This study investigates the concordance of clinical and demographic data under the Unified National Health System (SUS) in Brazil, comparing authorizations for high-complexity oncological procedures in breast cancer with a random sample of 310 corresponding medical records in Rio de Janeiro State, Brazil. Concordance was verified using kappa (K), prevalence-adjusted bias-adjusted kappa (PABAK), weighted-kappa (KW), intra-class correlation (ICC), and relative frequency statistics. 95% confidence intervals were estimated, and Landis and Koch criteria were applied for concordance interpretation. Assuming PABAK as the most appropriate statistic for the variables analyzed, we observed a moderate concordance for cancer staging, PABAK = 0.63 (95%CI: 0.56-0.70), and for morphological diagnosis, PABAK = 0.84 (95%CI: 0.72-0.96). For date of birth and date of diagnosis, ICCs were 0.96 (95%CI: 0.95-0.96) and 0.92 (95%CI: 0.90-0.94), respectively. Concordance regarding municipality of residence was 52.9%. The study indicates that, except for municipality of residence, the data concordance comparing authorizations for procedures and patient medical records in Rio de Janeiro allows its application to epidemiological and health services research, as well as for health planning purposes.
Women's Health; Breast Cancer; Reliability and Validity
