Validity of self-reported smoking status using biologic markers (pilot phase of Azar cohort study)

Abstract

Background: Accurate self-reported questionnaires to assess smoking status are conceptually important to evaluate associations between smoking status and disease outcomes in population-based studies. Noticing the importance of measuring smoking status in population level, this study was conducted to assess the validity of self-reported smoking status using biologic markers for the pilot phase of the AZAR Cohort study. Methods: In this cross-sectional study, all 35-70 years old people living in Khameneh, a city in East Azarbaijan, Iran were invited by informed consent to take part in the pilot phase of the AZAR Cohort study. A total of 952 people completed the self-reported smoking status questionnaire. The validity of self-reported smoking status was obtained by measuring urine cotinine. Chi-square test and Kappa agreement index were used to measure the association between smoking status and urine cotinine. Data analysis was done using SPSS software version 23. Results: In this study, 952 subjects aged between 35 and 70 years old (mean and standard deviation: 52, ±8.6). The majority of included subjects were 45-55 years old (42.1%), female (54.2%), married (92%), housewives (40%) and had passed the primary education till diploma degree (38.9%). Non-smokers had the highest self-reported smoking (58.1%). Our study showed that the prevalence of self-reported smoking was 15.8%, of which 34% were men and 0.2% were women; therefore it was significantly higher in men than women (P <0.001) based on Chi-square test. The age range of trying the smoking for the first time starts from 7 years old up to 50 years old and the highest rate of smoking was reported in the age group of 55 to 65 years old (21.7%). Based on the cut-off point of 50 ng/ml, the sensitivity and the specificity of the self-reported questionnaire (95% confidence interval) in total were respectively 27.7% (21.5% -34.7%) and 95.8% (85.8% -% 99.5), in men were respectively 48.1% (38.4% -58%) and 91.3% (98.9% -0.72%). With two cutting points of 50 and 550 ng/ml the Kappa index was obtained in total and men 25.29% and 30.56% respectively. Conclusion: Our findings yielded that self-reported smoking doesn’t show the real condition of smoking in target population. In the study, incorrect categorization is observed in all of the groups and indirect exposure to cigarette smoke can’t be a realistic indication; therefore, it seems that self-reporting alone doesn’t show a proper picture of the smoking condition and needs to use alternative markers in this field.