Geospatial Variations and Predictors of Low Birth Weight in Sub-Saharan Africa: Further analysis using evidence from Demographic Health Survey 2015-2024.

Abstract

Introduction: Low birth weight (LBW), defined as less than 2.5 kg (5.5 lbs) at birth, remains a critical global public health challenge. It significantly increases the risk of neonatal mortality and immediate complications such as sepsis and hypothermia, along with lifelong consequences including childhood disabilities and adult-onset chronic diseases. However, there was a limited study that described the spatial distribution and predictors of low birth weight in sub-Saharan Africa. Objective: To assess geospatial variation of low birth weight and associated factors in Sub-Saharan Africa (SSA) using evidence from 2015 to 2024 Demographic and Health Survey (DHS) data. Method: A community-based cross-sectional study design based on DHS (2015-2024) data, comprising a weighted sample of 138,164 women aged 15–49 years with live births among 28 sub-Saharan African countries, was included in the study. The analysis was performed by STATA 17, GeoDa 1.14, SatScan 9.7, and ArcGIS Pro 3.1. Global Moran’s I was calculated to determine overall clustering of low birth weight. Statistically significant hot spot and cold spot areas of low birth weight were determined by Getis-Ord G* statistics. Ordinary least squares, spatial lag, spatial error, geographically weighted regression, and multiscale geographically weighted regressions were utilized to determine predictors of low birth weight. The best-fitting models were determined by the highest R² and the lowest AICc values. Finally, the statistically significant predictors from the final model were displayed on a map. Results: Low birth was clustered (Moran’s I 0.23, z-score 50.2, p-value < 0.01) in the study area. Significant hotspot areas were depicted in Mauritania, Mali, Senegal, Burkina Faso, Nigeria, Gabon, Angola, Madagascar, South Africa, Lesotho, Malawi, and Ethiopia. Conversely, low-risk cold spots were observed in Uganda, Kenya, Rwanda, Burundi, Tanzania, Zambia, Zimbabwe, Cameroon, and Sierra Leone. Short birth interval, no visit to a health facility in the last year, twin birth, no media exposure, and unemployed women were significant predictors of low birth weight. Conclusion and Recommendation: There is spatial variation of low birth weight across different regions in SSA. Short birth interval, no visit to a health facility in the last year, twin birth, no media exposure, and unemployed women were significant predictors of low birth weight. Targeted maternal health interventions, improved healthcare access, health education using mass media, and economic empowerment for women are recommended to reduce low birth weight.