Background: Maternal self-report is the most common method for assessment of past cigarette exposure to assess birth defect risk. This study compared maternal smoking prior to and during pregnancy based on self-reports obtained from the medical records abstracted for the Utah Birth Defect Network (UBDN), the birth certificate, and the computer-assisted telephone interview (CATI) in the National Birth Defects Prevention Study (NBDPS). The study also investigated how the different sources for maternal smoking data affect estimates in an empirical study.
Methods: A total of 1774 case and 618 control mothers who had participated in the NBDPS and whose live born infants were delivered between January 1, 2003, and December 31, 2007, were included in this study. Among the case mothers, we compared data from all three sources, whereas for control mothers only two data sources were available for comparison (i.e., birth certificate and CATI).
Results: Smoking prevalence was highest in the CATI. Compared to the CATI, data from the UBDN had a higher sensitivity (61.3%) and better agreement (kappa = 0.63) than birth certificates (51.8%; kappa = 0.56). Adjusted odds ratios for all and specific birth defects (i.e., holoprosencephaly, hydrocephalus, anophthalmia/microphthalmia, anotia/microtia, total anomalous pulmonary venous return/partial anomalous pulmonary venous return [TAPVR/PAPVR], heterotaxy, and gastroschisis) were different between the birth certificate and CATI. The change in the effect estimates between the two sources ranged from 19% to 56%.
Conclusions: Based on our findings, maternal smoking exposure from interview data was shown to be of higher quality with less misclassification compared to data obtained from medical records or birth certificates.
Copyright © 2012 Wiley Periodicals, Inc.