[Validity of information on birth weight for studies based on fetal programming]

Birth weight is essential information in fetal programming studies and is generally obtained retrospectively. In Brazil's Information System on Live Births (SINASC), birth weight is valid information but is not always accessible. The study aimed to establish an algorithm for the selection of the most reliable data source for birth weight in the absence of information in the SINASC database. In a cross-sectional study of 6-14-year-old schoolchildren in Niterói, Rio de Janeiro State, Brazil, in 2010, birth weight was collected through a self-completed questionnaire, interview, medical records from the Family Physician Program, and the SINASC database. We calculated intra-class correlation coefficients (ICCs) and differences in mean birth weight. ICCs varied from 0.90 to 0.99. All the other sources showed higher mean birth weight than SINASC, allowing differences up to 100g. Birth weight is recorded at birth (SINASC) or close to it (Family Physician Program), and in the absence of these sources, birth weight as retrieved at 6-14 years of age is a reliable option. To complement information on birth weight in the absence of SINASC, we recommend the following order: Family Physician Program, interview, and questionnaire.