Bayesian hierarchical modeling of infant mortality in Nigeria

Abstract

Infant mortality and its risk factors in Nigeria was investigated using Bayesian hierarchical modeling. The hierarchical nature of the problem was examined to detect the within and between groups (states and regions) variations in infant deaths. The effect of individual level variables on the risk of a child dying before the age of one was determined using data collected from the fifth round Multiple Indicator Survey (MICS5, 2016-2017). Infants in Northern Nigeria had a higher risk of dying than others, especially in North West, while South West had the lowest risk of infant deaths. Ten percent of the variations in infant deaths was explained by differences between states while differences between regions explained only seven percent of the variations. Also, factors such as urban place of residence, mothers with secondary and tertiary education, first birth and birth interval above 2 years were associated with a decreased risk of infant deaths. Male infants, birth interval of less than 2 years, mothers with primary and no education, teenage mothers and mothers that gave birth at age 35 years and above were associated with a higher risk of infant mortality.