Accuracy of self-reported smoking abstinence in clinical trials of hospital-initiated smoking interventions

Aims: To estimate the prevalence and predictors of failed biochemical verification of self-reported abstinence among participants enrolled in trials of hospital-initiated smoking cessation interventions. Design: Comparison of characteristics between participants who verified and those who failed to verify self-reported abstinence. Settings: Multi-site randomized clinical trials conducted between 2010 and 2014 in hospitals throughout the United States. Participants: Recently hospitalized smokers who reported tobacco abstinence 6 months post-randomization and provided a saliva sample for verification purposes (n = 822). Measurements: Outcomes were salivary cotinine-verified smoking abstinence at 10 and 15 ng/ml cut-points. Predictors and correlates included participant demographics and tobacco use; hospital diagnoses and treatment; and study characteristics collected via surveys and electronic medical records. Findings: Usable samples were returned by 69.8% of the 1178 eligible trial participants who reported 7-day point prevalence abstinence. The proportion of participants verified as quit was 57.8% [95% confidence interval (CI) = 54.4, 61.2; 10 ng/ml cut-off] or 60.6% (95% CI = 57.2, 63.9; 15 ng/ml). Factors associated independently with verification at 10 ng/ml were education beyond high school education [odds ratio (OR) = 1.51; 95% CI = 1.07, 2.11], continuous abstinence since hospitalization (OR = 2.82; 95% CI = 2.02, 3.94), mailed versus in-person sample (OR = 3.20; 95% CI = 1.96, 5.21) and race. African American participants were less likely to verify abstinence than white participants (OR = 0.64; 95% CI = 0.44, 0.93). Findings were similar for verification at 15 ng/ml. Verification rates did not differ by treatment group. Conclusions: In the United States, high rates (40%) of recently hospitalized smokers enrolled in smoking cessation trials fail biochemical verification of their self-reported abstinence.