Survival Modeling of Infant Mortality. Evidence from Nandom District

摘要

Survival analysis is employed in estimating the survival proportions of infants born in the Nandom district, Ghana and the infant mortality count data fitted to four standard loss distributions. The four loss distributions were Poisson, Binomial, Negative Binomial and Geometric distributions. The results showed that, the first twenty-four (24) hours and month after delivery are the most critical periods in the life of an infant child in the Nandom District. Also, about 98 percent survived up to the neonatal stage and only 97 percent made it to the post neonatal stage. The conditional proportion of infants dying within the early neonatal stage recorded the highest proportion of about 2.17 percent. The Negative Binomial Distribution was the best model to fit the infant mortality data with log-likelihood and A.I.C values as -31.7365 and 65.4730 respectively with an expectation of about 51 deaths per year. This implies that, the district is still facing a bigger challenge in terms of infant mortality and should therefore work towards achieving the goal as stipulated in the SDGs.