“Patients with a diagnosis of pneumonia were studied at 16 randomly selected hospitals. The services and outcomes studied include five measures of the intensity of diagnostic and therapeutic services received by patients, and death rates during hospitalization….Our findings suggest that nonwhite pneumonia patients received fewer hospital services than expected on the basis of their health characteristics, and that their hospital lengths of stay were longer than expected…No consistent differences in death rates were apparent… We conclude that patient race remains a potentially significant characteristic in determining the intensity of care provided to patients in hospitals, which is not explained by differences among racial groups in health status, source of payment, or site of hospitalization.”

“The 17 hospitals were selected by a stratified random procedure, designed to ensure a sample that varied by size, teaching status, and expenditures per patient day.… The four measures of diagnostic and therapeutic services for pneumonia were radiographic procedures, consultations, surgical procedures, and intensive care unit use. .. We adjusted our measures of service intensity or outcome for differences in patient health prior to most of our comparisons between race categories.”

“Nonwhites received fewer consultations, fewer surgical procedures, and less intensive care than expected and experienced lower levels than whites. However, the only statistically significant difference was for intensive care, with a marginally significant difference for surgical procedures. For nonwhites, length of stay was longer and mortality slightly lower, but the differences were not statistically significant for either variable.”

“Not all of the differences appear to be due to hospital, since we found some indication of racial differences within a given hospital….From our data, we conclude that patient race may still represent a significant characteristic in determining the type and amount of care provided in hospitals. In particular, we found evidence to suggest that health differences, source of payment, or site of hospitalization may not account doe all the service-intensity variation by race.
A limitation of our analysis is that we are unable to determine in detail how barriers to service intensity might work within hospitals. We can speculate that nonwhites may be more likely to receive care as ward patients rather than as private patients, through less frequent alliances with private physicians…. Alternatively, physicians may discriminate in the care they deliver and arrange additional or more extensive services for whites….Cultural differences among patients may also result in variation in patient expectations that could influence service intensity.”