Working Papers

Does Research Save Lives? The Local Spillovers of Biomedical Research on Mortality" with Bruce A. Weinberg

NBER working paper No. 29420. 

Abstract: This paper investigates the local impact of biomedical research on mortality in the USA. Causally estimating the marginal value of biomedical research is challenging due to a lack of micro data linking health outcomes to plausibly exogenous variation in research. We create a new linkage between a research database (PubMed) and administrative death records that enables research to be related to mortality at the geographic, disease and time level. We then estimate the marginal impact of biomedical research on mortality using hospital market (HRR) level shocks to research activity by disease. Our identification strategy builds on the literature on the dissemination of knowledge, specifically that of local knowledge spillovers. By utilizing variation across diseases, time and distance from research we control for additional trends relative to the current literature. Our results show that an additional research publication on average reduces local mortality from a disease by 0.35%. Our results also provide novel evidence that there are health benefits to the local communities (local spillovers) in which biomedical research is conducted.

“Medical Practice Shutdowns and Healthcare Utilization: Evidence from the COVID-19 Pandemic"  with Xuechao Qian 

Abstract: This paper studies the effect of the permanent closure of office-based physician practices on patient healthcare utilization. Using a nationwide claims database we first show that the exit rate of office-based physician practices increased in 2020. Using the supply shock caused by the financial impact of COVID-19, we show that permanent physician practice closures increase the probability of visiting hospitals, emergency departments and the cost per service. People from areas with lower income levels and higher shares of minority population, senior people, and patients using public insurances are disproportionately affected.