Agreement between maternal self-report of birth-related information and medical records in Brazil: A comparison study between public, mixed, and private hospitals

Abstract

Background: The study aims to assess agreement between data obtained from interviews with postpartum women and their health records about labor and birth characteristics, newborn care, and reasons for cesarean birth. Methods: The present study analyzes the Birth in Brazil study dataset, a nationwide hospital-based survey that included 23,894 postpartum women. Reliability was assessed using kappa coefficients and 95% confidence intervals. We also calculated the proportion of specific agreement: the observed proportion of positive agreement (Ppos) and the observed proportion of negative agreement (Pneg). Results: In terms of labor and birth characteristics, more significant discrepancies in prevalence were observed for fundal pressure (1.4%–42.6%), followed by amniotomy, and augmentation. All of these variables were reported more frequently by women. Reliability was nearly perfect only for mode of delivery (kappa 0.99–1.00, Ppos and Pneg >99.0%). Higher discrepancies in reasons for cesarean prevalence were observed for previous cesarean birth (CB) (3.9%–10.4%) and diabetes mellitus (0.5%–8.5%). Most kappa coefficients for CB reasons were moderate to substantial. Lower coefficients were seen for diabetes mellitus, induction failure, and prelabor rupture of membranes and Pneg was consistently higher than Ppos. Discussion: Our findings raise relevant questions about the quality of information shared with women during and after the process of care for labor and birth, as well as the information recorded in medical charts. Not having access to full information about their own health status at birth may impair women's health promotion behaviors or clear disclosure of risk factors in future interactions with the healthcare system.