Essays on Health Policy Methods

Abstract

This dissertation includes 4 chapters on methods for health policy research. In the first 3 chapters, we consider common approaches to observational policy analyses. We discuss statistical and practical problems with the parallel trends test frequently used for difference-in-differences. Specifically, researchers often test the null hypothesis of "no violation" of this assumption. Failing to find evidence to reject this null, they may conclude that the assumption holds, especially if the point estimate of the violation is small. However, this approach inappropriately reverses Type I and Type II error. In many cases, it may miss important violations due to lack of sufficient statistical power. In the first two chapters, we provide alternatives to this test and to the traditional ad hoc model selection pathway in quasi-experimental analysis. In the third chapter, we consider in more detail the length of the pre-intervention period. We argue that rather than focusing on parallel trends over a long time horizon, researchers should optimize prediction of the treatment group by the comparison group. Based on this criterion, we present an estimator that leverages time-series cross validation to select optimal pre-intervention period weights. The last chapter arose as a result of the unexpected events of 2020. During the COVID-19 pandemic, all 50 states recommended or mandated school closures. To help guide school reopening policies, we developed an agent-based model of school-based transmission to assess and compare the impact of mitigation strategies on outbreak risk. Across age groups, we found that the risk of large outbreaks can be mitigated by investment in personal protective equipment to reduce classroom transmission potential, in surveillance testing with rapid turnaround, and across our society at large to suppress community transmission and prevent the introduction of infections into school settings.