Do hospitals provide lower quality of care to black patients for pneumonia?

Abstract

Objectives: Recent studies reported lower quality of care for black vs. white patients with community-acquired pneumonia and suggested that disparities persist at the individual hospital level. We examined racial differences in emergency department and intensive care unit care processes to determine whether differences persist after adjusting for case-mix and variation in care across hospitals.

Design: Prospective, observational cohort study.

Setting: Twenty-eight U.S. hospitals.

Patients: Patients with community-acquired pneumonia: 1738 white and 352 black patients.

Interventions: None.

Measurements: We compared care quality based on antibiotic receipt within 4 hrs and adherence to American Thoracic Society antibiotic guidelines, and intensity based on intensive care unit admission and mechanical ventilation use. Using random effects and generalized estimating equations models, we adjusted for case-mix and clustering of racial groups within hospitals and estimated odds ratios for differences in care within and across hospitals.

Main results: Black patients were less likely to receive antibiotics within 4 hrs (odds ratio, 0.55; 95% confidence interval, 0.43-0.70; p < .001) and less likely to receive guideline-adherent antibiotics (odds ratio, 0.72; 95% confidence interval, 0.57-0.91; p = .006). These differences were attenuated after adjusting for casemix (odds ratio, 0.59; 95% confidence interval; 0.46-0.76 and 0.84; 95% confidence interval, 0.66 -1.09). Within hospitals, black and white patients received similar care quality (odds ratio, 1; 95% confidence interval, 0.97-1.04 and 1; 95% confidence interval, 0.97-1.03). However, hospitals that served a greater proportion of black patients were less likely to provide timely antibiotics (odds ratio, 0.84; 95% confidence interval, 0.78-0.90). Black patients were more likely to receive mechanical ventilation (odds ratio, 1.57; 95% confidence interval, 1.02-2.42; p = .042). Again, within hospitals, black and white subjects were equally likely to receive mechanical ventilation (odds ratio, 1; 95% confidence interval, .94-1.06) and hospitals that served a greater proportion of black patients were more likely to institute mechanical ventilation (odds ratio, 1.13; 95% confidence interval, 1.02-1.25).

Conclusions: Black patients appear to receive lower quality and higher intensity of care in crude analyses. However, these differences were explained by different case-mix and variation in care across hospitals. Within the same hospital, no racial differences in care were observed.

Figure 1

Odds ratios (OR) and 95% confidence intervals of different care processes in the (A) emergency department and (B) intensive care unit (ICU) in black compared to white subjects. The OR are estimated in three models: an unadjusted model, a model adjusting for case-mix, and a model adjusting for case-mix and different clustering of black and white subjects among hospitals using random effects. For case-mix adjustment for all quality and intensity measures, we adjusted for age, sex, insurance status, and Charlson and pneumonia severity score scores. The among-center OR reflects differences in treatment among hospitals serving the lowest and highest proportion of black patients (hospitals with >80% black patients compared to those with no black patients).