Several studies have documented racial differences in medical services; however, two limitations affect the ability to draw definitive conclusions from these studies: first, many studies have not adequately controlled for potential differences between blacks and whites with regard to illness prevalence, and, second, studies that have considered current illness have not adequately controlled for other potential confounders.

This study examined black-white differences in the use of intensive treatments for six different conditions – angina, nonangina coronary heart disease (CHD), stroke, hip fracture, colon cancer, and breast cancer – within an economically homogeneous group (all were Medicare enrollees and drawn from a sample of beneficiaries matched on residential zip code, and eligibility and median county income were controlled). The conceptual model offered by the authors was as follows: the probability that a service/procedure will be offered is a function of disease prevalence, severity, ability to pay, probability of treatment seeking, compliance, and provider practice patterns. The authors hypothesized that racial differences in procedure use are attributable to the other variables in the model.

Data for this study were drawn from the 1989 Part A and Part B Medicare claims data for 10 states and the District of Columbia. A sample was constructed that matched equal numbers of whites and blacks based on zip code of residence, age, gender, and Medicaid eligibility. Additionally, the 1990 American Hospital Association Survey of Hospitals was used to determine the characteristics of the hospital in which each index hospital admission occurred, and Census data were used to determine county-level income and race distributions.

For breast cancer, colon cancer, and hip fracture patients, after adjusting for relevant factors, there were no black-white differences in any procedure examined. However, for heart disease and stroke, substantial racial disparities persisted after controlling for age, gender, Medicaid eligibility, comorbid illnesses, hospital characteristics, county characteristics, region, and distance traveled to hospital. The largest racial difference was found in stroke procedures: the white/black odds ratio (OR) for endarterectomy was 5.26 for patients at age 65 years and 3.35 for patients at 75 years; for cerebral angiography, the OR’s were 3.36 and 2.42; and, for noninvasive imaging procedures, 1.48 and ~1.00. CT of the brain was performed significantly more often among blacks at age 65, and there were no differences in magnetic resonance imaging (MRI) of the brain. With regard to coronary vascular disease (CVD) procedures, the odds for whites receiving coronary artery bypass grafting (CABG) surgery was 1.66 times that of blacks at age 65 and 2.51 at age 75. For angiography, the odds for whites was 1.44 that of blacks at age 65, and not significantly greater at age 75.

Further analyses showed that the racial disparity in angiography for angina patients diminished with age and that racial differences in CABG surgery increased with age. Gender differences were also found: older women were less likely to undergo angiography than older men, and all but the oldest women were less likely to undergo CABG surgery than their male counterparts. Comorbidity, hospital type, geographic region and distance to hospital all had a significant effect on CABG use as well.

Finally, racial differences were found in the total Medicare allowed charged during the interval 90 days before and after an index hospitalization. For angina patients, $568 less (adjusted for covariates) was billed for black versus white patients, and for nonangina CAD patients $469 less was billed for black patients.

The authors review possible reasons for these racial patterns in procedure use for CAD and stroke disease, including: blacks may have less extensive disease, blacks may have a lower propensity to undergo recommended surgery, and whites may be treated excessively. Although this study reported no (adjusted) racial disparities in procedures for colon and breast cancer and hip fracture, the authors note that, when the analyses were not conditioned on an episode of illness, these procedures were performed more frequently among whites than blacks. The authors conclude that it is likely that blacks are less likely to be screened for cancer, and, thus, still receive inadequate treatment, despite the results previously reported in this study. With regard to hip fracture, the authors concluded that the previously reported differences in procedures for hip fracture are likely to be due to racial differences in incidence.