Background and methods: Previous studies at individual hospitals have reported differences in the use of major diagnostic and therapeutic procedures for women and men with coronary heart disease. To assess whether these differences can be generalized, we performed retrospective analyses of coronary angiography and revascularization (coronary-artery bypass surgery or percutaneous transluminal coronary angioplasty) in women and men hospitalized for coronary heart disease in 1987, using abstract data on 49,623 discharges in Massachusetts and 33,159 discharges in Maryland. We used multiple logistic regression to estimate the adjusted odds of the use of a procedure, controlling for principal diagnosis, age, secondary diagnosis of congestive heart failure or diabetes mellitus, race, and insurance status.
Results: The adjusted odds of undergoing angiography were 28 percent and 15 percent higher for men than for women in Massachusetts and Maryland, respectively (95 percent confidence intervals for the odds ratios, 1.22 to 1.35 and 1.08 to 1.22). The respective adjusted odds of undergoing revascularization were 45 percent and 27 percent higher for men than for women (95 percent confidence intervals, 1.35 to 1.55 and 1.16 to 1.40). Because these differences could be related to differing thresholds for hospital admission, we performed a second analysis limited to patients with diagnosed acute myocardial infarction (11,865 discharges in Massachusetts and 6894 discharges in Maryland), a group in which all patients would be expected to receive hospital care. The male-to-female odds ratios in both states remained similar in magnitude and were statistically significant for angiography and revascularization.
Conclusions: These findings demonstrate that women who are hospitalized for coronary heart disease undergo fewer major diagnostic and therapeutic procedures than men. These differences may represent appropriate levels of care for men and women, but it is also possible that they reflect underuse in women or overuse in men. Further study should assess the cause of these differences and their effect on patients' outcomes.