MODELING AGE-AT-DEATH IN ADAMAWA STATE, NIGERIA, USING PROBABILITY DISTRIBUTION FITS

Abstract

High mortality rates and low life expectancy remain critical public health challenges in Nigeria, yet regional variations in mortality patterns are poorly understood. This study models age at death as a proxy for life expectancy in Adamawa State, Nigeria, using probability distribution fitting approaches. Drawing on secondary data from the Specialist Hospital Yola, the study analyses age-at-death records over 61 years. Four continuous lifetime distributions—Log-normal, Log-logistic, Generalized Extreme Value (GEV), and Generalized Pareto—were evaluated for best fit using Kolmogorov-Smirnov, Anderson-Darling, and Chi-square tests. Results indicate that age-at-death is right-skewed and multimodal, with an average of 26.89 years, far below Nigeria's national average. Log-normal and GEV distributions provided the best fits among the fitted models based on multiple goodness-of-fit criteria. Regression analysis using these distributions revealed that age-at-death significantly varied by cause of death and local government area, but not by gender. The findings highlight the importance of region-specific health policy interventions and support the use of lifetime distributions in modelling mortality patterns for public health planning.