Long-term MI outcomes at hospitals with or without on-site revascularization

Context — Many studies have found that patients with acute myocardial infarction (AMI) who are admitted to hospitals with on-site revascularization facilities have higher rates of invasive cardiac procedures and better outcomes than patients in hospitals without such facilities. Whether such differences are due to invasive procedure rates alone or to other patient, physician, and hospital characteristics is unknown.

Objective — To determine whether invasive procedural rate variations alone account for outcome differences in patients with AMI admitted to hospitals with or without on-site revascularization facilities.

Design — Retrospective, observational cohort study using linked population-based administrative data from a universal health insurance system.

Setting — One hundred ninety acute care hospitals in Ontario, 9 of which offered invasive procedures.

Patients — A total of 25 697 patients hospitalized with AMI between April 1, 1992, and December 31, 1993, of whom 2832 (11%) were in invasive hospitals.

Main Outcome Measures — Mortality, recurrent cardiac hospitalizations, and emergency department visits in the 5 years following the index admission, adjusted for patient age, sex, socioeconomic status, illness severity, and index revascularization procedures; attending physician specialty; and hospital volume, teaching status, and geographical proximity to invasive-procedure centers and compared by hospital type.

Results — Patients admitted to invasive-procedure hospitals were much more likely to undergo revascularization (11.4% vs 3.2% at other hospitals; P<.001). However, many other clinical and process-related factors differed between the 2 groups. Although mortality rates were similar between the 2 institution types, the nonfatal composite 5-year event rate (ie, recurrent cardiac hospitalization and emergency department visits) was lower for patients initially admitted to invasive-procedure hospitals (71.3% vs 80.4%; unadjusted odds ratio [OR], 0.65; 95% confidence interval [CI], 0.52-0.82; P<.001). This advantage persisted after adjustment for sociodemographic and clinical factors and procedure utilization (adjusted OR, 0.68; 95% CI, 0.53-0.89; P<.001). However, the nonfatal outcome advantages of invasive-procedure hospitals were explained by their teaching status (adjusted OR, 0.98; 95% CI, 0.73-1.30; P = .87).

Conclusions — In this sample of patients admitted with AMI, the differing outcomes of apparently similar patients treated in 2 different practice settings were explained by multiple competing factors. Researchers conducting observational studies should be cautious about attributing patient outcome differences to any single factor.