Segmentation of public health status using finite mixture partial least square structural equation model

This study aims to investigate the heterogeneity in the public health status model in Central Java Province, Indonesia. By using Hendric L. Blum's theory, the latent variable of health status is influenced by environment, medical care services, healthy living behavior, and genetics. The structural equation model is used to determine the mechanism of the relationship among these latent variables. Furthermore, the segments of the unit of observation that represent heterogeneity in the SEM model were identified using finite mixture partial least square (FIMIX PLS) method. The results show that districts/cities in Central Java Province can be grouped into five segments. The first segment is characterized by a structural model in which public health status is positively influenced by the environment and genetics and negatively influenced by healthy lifestyles and medical care services. The second segment is characterized by a structural model where public health status is positively influenced by medical care services and negatively influenced by environment, genetics and healthy living behavior. The third segment is characterized by a structural model where public health status is positively influenced by the environment and healthy living behavior is negatively influenced by genetics and medical care services. The fourth segment is characterized by a structural model where health status is positively influenced by healthy living behavior and medical care services and negatively influenced by environmental and genetic factors. While the fifth segment is characterized by a structural model where health status is positively influenced by healthy living behavior, medical care services, environment and genetics. The structural model in each segment has a determinant coefficient more than 90%.