Use of cancer registry data for research on patterns of breast cancer care of individuals with Medicaid insurance

Abstract

6021 Background: Data from Tumor Registries are used to guide important decisions regarding cancer care and research, policy and funding. Inferences are drawn about the adequacy and timeliness of treatment and whether such treatment is disproportionately allocated by patient characteristics such as race, age, sex, insurance type and others. This study asked whether these data were adequate, and whether supplemental information from medicaid improved the overall quality and completeness of patient treatment information. Methods: To study breast cancer care of low-income insured individuals, we developed a Medicaid-Claims and tumor registry linked database using data on 851 new cases of breast cancer diagnosed in calendar years 1997–1999 who were continuously enrolled in Medicaid in the state of North Carolina for 1 year post-diagnosis. The validity of this linked database was tested using an algorithm designed to jointly use the information from registry and claims data sources. Results: Substantial and systematic gaps in the tumor registry data were found, particularly for treatments such as chemotherapy, that occur during the initial period of diagnosis of a cancer. Approximately 75 percent of cases in the claims-linked dataset had agreement with tumor registry, and an additional 4 percent showed agreement with an absence of surgery data for a total agreement rate of 79 percent. Agreement between Medicaid and tumor registry for individual components of care ranged from 65% to 89%. Conclusions: Consideration should be given to establishing automated, validated integration of Medicaid claims with state tumor registry systems to support studies on cancer care of individuals enrolled with this important form of public health insurance. No significant financial relationships to disclose.