What can we glean from linking registry and Medicaid claims data?

Abstract

6055 Background: Cancer quality of care problems are concentrated among the indigent. Although Medicare claims linked to tumor registry data have become an important source for evaluating cancer care, parallel efforts to use Medicaid claims are not well developed. Objective: To evaluate the feasibility and potential value of linking the California state cancer registry with Medicaid claims for health services research. Methods: Cases reported to the California Cancer Registry (CCR) between 1988 and 2000 were matched with California Medicaid enrollment and claims data. We assessed the number of cancer cases that appeared in the Medicaid enrollment data. We evaluated the sensitivity of Medicaid claims to capture cancer diagnoses and treatments reported in the registry data. Results: 6,800 patients ages 18–64 had records in the 1998 CCR data that were matched to the 1998 Medicaid enrollment file; Medicaid enrollees accounted for 34.4% of patients with hepatoma and 21.1% of patients with cervix cancer, but only 8.9% of breast and 8.9% of testicular cancer patients in this age bracket. Medicaid data had only modest sensitivity to capture cancer diagnoses and procedures. For 5 cancers, Medicaid claims included procedure codes for about 70% of surgeries reported in the registry data. Missing diagnoses and procedures could be attributed to discontinuous enrollment, Medicare crossover and enrollment in pre-paid health plans. Conclusion: Linkage of CCR and Medicaid enrollment data identified sizable populations of indigent patients. However, the modest sensitivity of Medicaid claims to capture cancer diagnoses and procedures currently limits the utility of Medicaid data to assess cancer treatment and outcomes. No significant financial relationships to disclose.