Abstract
Study Aim: To compare 5-year survival, by Medicaid status, in adults diagnosed with a curable cancer, defined as 5-year survival without relapse. We hypothesized that adjusting for patient attributes, as well as cancer site and stage, Medicaid beneficiaries affected with these cancer experienced greater mortality than their non-Medicaid counterparts.
Methods: Using a database developed by linking records from the Ohio Cancer Incidence Surveillance System (OCISS) with Ohio Medicaid and death certificate files, we identified individuals 15-54 years of age and diagnosed with any of the following incident cancers in the years 1996-2002: cancer of the testis; Hodgkin's and non-Hodgkin's lymphoma; early-stage colon, lung, and bladder cancer; pediatric malignancies; and early-stage melanoma (n=13,004). Medicaid status was ascertained if the patient was enrolled in Medicaid in the year s/he was diagnosed with cancer. Death certificate data were used to ascertain the patient's vital status through 2007, and to retrieve the date and cause of death. The individual's income was approximated to that of the census block group in which s/he resided at the time of diagnosis. Using logistic regression analysis and Cox regression models, respectively, we examined the association between Medicaid status and each of 5-year mortality and hazard of death after adjusting for patient demographics, income, geographic area of residence, and cancer site and stage at diagnosis.
Results: There were 12,703 subjects in our population 60% of whom were 40 years of age or older; 57% were male; 6% were African American, and 10% were identified as Medicaid beneficiaries. The most common types/anatomic sites of cancer were melanoma (36.4%), colon (22.3%), testis (14.5), and Hodgkin's (7.0%), and non-Hodgkin's lymphoma (9.1 %). The proportion of patients dying of cancer within 5 years of diagnosis was highest among African Americans (18.3%), Medicaid beneficiaries (21.9%), those diagnosed with lung cancer (33.9%) or pediatric malignancies (31.2%), and those diagnosed with more advanced stages of cancer (20.9% and 22.2% among those diagnosed with regional, and distant-stage cancer, respectively). The findings from the multivariable logistic analysis indicated that Medicaid beneficiaries were nearly twice as likely as their non-Medicaid counterparts to die of cancer within the 5 years following diagnosis (adjusted odds ratio: 1.84,95% Confidence Interval: 1.54,2.21). Similarly, results obtained from the survival analysis indicated that Medicaid beneficiaries experienced significantly greater hazard of death from cancer (adjusted hazard ratio: 1.78 (1.53,2.07)).
Conclusion: Consistent with our hypothesis, we found Medicaid status to be associated with significantly greater likelihood of 5-year cancer specific mortality and hazard of death, even after adjusting for potential confounders. In addition to serving the poor and the disabled, the Medicaid program is the primary safety net to the near poor who resort to it when they are diagnosed with a catastrophic disease such as cancer. In this study, nearly 45% of Medicaid patients joined the program upon, or after being diagnosed with cancer. Future studies should determine to what extent higher mortality among Medicaid beneficiaries is attributed to advanced cancer stage at diagnosis or cancer relapse due to inadequate treatment and surveillance.
Citation Information: Cancer Epidemiol Biomarkers Prev 2010;19(10 Suppl):A62.
RSS Feeds
