Completeness of American Cancer Registry Treatment Data: implications for quality of care research.

BACKGROUND: Evaluating and improving the quality of cancer care requires complete information on cancer stage and treatment. Hospital-based registries are a key tool in this effort, but reports in the 1990s showed that they fail to identify a major fraction of outpatient-administered treatment, including chemotherapy, endocrine therapy, and radiation. This can limit their value for evaluating patterns and quality of care. To determine the completeness of registry data in more recent years, we linked administrative claims from 2 private payers in Ohio to the National Cancer Data Base and Ohio Cancer Incidence and Surveillance System.
METHODS: Incident breast and colorectal cancers among Ohio residents diagnosed in 2004-2006 were identified from linkage of the National Cancer Data Base, Ohio Cancer Incidence and Surveillance System, and payer insurance claims using ICD-9 and CPT procedure codes, and ICD-9 diagnosis codes. Linkage was accomplished using patient demographics, surgery dates, and hospital facility. Treatment found in claims and registry data were compared and assessed using the κ statistic.
RESULTS: The analytic cohort included 2,552 breast and 822 colorectal cases. Results showed high agreement for breast surgery type, and moderately high agreement for colorectal surgery type. For breast cases, the registries captured 87% of chemotherapy, 86% of radiation, and 64% of endocrine treatment in claims. For colorectal cases, the registry captured 83% of chemotherapy and 84% of radiation in claims.
CONCLUSIONS: Hospital-based registries for breast and colon cancer diagnosed in 2004-2006 captured about 85% of radiation and chemotherapy data compared with claims data, a higher percentage than earlier reports. These findings provide direction and a cautionary note to those using registry data for study of patterns and quality of systemic and radiation therapy care.