Use Of The Shared Frailty Model To Identify The Determinants Of Infant And Child Survival In Rwanda Using DHS Data

Abstract

The infant and child mortality have much fluctuated in Rwanda since the year 1992 and decreased considerably during 5 years preceding the 2005 RDHS. However, the current indicators are still high compared to the goals to achieve and current global standards while factors that have contributed to the observed achievements remain not well known and diverse resources invested in improving the children's health conditions keep becoming more scarce. Therefore, it is very important to ascertain the key determinants of infant and child survival in Rwanda and hence mark out the sectors that should be prioritized in order to realize more improvements of under age five children's survival with limited resources. The analysis of semi-parametric gamma shared frailty models utilized to analyze separately infant and child mortality in Rwanda revealed that frailty effects were significant in childhood while its effects in infancy were not significant. The causes of deaths in infancy and childhood were quite different. The hazard rates and hazard ratios indicated that infant deaths were strongly related to proximate and biological factors including place of delivery, preceding birth interval and mother's age at child's birth while child deaths were at a very large extent due to socioeconomic and demographic factors such as household socioeconomic status being the most important determinant and province of residence the least. Key socioeconomic determinants of infant mortality were mother partner's education and household's socioeconomic status. Multiple births were at higher risk of-death than others both in infancy and childhood but the risk decreases with age.