Survival analysis of mortality data among elderly patients in University Of Ilorin Teaching Hospital Ilorin, Nigeria

Abstract

A study on the mortality among old patients 60 years or more, admitted at University of Ilorin Teaching Hospital (UITH), Ilorin was carried out using survival analysis approach. Results revealed that the median survival time, which is the time beyond which half of the patients are expected to stay in hospital before death was higher for male patients than female patients. This was also found to be higher for the patients of age less than 70 years than those who are 70 years or more. It was found that interaction effects of age and gender of patients was significant. Older female patients were more likely to die earlier than their older male counterparts. Results of Akaike Information Criterion (AIC) revealed that models with interactions provided better fits than those without interaction. Models stratified on variable weekdays on which proportional hazards assumption was violated were found to have improved model fits than those that ignored stratification. Cox models were observed to outperform the Weibull models in all the models specified in the study.

Keywords: Survival analysis, Cox Model, Weibull model, interaction, stratified model