Validity of maternal birthweight recall among Colombian children

Low birthweight and preterm birth are associated with adverse health outcomes later in life, but acquisition of accurate birthweight information is not always feasible in large epidemiological studies. We examined the validity of child birthweight and gestational age recall by mothers, and the extent to which recall bias affects associations between birthweight and childhood obesity in children from Bogotá, Colombia. We surveyed mothers of 3,202 schoolchildren aged 5-12 years about child's weight and gestational age at birth, and sociodemographic characteristics. In a subsample of 279 children, we obtained hospital birth records and extracted birthweight, gestational age, and other perinatal information. Mean birthweight (SD) was 3,106 (739) grams according to maternal recall and 2,977 (462) grams according to hospital records (difference 129 g; 95% CI = 55, 203). Thirty-three percent of mothers recalled their children's birthweights exactly as they appeared in hospital records. Mother's age and fewer years of education were each significantly associated with greater birthweight recall bias. Specificity of low birthweight (<2,500 g) and preterm birth (<37 weeks gestation) from maternal recall was 0.95 and 0.86, respectively; however, sensitivity was lower (0.66 and 0.67, respectively). Associations between recalled birthweight and BMI-for-age or overweight during school age were weaker than those with hospital record birthweight. Maternal birthweight recall 5-12 years after birth differs from hospital record birthweight by a clinically meaningful amount. Birthweight recall should be used with caution in epidemiological studies conducted in this and comparable settings. Associations between birthweight and obesity may be stronger than they appear when using recalled birthweight.