Analysis of self-reported versus biomarker based smoking prevalence: methodology to compute corrected smoking prevalence rates

Context: Prevalence of smoking is needed to estimate the need for future public health resources.

Objective: To compute and compare smoking prevalence rates by using self-reported smoking statuses, two serum cotinine (SCOT) based biomarker methods, and one urinary 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL) based biomarker method. These estimates were then used to develop correction factors to be applicable to self-reported prevalences to arrive at corrected smoking prevalence rates.

Materials and methods: Data from National Health and Nutrition Examination Survey (NHANES) for 2007–2012 for those aged ≥20 years (N?=?16826) were used.

Results: Self-reported prevalence rate for the total population computed as the weighted number of self-reported smokers divided by weighted number of all participants was 21.6% and 24% when computed by weighted number of self-reported smokers divided by the weighted number of self-reported smokers and nonsmokers. The corrected prevalence rate was found to be 25.8%.

Discussion and conclusions: A 1% underestimate in smoking prevalence is equivalent to not being able to identify 2.2 million smokers in US in a given year. This underestimation, if not corrected, could lead to serious gap in the public health services available and needed to provide adequate preventive and corrective treatment to smokers.