Validity and reliability of survey items and pictograms for use in a national household survey of tobacco use in Cambodia.

The validity of survey measures of smoked and smokeless tobacco use in the Western Pacific Region is often unknown. We conducted a validation study (n = 201) in a random sample of rural adults in Cambodia. A comparison with salivary cotinine indicated (1) that survey items and pictograms of current tobacco use had an 87% (95% confidence interval [CI] = 78%-93%) sensitivity, 94% specificity (95% CI = 87%-98%), and 93% (95% CI = 85%-97%) positive predictive value in detecting cotinine levels >10 ng/mL; (2) a positive correlation with number of cigarettes smoked (R = 0.34; P = .01); and (3) a positive correlation with the amount of tobacco chewed (R = 0.44; P = .02). The validity of the index for the amount of smokeless tobacco used was enhanced by adding to the index the data from pictograms that were utilized to help participants estimate the amount of loose tobacco used per session. These tobacco items and pictograms were found to have excellent reliability (κ = 0.80-1) over 2 to 3 weeks. Interviewer-administered survey items and pictograms can provide an accurate, quantitative measure of smoked and smokeless tobacco use in rural Cambodia.