Hospital Quality, Minority Patients and Other Hospital Characteristics

Objectives: Several studies have identified minority patients' use of low quality providers as an explanation for disparities in hospital care. These studies note the concentration of minority patients in a relative few hospitals and point out the differences in quality provided by these hospitals relative to other hospitals that serve practically an all-white patient population. The policy recommendation usually offered is to target quality improvement and disparity reduction efforts to minority serving hospitals. This study explores whether using the racial/ethnic composition is a better proxy for identifying low quality hospitals than using other hospital characteristics.

Methods: Using 3 years of state inpatient discharge data from 13 states and the AHRQ hospital quality indicators, we computed composite inpatient quality and patient safety measures. We linked this data to the AHA hospital survey and Medicare Cost Reports. We computed correlations between the composite measures and the individual IQI and PSI to determine how well they track the variation in the individual measures. To explore the association between hospital characteristics and hospital quality, we conducted bivariate and multivariate analyses. We measured patient race/ethnic composition, organizational structure, payer mix, facility resources, and financial health of each hospital. For the bivariate analyses, we categorized hospitals into thirds (high, medium and low quality) based on their composite scores. For each composite score categorical variable, we conducted 2 by N chi-squared analyses for the categorical data and ANOVAs for the continuous variables. We used the continuous composite scores for the multivariate analyses. We estimated least squares models including all of the hospital characteristics with state level fixed effects.

Results: The percent of minority patients was not uniformly associated with lower quality. We found that the percent of black patients was associated with low quality for surgical procedures, but the percent of Hispanic patients was associated high quality for surgical procedures and patient safety. Better indicators of low quality hospitals were payer mix, occupancy rate, nursing ratios and average cost per discharge.

Conclusion: Payer mix, occupancy rate, nursing ratios and average cost per discharge instead of race/ethnic composition may be better ways of identifying hospitals that need assistance providing high quality care. Given the difference in reimbursement, it is easier to understand why hospitals serving high proportions of publicly insured and uninsured patients may lack the resources to provide high quality care. The associations between occupancy rate, patient-to-nurse ratios and average costs were also related to resource constraints.