Misclassification of smoking status in the CARDIA study: a comparison of self-report with serum cotinine levels

Am J Public Health. 1992 Jan;82(1):33-6. doi: 10.2105/ajph.82.1.33.


Background: Although widely used in epidemiological studies, self-report has been shown to underestimate the prevalence of cigarette smoking in some populations.

Methods: In the CARDIA study, self-report of cigarette smoking was validated against a biochemical marker of nicotine uptake, serum cotinine.

Results: The prevalence of smoking was slightly lower when defined by self-report (30.9%) than when defined by cotinine levels equal to or greater than 14 ng/mL (32.2%, P less than .05). The misclassification rate (proportion of reported nonsmokers with cotinine levels of at least 14 ng/mL) was 4.2% and was significantly higher among subjects who were Black, had a high school education or less, or were reported former smokers. Possible reasons for misclassification include reporting error, environmental tobacco smoke, and an inappropriate cutoff point for delineation of smoking status. Using self-report as the gold standard, the cotinine cutoff points that maximized sensitivity and specificity were 14, 9, and 15 ng/mL for all, White, and Black subjects, respectively. The misclassification rate remained significantly higher in Black than in White subjects using these race-specific criteria.

Conclusions: Misclassification of cigarette smoking by self-report was low in these young adults; however, within certain race/education groups, self-report may underestimate smoking prevalence by up to 4%.

Publication types

  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Adolescent
  • Adult
  • Bias
  • Biomarkers / blood
  • Black or African American
  • Confounding Factors, Epidemiologic
  • Cotinine / blood*
  • Educational Status
  • Female
  • Humans
  • Male
  • Prevalence
  • ROC Curve
  • Reproducibility of Results
  • Sensitivity and Specificity
  • Seroepidemiologic Studies
  • Smoking / blood
  • Smoking / epidemiology*
  • Surveys and Questionnaires / standards*
  • Truth Disclosure
  • White People


  • Biomarkers
  • Cotinine