Smoking status in pregnant women assessment of self-report against carbon monoxide (CO)

Addict Behav. 2001 Jan-Feb;26(1):1-9. doi: 10.1016/s0306-4603(00)00070-8.

Abstract

Smoking according to self-report and to carbon monoxide (CO) were examined in 7,405 pregnant women who agreed to complete a breath test and questionnaire. The sensitivity and specificity of self-report against CO, and predictors of a high CO among reported nonsmokers, were examined. The prevalence of reported smoking was 23% (95% CI: 19.4-26.4), and 20% had a CO of 9 or more (95% CI: 17.0-23.1). The sensitivity and specificity against CO were 87% (95% CI: 83.1-91.6) and 93% (95% CI: 91.9-94.6). The positive predictive value was 76% (95% CI: 73.2-79.8), reflecting CO's inability to detect light smoking. The negative predictive value was 97% (95% CI: 95.6-97.8). The best predictors of high CO among reported nonsmokers were being divorced/separated (9.8% had high CO), quitting during pregnancy (8.5% had high CO), being third trimester, and high passive exposure. Some pregnant women may report inaccurately despite validation. Difficulties remain in determining the contributions of passive exposure and inaccurate report to inconsistencies between biochemical and self-report measures.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Breath Tests
  • Carbon Monoxide / analysis*
  • Demography
  • Female
  • Humans
  • Predictive Value of Tests
  • Pregnancy
  • Pregnancy Complications
  • Pregnancy Trimesters
  • Psychometrics
  • Smoking*
  • Surveys and Questionnaires
  • Truth Disclosure

Substances

  • Carbon Monoxide