Statistical Analysis of Infant Mortality Rates in Rivers State Using Regression Models

Infant mortality remains a critical public health indicator in Rivers State, situated in the South South geopolitical zone of Nigeria, where access to quality maternal and neonatal care is unevenly distributed across urban and rural communities. This study employs multiple linear regression and Poisson regression models to identify statistically significant determinants of infant mortality rates across the 23 local government areas of Rivers State. Secondary data were sourced from the Rivers State Ministry of Health and the National Population Commission spanning 2019 to 2023. Predictor variables include maternal education level, distance to the nearest health facility, antenatal care attendance rate, skilled birth attendance proportion, and household income quintile. Diagnostic tests for multicollinearity, heteroscedasticity, and residual normality are conducted to validate model assumptions. The Poisson regression model outperforms the linear regression alternative in terms of Akaike Information Criterion, confirming its suitability for count outcome data. Results indicate that skilled birth attendance and antenatal care attendance are the strongest negative predictors of infant mortality, while distance to health facility is the most significant positive predictor. Local government areas in the riverine districts exhibit mortality rates 1.7 times higher than upland LGAs after controlling for socioeconomic covariates. The study recommends targeted deployment of mobile health units and community midwifery programmes in riverine communities. Keywords: infant mortality, regression analysis, Rivers State, public health, Poisson model.