Abstract
In Zambia, neonatal mortality still remains a noteworthy public health problem with a current rate of 27 deaths per 1000 and ranking 162 out of 195 countries globally. The study aimed to investigate the spatial variations and predictors of neonatal mortality in rural Zambia among HIV-infected and non-infected mothers using the national-level data from the 2018 Zambia Demographic and Health Survey (ZDHS). Statistical analyses were conducted using the Rao - Scott Chi-square test to assess associations between neonatal mortality and categorical variables. Additionally, a multilevel mixed effect logistic regression model was used to examine predictors of neonatal mortality. Geospatial variations of neonatal mortality across Zambia's ten provinces were mapped using Quantum Geographical Information System (QGIS) version 3.34.1. Data analysis was performed using R and Stata version 14.2. This study examined the spatial variations and predictors of neonatal mortality among HIV-infected and non-infected mothers in rural Zambia using the 2018 Zambia Demographic Health Survey dataset. Key findings include the protective role of maternal education, with those having secondary or higher education showing reduced odds of neonatal mortality. Women aged 20-24 years in the study had higher odds of neonatal death compared to younger mothers, while delivering in public health facilities was associated with increased neonatal mortality. Maternal HIV status had no significant impact on neonatal outcomes. Spatial analysis revealed significant regional disparities, with high mortality rates in Central, Southern, and Eastern provinces, while North Western Province had lower rates. These results emphasize the need for improved healthcare quality, targeted maternal education programs, and region-specific interventions to address neonatal mortality in Zambia.