Ayanian JZ, Weissman JS, Chasan-Taber S, and Epstein AM.
Quality of care by race and gender for congestive heart failure and pneumonia.
Medical Care 1999;37(12):1260-9.

There is considerable literature showing that blacks are less likely to undergo major procedures for a range of health conditions; however, it is uncertain whether these findings indicate that they are also receiving lower quality of care. This study analyzed quality of care as recorded in medical records for a large cohort of Medicare recipients aged 65 years or older discharged during the period from September 1991 to August 1992 from hospitals in Illinois, New York and Pennsylvania with a primary diagnosis of congestive heart failure or pneumonia. The sampling process was based on diagnosis, readmission status, and race. Multiple admissions for the same patient were not included.

Among those with heart failure and pneumonia, black patients were younger and more likely to be female, covered by Medicaid, residents of poor communities, and treated in teaching hospitals. Black patients were less likely to have do-not-resuscitate orders and were treated in rural hospitals less often than non-black patients. Additionally, black patients with congestive heart failure had lower sickness at admission scores, and black patients with pneumonia were less likely to be admitted from a nursing home.

With regard to gender differences, women were older, more often of black race, and more often admitted to teaching and urban hospitals. Women with pneumonia had lower sickness at admission scores and were more often covered by Medicaid and admitted from nursing homes. There were no gender differences in likelihood of residing in poor communities and having do-not-resuscitate orders.

Quality of care was measured in two ways. The first method utilized implicit measures. Physician reviewers from each state conducted implicit assessments using a 50-item structured review of patients' medical charts. Two global questions were asked: "Considering everything you know about this patient, how would you rate the overall quality of care?" and "Based on the care provided to this patient, would you send one of your patients to these physicians in this hospital?" Likert-like responses were obtained, and the average response from these two questions was used in analyses. Quality of care was also measured by explicit process criteria. Four subscales, shown in previous research to be associated with better quality of care and lower 30-day mortality for congestive heart failure and pneumonia, were rated: the physician cognitive scale, the nurse cognitive scale, the technical diagnostic scale, and the technical therapeutic scale. The ratio of the number of criteria met on these scales to the number of applicable criteria was calculated for each patient. Both the implicit and explicit measures were standardized.

In unadjusted analyses for patients with congestive heart failure, black patients had a lower overall quality of care by implicit review (-0.13 standardized units, p=0.05) but not by explicit criteria (-0.01 standardized units, p=0.86). For patients with pneumonia, black patients had a lower overall quality of care by explicit criteria (-0.15 standardized units, p=0.04) but not by implicit review (-0.11 standardized units, p=0.16). In adjusted analyses, black patients received lower quality of care by both measures for both congestive heart failure and pneumonia. In adjusted analyses for congestive heart failure, there were no significant racial differences on any individual subscale of explicit criteria. For pneumonia, blacks had lower ratings on two of the subscales of explicit criteria – the technical diagnostic and technical therapeutic subscales. (Blacks were less likely to have blood cultures collected on the first or second hospital day and non-immunocompromised black patients were less likely to receive antibiotics within six hours of admission.) The deficits in overall implicit quality were greater in non-teaching hospitals than in teaching hospitals for blacks than non-blacks with congestive heart failure (-0.59 standardized units, p<0.001).

In unadjusted analyses for patients with congestive heart failure, women had a lower overall quality of care by implicit review (-0.13 standardized units, p=0.04) but not by explicit criteria. There were no gender differences for pneumonia. The results were the same in adjusted analyses. Examining the subscales separately, the adjusted rating of the physicians' cognitive subscale was lower for women. (For example, among patients with congestive heart failure, physicians were less likely to record physical findings from the initial cardiac examination.) Ratings of technical therapeutic care were higher for women than men with congestive heart failure, and ratings of nurses' cognitive care were higher for women than men with pneumonia. Deficits in overall implicit quality were greater in non-teaching hospitals than in teaching hospitals for women than men with congestive heart failure (-0.29 standardized units, p=0.02).

The authors conclude that quality of care was lower for blacks than whites both among those with congestive heart failure and among those with pneumonia. Because the services measured in this study are noninvasive and entail little or no risk, the authors rejected the possibility that the observed differences were due to patient preferences. The data also showed that they were not due to socioeconomic factors.

A more complex picture emerged with regard to gender differences. Women with congestive heart failure had lower quality of care based on implicit review only. To understand this better, the authors examined the specific subscales of explicit review, which showed that physicians might have provided more thorough evaluations for men and nurses might have provided more complete assessments for women. The authors suggested that, because physicians provided the implicit reviews, they might have focused on physician behavior when making ratings, thus explaining the incongruous results from the two measures. However, the technical therapeutic subscale rating was higher for women than men. Overall, the authors conclude that "future studies should be tailored to specific conditions and types of services to gain a more complex comparative understanding of how women and men are treated."

The authors note several limitations, particularly the lack of measures on functional outcomes and survival, which would be useful in validating the quality of care measures used in this study. However, they note that, based on previous research, the differences in overall explicit quality observed by race would be associated with adjusted 30-day mortality rates approximately 0.5% higher for blacks than non-blacks with congestive heart failure or pneumonia.