Objective: to assess the coverage, completeness and reliability of data on live births in public maternity wards in São Paulo, Brazil.
Methods: data recorded in the Information System on Live Births (Sinasc) were compared with data collected in a field study (gold standard) during three months in 2011 in four maternity wards in hospital from the SUS network; kappa coefficient was calculated to assess agreement.
Results: 5,785 birth records were analyzed; Sinasc coverage was 99.8% and completeness was 96.2%; kappa values showed excellent and good agreement for maternal age (0.99), type of pregnancy, sex and newborn sex and birth weight (0.98), type of delivery (0.97), 1 minute (0.96) and 5th (0.95) minutes, previous live births (0.87) and education level (0.62); regular agreement for prenatal care visits (0.60) and gestational age (0.56); and weak agreement for previous stillbirths (0.09).
Conclusion: in the assessed hospitals, Sinasc presented high coverage, completeness and reliability.