Survival analysis of acute myocardial infarction patients using non-parametric and parametric approaches

Abstract

In this paper, an investigation about the survival pattern of acute myocardial infarction patients was explored, using non-parametric and parametric modeling strategies. Median survival time and their associated confidence intervals are often used to summarize the survival pattern of a group of patients in clinical data with failure- time end points. Although there is an extensive literature on this topic but to the best of our knowledge no study exists which compare the survival pattern of AMI patients using the non-parametric and parametric modeling strategies. Life table estimation of cumulative survival function of AMI patients stratified by age and marital status show a poor prognosis for older and married patients respectively. The estimated median survival time of overall AMI patients by clinical life table method is 3.31(95% confidence interval, 2.80-3.82) years. Probability plotting and Anderson-Darling goodness of fit test were used to compare the theoretical distributions viz. Weibull and Gamma distributions. Among these, the Weibull distribution is found to be the best fit to the observed data. The median survival time of overall AMI patients using Weibull distribution is 2.45(95% confidence interval, 1.87-3.03) years. Knowledge of median survival time will enable the physicians to develop strategies for preventive cardiology so that survival time of new AMI patients could be improved. Also it can be considered as a baseline for further studies.