Defining care provided for breast cancer based on medical record review or Medicare claims: Information from the Centers for Disease Control and Prevention Patterns of Care Study

Background: Description of care patterns is important as evidence-based guidelines increasingly dictate care. We explore the level of agreement between claims and record abstraction for guideline concordant multidisciplinary breast cancer care. Methods: From the U.S. Centers for Disease Control and Prevention's National Program of Cancer Registries Patterns of Care study, in which medical record abstraction of breast cancer and treatment was accomplished, cases include breast cancer where Medicare claims were available. Components of care were breast-conserving surgery (BCS), mastectomy, node assessment, radiation (RT), and chemotherapy (CTX), including specific chemotherapeutic agents, and combinations. We compared Medicare claims with record abstraction, and measured concordance using the kappa statistic and sensitivity. Results: The study sample consisted of 1762 women with stage 0 to 4 breast cancer. Level of agreement was excellent for surgery type (kappa = 0.84) and CTX (kappa = 0.89); agreement for RT therapy was slightly lower (kappa = 0.79). For standard multicomponent strategies, sensitivities and specificities were high; for example, 88.8%/93.5% for mastectomy plus nodes and 86.6%/95.4% for BCS plus nodes and RT. For selected, standard, multi-agent, adjuvant CTX regimens, sensitivities ranged from 66.3% to 68.8% (kappa 0.63-0.73). Conclusions: Medicare claims, compared with chart abstraction, is a reliable method for determining patterns of multicomponent care for breast cancer.