Does the Patient-Provider Match Matter? Evidence from Gender Matching
AbstractThere is a long-held belief that the relationship between a patient and their provider is fundamentally important to the delivery of high-value healthcare. However, there is little empirical evidence on the effects of the patient-provider relationship on patient health outcomes. Existing research has estimated the impacts of patient-provider concordance in demographic characteristics in emergency delivery (e.g. Greenwood et al. 2018) and preventive care uptake (e.g. Alsan et al. 2019), but there is a paucity of evidence in settings where such relationships arguably matter the most—namely, outpatient care and management of chronic disease. In this paper, I begin to fill this gap in understanding by providing evidence on the impact of patient-provider gender concordance on utilization and quality of care in the outpatient setting for Medicaid enrollees with chronic conditions.
Specifically, I use data from a novel and largely unutilized national Medicaid claims database, the Medicaid Analytic eXtract (MAX), together with an instrumental variables design to address patient sorting across providers. My instrument for the gender of a patient’s provider is the share of providers of that gender treating other enrollees in that patient’s zip code, year, and month of enrollment. I evaluate the effect of gender concordance on several measures of quality of care, such as days from first visit to diagnosis and number of preventable condition-specific ambulatory care sensitive condition (ACSC) hospitalizations; and utilization, such as general inpatient hospitalizations and ED visits.
Preliminary results indicate that male patients have significantly worse health outcomes when treated by male providers than female providers, both in terms of overall utilization and quality of care indicators. There is no significant difference in outcomes for female patients treated by female versus male providers, though female patients tend to be diagnosed more quickly when treated by female providers.
These findings suggest that gender concordance for male patients and male providers may be problematic for certain patient populations, particularly for male patients who are exogenously matched to male providers. Furthermore, my preliminary evidence suggests that male providers who accept Medicaid patients may be adversely selected on the margin and have worse outcomes than their female colleagues.