Predictive Factors of Hospital Mortality Due to Myocardial Infarction: A Multilevel Analysis of Iran’s National Data

Abstract

Background: Regarding failure to establish the statistical presuppositions for analysis of the data by conventional approaches, hierarchical structure of the data as well as the effect of higher‑level variables, this study was conducted to determine the factors independently associated with hospital mortality due to myocardial infarction (MI) in Iran using a multilevel analysis.

Methods: This study was a national, hospital‑based, and cross‑sectional study. In this study, the data of 20750 new MI patients between April, 2012 and March, 2013 in Iran were used. The hospital mortality due to MI was considered as the dependent variable. The demographic data, clinical and behavioral risk factors at the individual level and environmental data were gathered. Multilevel logistic regression models with Stata software were used to analyze the data.

Results: Within 1‑year of study, the frequency  (%) of hospital mortality within 30  days of admission was derived 2511 (12.1%) patients. The adjusted odds ratio (OR) of mortality with (95% confidence interval [CI]) was derived 2.07 (95% CI: 1.5–2.8) for right bundle branch block, 1.5 (95% CI: 1.3–1.7) for ST‑segment elevation MI, 1.3 (95% CI: 1.1–1.4) for female gender, and 1.2 (95% CI: 1.1–1.3) for humidity, all of which were considered as risk factors of mortality. But, OR of mortality was 0.7 for precipitation (95% CI: 0.7–0.8) and 0.5 for angioplasty (95% CI: 0.4–0.6) were considered as protective factors of mortality.

Conclusions: Individual risk factors had independent effects on the hospital mortality due to MI. Variables in the province level had no significant effect on the outcome of MI. Increasing access and quality to treatment could reduce the mortality due to MI.

Keywords: Mortality, multilevel analysis, myocardial infarction