A Predictive Model of Regional Dementia Prevalence Using Geographic Weighted Regression Analysis

Abstract

Globally, dementia is one of the highest priority public health policy issues. This study was conducted to analyze the spatial distribution pattern of dementia prevalence using geographic weighted regression analysis and to identify preventable risk factors at the regional level of dementia prevalence. For the data to be analyzed, this work used the 2020 regional dementia prevalence index of the Korea Central Dementia Center and the regional health statistics of the Korea Centers for Disease Control and Prevention Agency (KDCA). Spatial autocorrelation analysis, hot spot analysis, and geographic weighted regression analysis were performed to identify regional associations of dementia prevalence, cluster regions with high dementia prevalence, and risk factors for regional dementia prevalence. As a result of the hot spot analysis, the regions corresponding to the hot spots with the high prevalence of dementia were found to be adjacent to each other, such as in Jeonnam, Jeonbuk, and Gyeongbuk, and the regions corresponding to the cold spots with the low prevalence of dementia were adjacent to each other, such as Seoul, Gyeonggi, Incheon, Busan, and Ulsan. The results of geographic weighted regression analysis showed that educational level, walking practice rate, hypertension prevalence, and a low-sodium diet preference were found to be risk factors for the prevalence of dementia. These results suggest that there is a need for a dementia prevalence management strategy to increase the walking practice rate and low-sodium diet preference rate, and decrease the hypertension prevalence, centering on the hot spot area, which is a cluster area with high dementia prevalence. This study is expected to be useful as basic data that can help in prioritizing health policies considering spatial characteristics for community health promotion.