ABSTRACT
BACKGROUND: Epidemiological research of events related to labor and delivery frequently uses maternal interview or birth certificates as a primary method of data collection; however, the validity of these data are rarely confirmed. This study aimed to examine the validity of birth certificate data and maternal interview of maternal demographics and events related to labor and delivery with data abstracted from medical records in a US setting. METHODS: Birth certificate and maternal recall data from the Iowa Health in Pregnancy Study (IHIPS), a population-based case-control study of risk factors for preterm and small-for-gestational age births, were linked to medical record data to assess the validity of events that occurred during labor and delivery along with reported maternal demographics. Sensitivity, specificity, positive and negative predictive values, and kappa scores were calculated. RESULTS: Postpartum maternal recall and birth certificate data were excellent for infant characteristics (birth weight, gestational age, infant sex) and variables related to labor and delivery (mode of delivery) when compared with medical records. Birth certificate data for labor induction had low sensitivity (46.3%) and positive predictive value (18.3%) compared to medical records. Compared to maternal interview, birth certificate data also had poor agreement for smoking and alcohol use during pregnancy. Agreement between all three methods of data collection was very low for pregnancy weight gain (kappa = 0.07-0.08). CONCLUSIONS: Maternal interview and birth certificate data can be a valid source for collecting data on infant characteristics and events that occurred during labor and delivery. However, caution should be used if solely using birth certificate data to gather data on maternal demographic and/or lifestyle factors.
