Purpose: Surgery is an important part of early stage breast cancer treatment that affects overall survival. Many studies of surgical treatment of breast cancer rely on data sources that condition on continuous insurance coverage or treatment at specified facilities and thus under-sample populations especially affected by cancer care inequities including the uninsured and rural populations. Statewide cancer registries contain data on first course of cancer treatment for all patients diagnosed with cancer but the accuracy of these data are uncertain.
Methods: Patients diagnosed with stage I-III breast cancer between 2003 and 2016 were identified using the North Carolina Central Cancer Registry and linked to Medicaid, Medicare, and private insurance claims. We calculated the sensitivity, specificity, positive predictive value, negative predictive value, and Kappa statistics for receipt of surgery and type of surgery (breast conserving surgery or mastectomy) using the insurance claims as the presumed gold standard. Analyses were stratified by race, insurance type, and rurality.
Results: Of 26,819 patients who met eligibility criteria, 23,125 were identified as having surgery in both the claims and registry for a sensitivity of 97.9% (95% CI 97.8%, 98.1%). There was also strong agreement for surgery type between the cancer registry and the insurance claims (Kappa: 0.91). Registry treatment data validity was lower for Medicaid insured patients than for Medicare and commercially insured patients.
Conclusions: Cancer registry treatment data reliably identified receipt and type of breast cancer surgery. Cancer registries are an important source of data for understanding cancer care in underrepresented populations.
Keywords: Breast cancer; Cancer registry; Epidemiology; Methods; Validity.
© 2021. The Author(s), under exclusive licence to Springer Nature Switzerland AG.