support networks, encouraged to visit and use local public health clinics, provided with information to improve health decision-making, and participate in microfinance programs. We first estimate human and economic losses, as well as impacts on risky sexual behaviors (including transactional sex), sexually transmitted infection (STI) testing, and STI incidence. We estimate impacts in a survey sample determined prior to the disaster, and are thus able to deal with selection biases (e.g., differential sample inclusion based on disaster exposure) that otherwise compromise treatment effect estimates. Impact estimates will exploit heterogeneity in hurricane exposure across study communities. Then, we will study how changes in individual preferences and psychological factors may mediate the effects of disasters on health decision-making. Examining motivated belief biases (in this context, over-optimism maintained by health risk denial) is particularly novel; it is a new behavioral economics frontier but has not been empirically evaluated in health-related contexts. Our measurement tools are piloted, refined, and ready for large-scale implementation. Finally, we will assess whether the FCC program helps shield communities from the negative impacts of disaster exposure on STI testing, treatment, and incidence. Causal inference will rely on the pre-disaster random assignment of the FCC program. We will also examine mechanisms through which the program achieves its protective effects. This project?s findings can help guide the design of public health programs in the future. Insights into the economics and psychology of post-disaster decision-making can suggest novel interventions to mitigate disaster-induced deterioration of health decisions. Evidence on whether and how public health programs shield people from disasters can provide additional rationales for such programs, and guidance on their optimal design.
