Comparative analysis of two tobacco surveillance questionnaires used in NHANES: accuracy of self-reported smoking status

National Health and Nutrition Examination Survey (NHANES) uses two tobacco use surveillance questionnaires. One is administered during an at-home interview (HI) and the other, during an examination session at a mobile examination center (MEC). NHANES data for the years 1999–2012 were used to evaluate the consistency of responses to tobacco surveillance questions in HI and MEC interview. In addition, accuracy of self-reported smoking status was evaluated. Of those who reported to be daily cigarette smokers during HI, 18.7% reported to be either some-day smokers or nonsmokers during MEC interview. Of those who reported to be some-day cigarette smokers during HI, 22.1% reported to be daily smokers and 17.5% reported to be nonsmokers. Also, 4.1% of those who reported to be nonsmokers during HI reported to be either daily or some-day smokers during MEC interview. Using serum cotinine measurements taken during MEC interview, 27.1% were found to be smokers and 72.9% were found to be nonsmokers. In general, a moderate to high agreement, as measured by the κ statistic, was found between the self-reported responses to tobacco use questions during the home and MEC interviews as well as between smoking statuses based on self-reported and serum cotinine measurements.