Identifying Some Risk Factors for the Time to Death of the Elderly Using the Semi-Parametric Blended Model of Survival Analysis With Competing Risks

Objectives As the population of elderly people in Iran is rising, determining the risk factors of their death is necessary. The purpose of this study was to identify the risk factors that reduce the survival time of elderly people.
Methods & Materials In a longitudinal retrospective study, data of 510 elderly people aged over 60 years, who were admitted to Kashan’s Golabchi nursing home from 2000 to 2012 were collected and analyzed. To identify some risk factors of time to death in elderly, semi-parametric mixture competing risk model in survival analysis was fitted to the data. To estimate the model parameters, Expand-Maximize-Compress (EMC) algorithm was used and parameters and their 95% confidence intervals were estimated using R software (version 3.3.1).
Results In separate one-variable fitted models, the variables like high blood lipids (HR=1.04; CI =1.00, 1.31), history of myocardial infarction (HR=0.90; CI=1.04, 1.10), stroke history (HR=0.95; CI=1.00, 1.14), and deaths of elderly people with cardiovascular diseases were significant. In the fitted multivariate model, renal problems had a significant effect (HR=1.58; CI=1.77, 2.83) on time to death of elderly.
Conclusion In single-variable fitting, age, history of myocardial infarction, history of stroke, and kidney problems were identified to have significant effects on the time to death of the elderly. Based on one-variable semi-parametric competing risk mixture fitted models, more significant risk factors for the time to death of elderly was identified when compared with a fitted multivariate mode to the data. This implies that the role of some independent variables can be explained by other independent variables.