The objective of this study was to identify factors related to use of cardiac catheterization, percutaneous transluminal coronary angioplasty (PCTA), and coronary-artery bypass surgery (CABG) among veterans with coronary artery disease hospitalized in Veterans Administration (VA) medical centers. The authors hypothesized that, because patients using the VA system have access to an equal package of services either at their local facility or through established referral networks, differences in the capabilities of the local VA medical centers to perform cardiac catheterization, PTCA, and CABG surgery would not significantly affect use of these procedures among patients with coronary artery disease.

Data were derived from the Department of Veterans Affairs Patient Treatment Files for fiscal year 1994. Patients with a primary diagnosis of coronary artery disease were included in this study. Control variables included sex, age, race, type of coronary artery disease (acute versus chronic), and a co-morbidity index (with up to ten co-morbid diseases). The main independent variables included geographic region of the patient’s home (based on primary service areas), size of this region, size of the local VA medical center, measure of hospital complexity (metropolitan area, hospital size, and affiliation with a medical school in six groups), and availability of a cardiac catheterization laboratory or surgical program.

Over half of the patients (55.8%) with coronary artery disease in the VA system underwent cardiac catheterization, 11.1% had PTCA, and 13.9% underwent CABG surgery. Patients undergoing each of these procedures were significantly more likely to be white than black (as well as be younger, be male (except PTCA), have acute coronary disease, and have a lower co-morbidity index).

After controlling for other clinical and demographic variables, white patients had a significantly increased odds compared with black patients of having each procedure. (For cardiac catheterization, OR=1.42, 95% confidence interval=1.32-1.52; for PTCA, OR=1.44, 95% confidence interval=1.27-1.63; and for CABG, OR=1.96, 95% confidence interval=1.73, 2.22). The availability of services, complexity of the hospital, and geographic region were also all significantly associated with having these procedures, and the effect of racial group increased after these hospital system variables were added to the models.

An important limitation to the study is that there were no detailed clinical data to determine the appropriateness of care. Additionally, as in many studies, the author relied on use of services to assess access to care.

The authors conclude that the “intent to provide equal access or usage in policy development is insufficient to ensure that it does, in reality, occur.” Furthermore, since the study design aimed to evaluate hospital system attributes, the authors also conclude that these variables were more important than clinical and demographic variables because the addition of system variables to models that only included clinical and demographic variables substantially improved the models’ discriminating ability.

Comment

Every NR, Ritchie JL.
Variations in the use of cardiac procedures: what is the explanation?
Am Heart J 1999;137(4 Pt 1):588-9.

This editorial was written as an introduction to the Mirvis and Graney study (Variations in the use of cardiac procedures in the Veterans Health Administration, American Heart Journal 1999; 137:706-13). The main finding reported in this study was a lower rate of cardiac procedure use among patients admitted to VA hospitals that did not have onsite catheterization facilities compared with patients admitted to hospitals that did have such facilities. This letter notes an important limitation in interpreting this study: the analyzed data was obtained from only the VA system, which leaves open the possibility that VA hospitals without the necessary facilities might regularly refer cardiac patients to local non-VA hospitals for catheterization. If so, the study findings would be biased by not taking these referrals into account. Nonetheless, the authors note that the distance of tertiary centers from some primary referring facilities in the VA system suggests that just because some patients had access to care did not ensure they had access to necessary procedures.

The authors argue that several improvements in the health care system might reduce the variation in use of procedures, and these improvements might be more manageably tested in the VA system than in the private sector. Such improvements might include nationally uniform electronic patient records for improving communication between primary and tertiary medical centers and the incorporation of clinical guidelines into the point of care in patient management. They conclude that "the cardiology and research communities must now define what the appropriate level of procedure use should be and take the steps necessary to achieve this level."