Semivariogram Modeling based on Provincial Clusters for Infectious Diseases and Mental Health in Indonesia

Abstract

Health is a valuable asset that profoundly impacts individuals and society as a whole, enhancing overall well-being and quality of life. Both internal and external factors, along with geographical location, play a crucial role in health. These factors exhibit spatial patterns that can be effectively analyzed through geostatistical methods, particularly semivariogram modeling. This study explores appropriate semivariogram models to depict disease distribution in Indonesian provinces using data from National Health Insurance Agency (NHIA). The provinces will be grouped into five clusters based on the Consumer Price Index (CPI), health claim amounts, the number of participants, and 23 disease groups through non-hierarchical cluster analysis. Three clusters, with the most provinces, will be selected for semivariogram modeling: exponential, Gaussian, and Spherical models. The best-fitting semivariogram models are anisotropic exponential for claim amounts and anisotropic Gaussian for CPI, number of participants, infectious diseases, and mental health issues. Meanwhile, the most suitable spherical model is identified for a specific cluster (Kalimantan and Nusa Tenggara regions). The results of this modeling can serve as recommendations for the inter-province radius of influence in disease prevention measures and the creation of a high-quality environment.