Modeling Time to Death of Patients with Ischemic Heart Diseases at Jimma University Medical Center: A Comparison of Parametric Shared Frailty Models

Abstract

Background: Ischemic heart disease (IHD) is a disorder of cardiac function caused by insu cient blood ow to the muscle tissue of the heart. The decreased blood ow is in most cases due to coronary arteriosclerosis or to obstruction by a thrombus of the coronary arteries. Acute myocardial infarction, unstable angina, and angina pectoris are manifestations of ischemic heart disease. Objective: The general objective of this study is to model the time to death of patients with ischemic heart disease using various parametric shared frailty models. Methods: Di erent parametric frailty models were compared using exponential, weibull, and log-logistic as baseline hazard functions and the gamma as well as the inverse Gaussian for the frailty distributions, with the goal of developing an appropriate survival model that adequately describes the ischemic heart disease dataset. All models were then compared using the AIC and BIC criteria. Results: The median time to death of the ischemic heart disease patients was about ve days, with a maximum death time of thirty days, of which about 35.37% died. The clustering e ect is signi cant in modeling the time to death of ischemic heart disease. The log-logistic model with an inverse Gaussian frailty distribution has the minimum AIC value among the models compared. According to the output of the model (loglogistic with inverse Gaussian frailty), diabetes mellitus, hypertension, obesity, smoking status, cholesterol, and other diseases were the main determinant factors of IHD. Conclusions: Compared to other distributions employed in this study, the log-logistic with inverse Gaussian frailty model provided a superior description of the ischemic heart disease dataset. The time to death of ischemic heart disease patients vary between woredas, indicating that frailty models must be used to take into account this clustering feature.