Physician-patient race-match reduces patient mortality

Abstract

This paper assesses the impacts of physician-patient race-match, especially Black patients paired with Black physicians, on patient mortality. We draw on administrative data from Florida, linking hospital encounters from mid-2011 through 2014 to information from the Florida Physician Workforce Survey. Focusing on uninsured patients experiencing unscheduled hospital admissions who are conditionally randomly assigned to physicians, we find that physician-patient race-match for Black patients reduces the likelihood of within-hospital mortality by 0.28 percentage points, a 27 % reduction relative to the overall mortality rate. An alternative identification strategy relying on instrumental variables provides a similar finding.