Smoking during pregnancy: how reliable are maternal self reports in New Zealand?

OBJECTIVE To determine the reliability of self reports of smoking during pregnancy. METHODS Residual sera from early and late antenatal blood samples were tested for cotinine for all pregnancies over a six month period. Over an overlapping 12 month period, a postal questionnaire on smoking was also sent to all new mothers (n = 4857) when their baby was 4-8 weeks old. Smoking status from obstetric booking notes was also obtained. RESULTS The cotinine-validated smoking prevalence was 31.3% for the first trimester and 27.7% for the third trimester. Questionnaire self reported prevalences were 19.2% and 15.7% for the first and third trimesters respectively, and 18.9% for obstetric booking. Of cotinine-validated smokers, 22% denied smoking-self deceivers. Of mothers who replied to the questionnaire, a half appeared to systematically under report the amount they smoked. CONCLUSIONS Nearly a quarter of smoking pregnant women did not report smoking. Moreover, of those who did, the amount smoked was often under reported. This tendency to under report may rise as pressures to stop smoking increase. Accurate measures of smoking prevalence in pregnant women will require objective testing.