Objectives: This study aimed to identify the regional factors affecting liver cancer mortality using geographically weighted regression (GWR). Methods: Data from 2017 to 2019 were utilized, including health examination statistics and regional medical utilization statistics from the National Health Insurance Corporation, regional infectious disease statistics and community health surveys from the Korea Centers for Disease Control and Prevention, mortality statistics and population census data from Statistics Korea. Spatial data were obtained from the Statistical Geographic Information Service, covering 229 cities, counties, and districts. Statistical analysis was performed using R version 4.3.2, with significant variables identified through ordinary least squares (OLS) analysis, followed by GWR to assess model improvement. Results: The OLS analysis identified percent of population aged 40+, incidence of hepatitis C, high-risk drinking rate, and national health insurance premium as significant factors. Moran’s Index showed a positive spatial autocorrelation (0.373, <i>p</i><0.001). GWR analysis increased the adjusted R² from 48% to 61% and reduced the AIC from 1,062.21 to 1,038.59, indicating an improvement in the model’s explanatory power and fit. In Jeollanam-do and Jeju-do, the explanatory power was high, and the percent of population aged 40+ had a strong influence in Gangwon-do. The incidence of hepatitis C had a strong influence in the southern coastal regions, the high-risk drinking rate was notably significant in the Yeongnam region, and the national health insurance premium had a strong influence in Busan. Conclusions: As significant factors vary by region, it is considered important to establish health care policies and projects that take into account the characteristics of each region, and through this, we can expect to reduce liver cancer mortality in the region and improve community health levels.
Read full abstract