Socio-economic, demographic and environmental determinants of under-five child mortality in Uganda: Analysis of trends and patterns (1995-2011)

Abstract

This study was aimed at identifying and documenting key factors that may be responsible for mortality amongst Ugandan children under five years using data from the Ugandan Demographic and Health Survey (UDHS) (UBOS and ICF International Inc 2012) for the years 1995, 2000, 2006 and 2011. The outcome variable took into consideration the children’s survival status as well as the month of their death or the last month they were known to be alive. The variables used in the study include: Educational level, type of place of residence (urban/rural), wealth index (quintile), mother’s age at child birth, mother’s current age, birth order, sex of the child, Preceding Birth Interval, Type of birth, region, religion, Floor type, household size, Altitude, Cooking fuel, Water Source, Roof material and Altitude. Survival data analysis approach was adopted across the study period. At bivariate level of analysis, associations were tested using the Log-Rank Chi-square and only factors with a P-value of 0.25 or less were included in the multivariate Cox regression model. Also, graphs of survivorship were adopted in establishing variations across groups using Kaplan Mier graphs. The results confirmed that all factors except mother’s age at first birth were eligible for further analysis at the multivariable stage (p < 0.25). Estimates from the Cox proportional hazard model showed a significant relationship between child’s sex, Toilet facility, child birth order, Preceding birth interval, Type of birth, mother’s age at first birth and household size with under-five child mortality (p<0.05) for all or some of the periods. Six of the eight hypotheses were partially supported as significant relationships were observed in some periods of the study. These were: age at first birth of the woman, household size, preceding birth interval, sex of the child, child birth order, and type of birth are consistent predictors of child mortality.