BAYESIAN SURVIVAL ANALYSIS OF DIABETES MELLITUSIN LAGOS STATE NIGERIA

Abstract

Diabetes Mellitus is an escalating public health issue in Nigeria, significantly affecting the population. This study analyzes the survival patterns of diabetes patients in Lagos State, Nigeria, using Bayesian survival analysis techniques, including the Cox Proportional Hazard model and Kaplan-Meier estimator, to identify key factors influencing survival rates and predict future outcomes. Drawing from a comprehensive, multi-year dataset that includes critical predictors and confounders, the study reveals a significant decline in survival probability over time. Posterior predictive checks confirm the models’ adequacy, showing strong alignment between observed data and simulations. The Cox Proportional Hazard model identifies age, gender, and insulin use as key contributors to the hazard rate, while other factors are found to have limited impact. These findings underscore the importance of early intervention strategies targeting high-risk factors, such as age and insulin dependency, to improve outcomes and reduce the diabetes burden in Lagos State, Nigeria.