Are non-responders in a quitline evaluation more likely to be smokers?

BACKGROUND: In evaluation of smoking cessation programs including surveys and clinical trials the tradition has been to treat non-responders as smokers. The aim of this paper is to assess smoking behaviour of non-responders in an evaluation of the Swedish national tobacco cessation quitline a nation-wide, free of charge service. METHODS: A telephone interview survey with a sample of people not participating in the original follow-up. The study population comprised callers to the Swedish quitline who had consented to participate in a 12 month follow-up but had failed to respond. A sample of 84 (18% of all non-responders) was included. The main outcome measures were self-reported smoking behaviour at the time of the interview and at the time of the routine follow-up. Also, reasons for not responding to the original follow-up questionnaire were assessed. For statistical comparison between groups we used Fischer's exact test, odds ratios (OR) and 95% confidence intervals (CI) on proportions and OR. RESULTS: Thirty-nine percent reported to have been smoke-free at the time they received the original questionnaire compared with 31% of responders in the original study population. The two most common reasons stated for not having returned the original questionnaire was claiming that they had returned it (35%) and that they had not received the questionnaire (20%). Non-responders were somewhat younger and were to a higher degree smoke-free when they first called the quitline. CONCLUSION: Treating non-responders as smokers in smoking cessation research may underestimate the true effect of cessation treatment.