Abstract
BACKGROUND: Treatment at a cancer center improves outcomes for individual patients, but the impact of cancer centers on the regions they serve is unknown. This study examines the association between cancer center presence and cancer mortality within health service areas (HSA), which define US regional care markets. METHODS: Cancer centers and cancer mortality were mapped to HSAs using geospatial analysis. The association between cancer center presence and cancer mortality was analyzed using a multivariable spatial lag model, adjusting for the percent of incident cases presenting at late stage, social vulnerability, environmental burden, and rurality. Results are reported as deaths per 100,000 people with 95% confidence interval (CI). RESULTS: Of 940 HSAs, 43% (n = 407) contained at least one cancer center. HSAs with a cancer center had lower cancer mortality than those without a cancer center [158 (95% CI, 155.9-159.5) versus 165 (95% CI, 162.5-166.6) deaths per 100,000 people]. In adjusted analyses, metropolitan HSAs with cancer centers had 4.4 fewer deaths (-8.4 to -0.4) than those without cancer centers whereas there was no observed effect of cancer center presence in nonmetropolitan HSAs (1.4; 95% CI, -1.4 to 4.3). Factors associated with higher cancer mortality were late stage at presentation and greater environmental burden. CONCLUSIONS: Cancer center presence within metropolitan healthcare markets was associated with modestly lower cancer mortality. IMPACT: There is a need to evaluate underlying mechanisms whereby cancer center presence may affect population-level cancer mortality in metropolitan, but not nonmetropolitan HSAs, including direct patient care and indirect effects such as community health promotion.