Municipal-level correlates and spatial variations of Korea’s national cancer screening program in 2018 and 2022: focus on weighted ordinary least squares regression and local indicators of spatial association

The Korean National Cancer Screening Program (KNCSP) addresses prevalent cancers; however, there is a lack of comprehensive understanding regarding municipal-level disparities in screening engagement and spatial patterns. This study aimed to analyze the associations between municipal health metrics and cancer screening participation rates, and to assess whether spatial clustering persisted even after adjusting for these metrics. We calculated screening rates for stomach, colorectal, liver, breast, and cervical cancers using a customized National Health Insurance Service (NHIS) KNCSP database. The final source populations included 21,879,658 individuals in 2018 and 21,477,470 individuals in 2022. We linked 10 municipal indicators from the Korean Community Health Status Indicators (K-CHSI) database to cancer screening rates for 2018 and 2022. We evaluated the unadjusted correlation between the screening rate and municipal indicators and estimated the adjusted associations using a weighted regression model. Residual Local Indicators of Spatial Association (LISA) were employed to identify clusters of municipalities with higher or lower participation and to detect emerging spots. In 2018, the screening rate was highest for breast cancer (77.6%) and stomach cancer (76.1%). By 2022, breast and stomach cancer screening rates remained the highest, with breast cancer at 62.3% and stomach cancer at 62.1%. The general health examination rate in 2022 was the strongest positive correlate of breast cancer screening (β = 0.582, SE = 0.067, P < 0.001) and stomach cancer screening (β = 0.502, SE = 0.086, P < 0.001). Conversely, the elderly population ratio was the most substantial negative factor for cervix cancer screening (β=−0.481, SE = 0.067, P < 0.001). Spatial analysis revealed LISA maps displaying localized high- and low-participation clusters in screening rates, while residual LISA identified clusters that deviated from the expected rates after regression adjustment. Factors influencing the municipal KNCSP screening rate between 2018 and 2022 were primarily linked to general health examinations and population aging. However, the persistent presence of cold spots/hotspots in the residual spatial analysis indicates the potential presence of other contributing factors beyond those considered in the analysis. Integrating population-weighted regression with residual spatial diagnostics can aid in identifying and prioritizing municipalities for focused local evaluations and customized implementation strategies.