Tradeoffs in the Distribution of a Scarce Health Resource: Evidence from a Large Randomized Natural Experiment

Our long-term goal is to inform evidence-based policies to improve public health when resources are scarce. We harness random assignment of early access to a scarce health resource to a large population through Michigan Medicine, a large hospital system with national reach. We can now examine years of detailed folllowup data. We will situate our results in a broader context now so that they have practical implications for health care delivery of resources that become scarce in the future. Aim 1: Estimate differences in takeup along many detailed dimensions of health and demographics using equality in early access via random assignment and rich data that we construct and disseminate. Aim 2: Quantify how much of the differences by health and demographics that persist given equality in early access via random assignment can be explained by various mechanisms that operate through health behaviors, health conditions, and social and environmental factors available in our electronic health records and survey data. Aim 3: Characterize how the composition of the treated population changes by health and demographics as the share treated expands with randomization, and relate those changes to differences in downstream impacts of treatment on health and economic wellbeing.