Using machine learning to estimate health spillover effects

Bruno Wichmann; Roberta Moreira Wichmann

doi:10.1007/s10198-023-01621-7

Using machine learning to estimate health spillover effects

Eur J Health Econ. 2023 Aug 6. doi: 10.1007/s10198-023-01621-7. Online ahead of print.

Authors

Bruno Wichmann¹, Roberta Moreira Wichmann^{2

3}

Affiliations

¹ Department of Resource Economics and Environmental Sociology, College of Natural and Applied Sciences, University of Alberta, 503 General Services Building, Edmonton, T6G-2H1, AB, Canada. bwichmann@ualberta.ca.
² World Bank, SCES Trecho 03, Lote 05, Ed. Polo 8, S/N, Brasília, DF, CEP 70200-003, Brazil.
³ Brazilian Institute of Education, Development and Research IDP, Economics Graduate Program, SGAS Quadra 607 - Modulo 49 - Via L2 Sul, Brasília, DF, CEP 70.200-670, Brazil.

PMID: 37543994
DOI: 10.1007/s10198-023-01621-7

Abstract

We develop a nonparametric model to study health spillover effects of policy interventions. We use double/debiased machine learning to estimate the model using data from 74 hospitals in Rio de Janeiro, Brazil, and examine cross-patient spillover effects during the COVID-19 pandemic. The pandemic forced hospitals to develop new protocols to offer intensive care to both COVID and non-COVID patients. Our results show that the need to care for COVID patients affects health outcomes of non-COVID patients. Controlling for a number of confounders, we find that mortality rates and length of stay of non-COVID ICU patients increase when hospitals simultaneously offer intensive care to both types of patients. Policy simulations suggest that an increase in the number of ICU beds can counter morbidity spillover, but it is unlikely to be a feasible approach to counter mortality spillover.

Keywords: Brazil; COVID-19 pandemic; Intensive care units; Machine learning; Non-COVID-19 patients; Spillover effects.