Objective: To assess criteria used for detecting underuse of coronary artery revascularization procedures in terms of patient outcomes.
Design: Retrospective cohort study using medical records supplemented by a telephone survey and review of county death records.
Setting: Four public hospitals and two academically affiliated private hospitals in Los Angeles County, California.
Participants: A total of 671 patients who had coronary angiography between June 1, 1990 and September 30, 1991, and who met explicit clinical criteria for the necessity of coronary artery bypass graft (CABG) surgery or percutaneous transluminal coronary angioplasty (PTCA).
Main outcome measures: For all patients (n = 671), we estimated the association between receipt of necessary revascularization and mortality (median follow-up after angiography, 797 days) after adjusting for potential confounders. For the patients completing the telephone interview (n = 374), we examined the relationship between receipt of necessary revascularization and frequency of chest pain.
Results: Patients who received necessary revascularization within 1 year of angiography had lower mortality than those who did not (8.7% vs 15.8%, P = .01), and this association persisted after adjustment for extent of coronary artery disease, clinical symptom complex, ejection fraction, and cardiac surgical risk index (adjusted odds ratio = 0.49; 95% confidence interval, 0.28 to 0.86). The same general results were obtained whether revascularization was received within 1 year or within 30 days of the catheterization, whether panelists' ratings or individual clinical variables were entered as covariates, and whether the statistical procedure used was logistic regression or Cox proportional hazards analysis. In addition, among patients responding to the telephone survey, those receiving necessary revascularization had less chest pain at follow-up (P = .03).
Conclusions: Among patients meeting criteria for the necessity of revascularization, those receiving a revascularization procedure within 1 year had lower mortality than those treated medically. These results support the validity of the RAND/UCLA criteria for detecting underuse of these procedures, but more research is needed to confirm the findings and to determine the validity of guidelines for other procedures.