Determination of factors that affect under five mortality rate using geographically weighted generalized Poisson Regression (GWGPR) in Lamongan

The statistical analysis approach that experienced overdispersion and considered the spatial effect was Geographically Weighted Generalized Poisson Regression (GWGPR). The GWGPR method can be applied to the health sector, such as the Under Five Mortality Rate (U5MR). The reduction in U5MR, which is one of the targets that must be achieved by the Sustainable Development Goals (SDGs) with the U5MR (<5 years) is 25/1000 births. However, Lamongan Regency experienced a fairly high U5MR increase in 2020 - 2021 with 11 cases to 81 cases. The U5MR data in Lamongan Regency have an overdispersion problem and it shows spatial heterogeneity. Thus, this study aims to determine the factors that influence the under Five Mortality Rate (U5MR) in Lamongan Regency in 2021 using GWGPR with the best adaptive kernel weighting. The results of the study used three adaptive kernel weights: the Adaptive Gaussian Kernel, Adaptive Bisquare Kernel, and Adaptive Tricube Kernel, showed that the best weighting for the GWGPR model with the smallest AIC of 61.0517 is the Adaptive Bisquare Kernel. Obtained 27 equation models, one of them is Kedungpring district with factors that affect U5MR, such as Exclusive Breastfeeding (X1), Low Birth Weight (X2), Thin Toddlers (X3), Toddler Tuberculosis (X4), Number of Posyandu (Integrated Service Post) (X5), and Number of Permanent Healthy Toilets (X6).