County-Level Medical Debt and Its Associations With Cancer Stage and Survival in the United States.

BACKGROUND: Medical debt is recognized as a social determinant of health. This study examines associations of county-level medical debt in collections with diagnosis stage and survival among individuals newly diagnosed with cancer in the United States. PATIENTS AND METHODS: A cohort of individuals aged ≥18 years newly diagnosed with cancer in 2011-2019 was identified from the National Cancer Database and followed through December 31, 2019. The county-level share of adults with medical debt in collections was combined with patient-level data. Hierarchical multivariable logistic and Cox proportional hazard models estimated associations of county-level medical debt with cancer diagnosis stage and overall survival, respectively, for all patients and by cancer type, clinical, and socioeconomic subgroups. RESULTS: A total of 7,558,658 individuals with cancer were identified, with median county-level medical debt of 18%, ranging from 0% to 56%. Patients in the highest medical debt quartile had the highest proportion with stage IV disease and lowest 5-year survival rate. After adjusting for other patient- and county-level characteristics, patients living in counties in the highest quartile of medical debt were more likely to be diagnosed with stage IV disease (odds ratio, 1.079; 95% CI, 1.064-1.094) and had poorer survival (hazard ratio, 1.072; 95% CI, 1.061-1.082) than those living in counties in the lowest quartile of medical debt, with a statistically significant dose-response relationship (P for trend <.001). Patterns were observed across major cancer types and were consistent across socioeconomic and clinical subgroups. CONCLUSIONS: County-level medical debt was associated with advanced-stage diagnosis and worse survival among individuals newly diagnosed with cancer. Future research is warranted to evaluate the recent and ongoing US policy changes on medical debt prevalence and their effects on cancer care and outcomes.