Accuracy of self-reported secondhand smoke exposure in NHANES 1988-2002

Background: The workplace remains a significant source of secondhand smoke (SHS) exposure. Occupational health studies often rely on self-reported SHS exposure; the accuracy of self-reports is unknown. This study examines the accuracy of self-reported tobacco use and exposure to SHS validated against serum cotinine levels, a nicotine metabolite.

Methods: Nationally representative data on serum cotinine and self-reported tobacco use and SHS exposure for US workers were extracted from the National Health and Nutrition Examination Surveys (NHANES). NHANES III, 1999-2000, and 2001-2002 surveys were used (n= 13,870). Serum cotinine was categorized into tobacco user (>15 ng/ml), SHS exposed (above the detectable limit but ≤15 ng/ml), and SHS unexposed (undetectable). The percent agreement between the self-reported and cotinine categories was calculated.

Results: Workers reporting tobacco use were 88% accurate while workers reporting work, home, or combination home/work exposures were 92-93% accurate. Workers reporting no SHS exposure were only 30% accurate. Accuracy levels did not vary much among worker groups (e.g. blue/white collar).

Conclusions: Workers accurately report their smoking status; positive reports of workplace and/or home SHS exposures are also accurate. However, large proportions of workers reporting "no exposures" have detectable levels of cotinine in their blood consistent with unreported SHS exposures which extend beyond home and workplace settings and/or reflect a worker inability to detect low-level workplace SHS exposure. Occupational health studies that rely on abbreviated self-reported SHS exposure instruments should be aware of some misclassification, and may consider more comprehensive assessment of these exposures both in and out of the worksite.