Modelling of Malaria Incidence data: Comparison of Shared frailty Model and Count Regression model

Abstract

Malaria is a major health challenge in most of the developing countries leading to high morbidity and mortality rates in the population. Seasonal variations, development of water projects and their operation have some long history of facilitating increased transmission of vector borne diseases. The study was motivated by the availability of the time-to-event dataset, which contains data on time-to-first P.falciparum malaria infection. The main interest of the study is to investigate the influence of seasons and distance to the Gilgel-Gibe hydroelectric dam reservoir on the time-to-first malaria infection in children aged less than 10 years. To incorporate the seasonal variation, a more flexible proportional hazards model was fitted by making some mild assumptions concerning the baseline hazard. The resulting piecewise constant hazards model is equivalent to a count regression model with aggregated event counts as response variable and the log of exposure time as an offset. However, due