Consistency of reported alcohol use by pregnant women: anonymous versus confidential questionnaires with item nonresponse differences.

Background: Self-report is the best method to attain information about moderate alcohol consumption. However, reported alcohol consumption before and during pregnancy is sensitive to response bias. Methods: We compared identical questionnaires completed either confidentially or anonymously by pregnant women. A representative half of the pregnant women in Oslo attending ultrasound screening between June 2000 and May 2001 were invited to join a confidential questionnaire study. Another group of pregnant women, selected in the identical manner, were asked to answer the same questionnaire anonymously during the summer of 2001. The questionnaire was completed by 1749 confidentially and 191 anonymously (response rate 93.4% vs 84.4%). For the current study, only women with Scandinavian ethnicity were included: 1707 and 178, respectively. Measures were T-ACE (screening measure for pregnancy risk drinking), reported frequency of alcohol use, Standard Units (SU) per occasion (po), SU per week, and binge drinking ( >= 5 SU po), before and during pregnancy. Results: There were no significant differences in reported alcohol consumption between the confidential and the anonymous group for the sample as a whole. Any alcohol consumption was reported by 22.6% and 23.3% during pregnancy week 7-12 and 13+ (confidential group) versus 22.5% and 25.8% (anonymous group). Differences between the groups tended to be larger for SU per week (p = 0.07 both before pregnancy and after week 12 of pregnancy) than for the indirect alcohol (T-ACE) and the binge drinking questions. In the confidential group, the women with lower education had close to twice as high item nonresponse on the direct alcohol questions during pregnancy compared with those with higher education. This difference was smaller in the anonymous group. The rate of smoking was not reported differently in the anonymous group. Conclusions: Among pregnant women, there was no significant difference in self-reported alcohol consumption obtained by confidential or anonymous questionnaires.