Time sereis analysis of under-five mortality in narok district hospital. Employing autoregressive integrated moving average (arima) model.

Abstract

This research work undertook time series analysis of under-five mortality in Narok District hospital for the period of 2003-2015 with objectives of:finding the extend at which each of the factors contributes to under-five mortality that is ranking the factors according to how it contributes to under-five mortality, establishing ifthere is trend in the mortality series from 2003-2015, investigating presence of seasonality in the mortality series, to analyze mortality differences in terms of sex, cause and period, to find out possible ways of reducing mortality and lastly to make predictions of under-five mortality for the period of 2016-2018. The study explores the application of time series autoregressive integrated moving average (ARIMA) in the study of under-five mortality. Analysis was based on data available at the Kenya Health information System for the period2003-2015. The data was directly entered into excel and imported to MINTAB, GRETL and EVIEWS for analysis. The study showed that malaria accounted for most of the deaths followed by Pneumonia and Diarrhea. Genital infection and oral disease accounted for the least number of deaths recorded for the period of 2003-2015. Augmented Dickey-Fuller Test also revealed that the mortality series was not stationary in levels for the recorded period of 2003-2015. Mortality series was also found to vary by gender, period and sex, where the male deaths were higher than the female deaths. Run’s test also revealed that the mortality series did not exhibit any trend over the period of study. Seasonality was also found to exist in the mortality series where most of the deaths were recorded in the month of June, February, December, July and August and the least in January and October.