Spatial disparities and determinants of perinatal mortality in Lesotho using multiscale geographically weighted regression

Abstract Perinatal mortality is a major public health challenge in many developing countries and has profound social, economic, and health implications for families and broader society. Despite its serious and devastating effects, there are a few studies that describe the spatial distribution and factors associated with perinatal mortality in Lesotho. Understanding the burden and geographic variation is crucial for implementing targeted interventions, allocating resources efficiently, and achieving the Sustainable Development Goals. Therefore, this study explores the spatial variations and predictors of perinatal mortality in Lesotho. This study employed a community-based cross-sectional design using data from the Lesotho Demographic and Health Survey (LDHS), comprising 2,096 weighted samples. Spatial analysis included Global Moran’s I to detect clustering, Getis-Ord Gi* to identify hotspots, and Kriging interpolation to estimate prevalence in areas without data, all conducted in ArcGIS. To identify potential factors, both Ordinary Least Squares (OLS), Geographically Weighted Regression (GWR), and Multi-scale Geographically Weighted Regression (MGWR) models were applied, with model fit evaluated using AICc and adjusted R-squared. Significant factors (p < 0.05) were mapped to show their spatial distribution. The overall prevalence of perinatal mortality in Lesotho was 46.1 per 1000 (95% CI: 35, 53), and its spatial distribution was found to be clustered. Hotspot areas were located in the western parts of Leribe, Berea, and Maseru, as well as in the northern parts of Butha-Buthe District. Cold spot areas were found in parts of Quthing, Qacha-Nek, and Mokhotlong Districts. The multi-scale geographically weighted regression model explained 54.73% of the spatial variation in perinatal mortality in Lesotho. Factors that are spatially associated with perinatal mortality include advanced maternal age (35–49), cigarette smoking, history of terminated pregnancy, childbearing before age 18, and male household sex. Perinatal mortality in Lesotho is high and varies across districts. Advanced maternal age (35–49), cigarette smoking, a history of terminated pregnancy, childbirth before age 18, and male household sex are significant spatial factors associated with perinatal mortality. Due to the spatial clustering of perinatal deaths, targeted, geographically focused interventions are crucial, especially in hotspot areas like Leribe, Berea, Maseru, and the northern parts of Butha-Buthe District, to lower perinatal mortality and promote equitable health outcomes. Governments and development partners should prioritize genetic counseling, improved access to tertiary care for delivery, and raising awareness of related risks, especially for older mothers in districts such as Maseru, Mafeteng, Berea, Mohale-Hoek, and Leribe, using location-specific strategies. Policymakers should focus on reducing smoking and early childbearing, particularly in Mokhotong and Butha-Buthe districts, and other hotspots.