Associations between Medicaid enrollment and diagnosis stage and survival among pediatric cancer patients

Background: Medicaid-associated disparities in childhood and adolescent (pediatric) cancer diagnosis stage and survival have been reported. However, a key limitation of prior studies is the assessment of health insurance at a single time point. To evaluate Medicaid-associated disparities more robustly, we used Surveillance, Epidemiology, and End Results (SEER)–Medicaid linked data to examine diagnosis stage and survival disparities in those (i) Medicaid-enrolled and (ii) with discontinuous and continuous Medicaid enrollment. Methods: SEER–Medicaid linked data from 2006 to 2013 were obtained on cases diagnosed from 0 to 19 years. Medicaid enrollment was classified as enrolled versus not enrolled, with further classifications as continuous when enrolled 6 months before through 6 months after diagnosis, and discontinuous when not enrolled continuously for this period. We used multinomial logistic and Cox proportional hazards regression models to determine associations between enrollment measures, diagnosis stage, and cancer death adjusted for covariates. Results: Among 21,502 cases, a higher odds of distant stage diagnoses were observed in association with Medicaid enrollment (odds ratio [OR] = 1.56, 95% confidence interval [CI]: 1.48–1.65), with the highest odds for discontinuous enrollment (OR = 2.0, 95% CI: 1.86–2.15). Among 30,654 cases, any Medicaid enrollment, continuous enrollment, and discontinuous enrollment were associated with 1.68 (95% CI: 1.35–2.10), 1.66 (95% CI: 1.35–2.05), and 1.89 (95% CI: 1.54–2.33) times higher hazards of cancer death versus no enrollment, respectively. Conclusions: Medicaid enrollment, particularly discontinuous enrollment, is associated with a higher distant stage diagnosis odds and risk of death. This study supports the critical need for consistent health insurance coverage in children and adolescents.