Validity of maternal recall of gestational age and weight at birth: Comparison of structured interview and medical records. Issue 2 (29th November 2018)

• Record Type: Journal Article
• Authors: Petersen, Julie M.; Mitchell, Allen A.; Van Bennekom, Carla; Werler, Martha M.
• Is Part Of: Pharmacoepidemiology and drug safety. Volume 28:Issue 2(2019)
• Page Start: 269
• Page End: 273
• Publication Date: 2018-11-29
• Subjects: pharmacoepidemiology -- pregnancy outcomes -- reliability and validity -- self‐report; Pharmacoepidemiology -- Periodicals; Chemotherapy -- Periodicals; Epidemiology -- Periodicals; 615.705
• DOI: 10.1002/pds.4699
• Languages: English
• ISSNs: 1053-8569

Abstract

Abstract: Purpose: To assess agreement between maternal recall and medical records for gestational age (GA) at birth as derived from dating information and birthweight. Methods: In the case‐control Slone Birth Defects Study, within 6 months of delivery, trained nurses conducted standardized telephone interviews with mothers of infants with and without major structural malformations. In a subset of approximately 5000 case and control mothers from five US centers (2008–2012), a research nurse abstracted subjects' medical records. GA at delivery was calculated as date of delivery minus (estimated date of confinement [EDC] minus 280); if EDC was unknown, last menstrual period served as a proxy for start of pregnancy. Positive and negative predictive values (PPV and NPV, respectively) were calculated, using medical records as the standard, for categories of GA at delivery (ie, early preterm <238, late preterm <258, and term ≥259 d) and birthweight (low <2500, normal 2500 to 4500, and high >4500 g). Results: The gestational age and birthweight validation samples comprised 3122 and 4760 women, respectively, with diverse characteristics. The PPV and NPV were high (>92% and >99%, respectively) for all categories of delivery GA and birthweight. Conclusions: Our findings suggest that mothers' recall may be a valid alternative to medical records to estimate delivery GA and birthweight. This study used standardized interviews conducted by trained research nurses, had a short recall period (<6 months post delivery), and for delivery GA, focused on date‐derived GA. Further research is needed on the potential impact of study design, population characteristics, and comparison to other data sources.