Answers are grounded in this paper and cite their sources.
National cancer screening programs aim to provide equitable access, but regional deprivation may still shape participation. Whether area-level gradients are consistent across cancer types within one national program, and how sensitive they are to the proportion of older residents and rurality, is unclear. We conducted an ecological panel study of Korean municipalities (si-gun-gu) from 2020 to 2024 using public national statistics (up to 253 municipalities per year). A study-specific deprivation index (DI) of eight indicators was conceptually informed by the US Area Deprivation Index and French European Deprivation Index. Screening rates for six cancers were modelled against DI using ordinary least squares with year fixed effects, generalized estimating equations (GEE), and binomial logistic regression of screened over eligible counts. Rural–non-rural contrasts were modelled with and without adjustment for the proportion aged ≥ 65 years. Sensitivity analyses used a common six-indicator index and an index excluding its age-related component, and assessed multicollinearity, facility density, and residual spatial autocorrelation. Exploratory random forests used repeated cross-validation. Unadjusted differences between the least and most deprived quintiles were 10.2% points (pp) for cervical, 3.7 pp for breast, and 2.1 pp for liver cancer screening, and were negligible for gastric, colorectal, and lung cancer. DI was strongly correlated with the proportion of residents aged ≥ 65 years ( r = 0.76). After adjustment for the proportion aged ≥ 65 years, the proportion of women, and rurality, breast and cervical cancer associations were attenuated toward the null, whereas liver (β = 2.34; 95% CI 0.86, 3.81) and lung (β = 4.83; 95% CI 2.06, 7.60) cancer coefficients became positive. These patterns were directionally consistent using the alternative indices. In exploratory random forests, whose predictive performance was limited for several cancers, the proportion aged ≥ 65 years had the highest relative importance for cervical (86.0%) and breast (53.0%) cancer screening. The adjusted rural–non-rural gap was largest for lung cancer (− 8.13 pp; 95% CI -11.91, -4.36; -7.06 pp after age adjustment). Deprivation–screening associations were cancer-specific and sensitive to adjustment for the proportion of older residents. This attenuation may reflect demographic differences and overlap between predictors and does not indicate that deprivation is unimportant. Assessing municipal screening needs may require considering deprivation, age composition, and rurality jointly; causal mechanisms and intervention effectiveness were not evaluated.
No takes yet. Share an insight, caveat, or question.
Rajaguru et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: