This study aimed to assess whether there is racial variation in the use of cardiac catheterization, percutaneous transluminal angioplasty (PCTA), and coronary artery bypass grafting (CABG) (as reported in previous studies) and to determine whether these treatment differences affect survival.

The study, conducted at a Veterans Affairs Medical Center, included male inpatients over 30 years of age who were discharged in 1993 with a primary diagnosis of cardiovascular disease or chest pain. Only blacks and whites were included. Patients were excluded if they had had a previous cardiac procedure within 90 days; repeat admissions within the study period were discounted. This cohort consisted of 1,043 men.

African Americans were less likely than whites to have cardiac catheterization performed during or within 60 days of the index hospitalization (unadjusted OR for blacks versus whites =0.37; 95% confidence interval=0.24 to 0.58), to have CABG within 90 days of the index admission (unadjusted OR=0.22; 95% confidence interval=0.08 to 0.63), and to have any cardiac procedure (unadjusted OR=0.32; 95% confidence interval=0.21-0.50). The difference in use of PTCA was not statistically significant. There were several demographic and clinical differences between blacks and whites in this cohort. Blacks were less likely to be married; less likely to have non-service VA eligibility; more likely to be county residents; more likely to have diagnoses of lung cancer, schizophrenia, dementia, renal disease, and diabetes mellitus; and had more severe disease than whites.

Multivariate analyses were not conducted, but comparisons between patients who received any procedure and those who did not showed that patients who received any procedure were slightly younger, more likely to be married, more likely to have non-service eligibility for the VA, and more likely to live outside of the county. There were significant differences in clinical characteristics as well. Patients were more likely to receive invasive cardiac procedures if they did not have comorbid diagnoses, were discharged from the hospital with a plan to later receive invasive procedures, and had lower disease severity.

Blacks and whites had nearly equivalent 30-day and 1-year survival rates (95% versus 96% for 30-day and 84% versus 88% for 1-year).

The authors were also interested in evaluating whether selection issues influenced previous studies. Thus, they assembled two additional cohorts. First, they examined the 511 patients who resided within the county in order to exclude those patients referred from elsewhere for elective surgery. Second, they examined the 497 patients who met narrower diagnostic criteria (acute myocardial infarction, unstable angina, angina, or chronic ischemia). Racial patterns in all of the above analyses were similar, although smaller in magnitude, using these two additional cohorts. Most comparisons did not reach statistical significance, perhaps due in part to the reduced sample size.

It is important to note that drawing conclusions from this study is limited by the lack of multivariate analyses (which were not included because of the small sample size).

The authors made note of several new findings in this study. First, patients residing outside of the county were more likely to receive cardiac procedures, and African Americans were less likely to reside outside of the county. A plausible explanation for this finding is that rural populations, which are predominantly white, sought medical care for early and acute symptoms of cardiac disease and were then referred for more invasive procedures to the study VA center. Thus, whites who were judged to have less severe cardiac disease (and were therefore inappropriate for cardiac procedures) and who were not referred to the VA study were not included in the present sample. Previous studies may have been similarly influenced by this selection issue. (The authors argue that the restricted sample, which included only local residents, also supported this claim, since the effect of race for catheterization reduced from OR=0.37 to OR=0.47 and the effect of race for CABG reduced from OR=0.22 to OR=0.30. They also argue that the racial difference in PTCA use was reversed. However, even though the OR for the full sample was 0.60 and for the local sample was 1.11, the 95% confidence intervals were overlapping and therefore might not indicate a reversal of effect.)

Second, refining the cohort to include only those with more narrowly defined ischemic heart disease also reduced the racial differences. The authors suggest that by limiting the cohort, they may also have eliminated the confounding conditions that predispose some patients to having invasive cardiac procedures. Further exploration of this is necessary.

Third, there were no race differences in survival. Thus observed treatment differences did not translate into clinically significant poorer survival among blacks. "Taken together, these results serve to remind us that we must not rely on process of care measures as surrogates for quality of care measures."

The authors conclude that potential explanatory factors for treatment disparity include the presence of more severe comorbid disease among blacks and limitations of study designs that have not controlled for referral status and specific cardiac diagnoses. These findings provide hypotheses that must be tested in future investigations."