Structured additive distributional zero augmented beta regression modeling of mortality in Nigeria

Area of Science:

• Public Health
• Demography
• Biostatistics

Background:

• Child mortality remains high in sub-Saharan Africa, with limited analysis considering exposure duration or effects on mortality distribution beyond the mean.
• Existing studies often link covariates to the mean response, potentially overlooking associations with higher moments of the mortality distribution.

Purpose of the Study:

• To analyze child mortality determinants in Nigeria by incorporating the duration of exposure using a child mortality index.
• To examine covariate effects on the probability of no child mortality, the conditional expectation of mortality, and mortality spread.
• To control for latent spatial associations in child mortality patterns.

Main Methods:

• Utilized the 2013 Nigeria Demographic and Health Survey data.
• Defined a child mortality index as the ratio of observed to expected child deaths.
• Employed a structured additive distributional beta regression model within a Bayesian inference framework using MCMC tools.

Main Results:

• Significant variations in the probability of experiencing no child mortality and in mortality spread were observed across Nigerian states.
• Child mortality is more concentrated among younger women compared to women aged 30 years and above, where it is more spread.

Conclusions:

• The duration of exposure to mortality risks and spatial factors are crucial in understanding child mortality patterns in Nigeria.
• Age is a significant factor, with younger women facing concentrated child mortality risks, highlighting the need for targeted interventions.