Validation of self-reported data on assisted conception in The Danish National Birth Cohort.

Background: An increasing number of children are born after assisted conception and in surveillance programmes information on mode of conception is often achieved via maternal self-report. We assessed the validity of self-reported assisted conception in The Danish National Birth Cohort (DNBC), a prospective pregnancy cohort. Here, the term assisted conception refers to IVF, ICSI, ovulation induction and insemination.
Methods: We compared self-reported assisted conception in the DNBC to corresponding data from Danish national registers; the IVF Register and Danish Drug Prescription Register, providing method of conception in the entire population. In the DNBC, 101,042 women accepted the invitation in early pregnancy from 1996 to 2002. Our final study population comprised 88,151 DNBC women aged 20 years and older who participated in the first DNBC interview with a pregnancy resulting in a live born child.
Results: In the DNBC, assisted conception was reported with a sensitivity of 83% and positive predictive value of 88%. Misclassification was largely explained by ambiguous phrasing of the DNBC interview question and interview skip patterns. Women with false negative reporting were more often multipara (P < 0.001) and older (P = 0.027 for IVF/ICSI and P = 0.002 for ovulation induction). The risk ratio (RR) for being born preterm in IVF/ICSI children was lower for children identified via the DNBC, RR 3.61 (95% confidence interval (CI) 3.31-3.94), than the IVF Register, RR 4.36 (95% CI 4.02-4.74).
Conclusions: There was a high positive predictive value of self-reported assisted conception in the DNBC, but the structure of the DNBC interview represented a problem and misclassification could introduce bias.