Community factors, hospital characteristics and inter-regional outcome variations following acute myocardial infarction in Canada.

BACKGROUND: While various community and hospital characteristics have been demonstrated to have an impact on individual cardiovascular outcomes, the extent to which such factors account for inter-regional and interhospital outcome variations following acute myocardial infarction (AMI) remains unknown.
OBJECTIVES: To examine the impact of community and hospital factors on individual AMI outcomes and procedure use, and to determine the extent to which such characteristics account for inter-regional and interinstitutional AMI outcome and procedure variations across Canada.
METHODS: Patients hospitalized with AMI between April 1, 1997, and March 31, 2000, across Canada were examined. The community and hospital characteristics studied included three indicators of socioeconomic status, two indicators of ethnicity, rural-urban status of residence, hospital academic affiliation, and the presence or absence of on-site angiography or revascularization capabilities at the admitting institution. Outcomes included in-hospital mortality, one-year cardiac readmissions and 30-day revascularization rates post-AMI. All analyses were adjusted for age, sex and age-sex interaction. The relationships between community/hospital factors and individual outcomes were examined using random-effects hierarchical logistic regression analysis, while the relationships between community/hospital characteristics and inter-regional/hospital risk-adjusted outcomes were examined using least squares regression and the coefficient of determination (r2).
RESULTS: After adjusting for demographic factors, a patient's neighbourhood socioeconomic status was inversely correlated with the likelihood of death and downstream cardiac readmissions (P<0.001); patients residing in lower educated regions were less likely to receive revascularization post-AMI (P<0.001). Patients living in regions with higher concentrations of new immigrants and/or visible minorities, as well as those admitted to academically affiliated hospitals or hospitals with on-site procedural capacity, had fewer cardiac readmissions (P<0.001) and greater use of revascularization post-AMI (P<0.001) after adjusting for age and sex. Despite their associations with outcomes on an individual patient level, community and hospital factors explained no more than 7% of the variation in the risk-adjusted outcomes across hospitals or regions. Finally, adjustments for community and hospital factors and procedure use, beyond adjustments for age and sex alone, had marginal impact on a province's risk-adjusted outcomes.
CONCLUSIONS: While community and hospital factors are important determinants of individual outcomes after AMI, they account for only a minimal degree of outcome variation across regions. Further studies are required to examine whether AMI outcome variations in Canada are explained by differences in patient clinical profiles and/or by differences in the decision-making behaviours of providers across jurisdictions.