Objectives: Although widely used in epidemiological studies, self-report has been shown to underestimate the prevalence of smoking among pregnant women. Objectives of this study were to examine the discrepancy between self-reported and cotinine-validated smoking status, and the sociodemographic characteristics associated with the misclassification of real smoking status among pregnant women in Tallinn, the capital of Estonia.
Methods: Serum cotinine assays were performed on a subsample (n = 1360) of the pregnant women, who had participated in a recent study of human papillomavirus type 16 (HPV-16) seroprevalence in Estonia. In the present study, serum concentrations > or = 15 ng/ml were used to distinguish current smokers from nonsmokers. The serum-validated smoking level was compared with the self-reported level in the records of the Estonian Medical Birth Registry. For the group of self-reported non-smokers, the differences between the cotinine-validated smokers and the cotinine-validated nonsmokers, with respect to their sociodemographic characteristics (age, ethnicity, educational level, employment status, marital status, parity), were estimated by logistic regression.
Results: Of 1239 women who reported being nonsmokers, 259 (20.9%) had serum cotinine levels > or = 15 ng/ml, and can be regarded as current smokers. Among self-reported nonsmokers, nondisclosure of current smoking was significantly more frequent in non-Estonian, less educated, socially inactive, cohabiting and multiparous women.
Conclusions: Self-reported data on smoking in pregnant women underestimates the real smoking prevalence in Estonia. Maternal unwillingness to declare smoking during pregnancy needs to be taken into account in the practice of maternal and child health to better target prenatal smoking cessation interventions.