Measures of abstinence in clinical trials: issues and recommendations.

A workgroup formed by the Society for Research on Nicotine and Tobacco reviewed the literature on abstinence measures used in trials of smoking cessation interventions. We recommend that trials report multiple measures of abstinence. However, at a minimum we recommend that trial: (a) report prolonged abstinence (i.e., sustained abstinence after an initial period in which smoking is not counted as a failure) as the preferred measure, plus point prevalence as a secondary measure; (b) use 7 consecutive days of smoking or smoking on > or = 1 day of 2 consecutive weeks to define treatment failure; (c) include non-cigarette tobacco use, but not nicotine medications in definitions of failure; and (d) report results from survival analysis to describe outcomes more fully. Trials of smokers willing to set a quit date should tie all follow-ups to the quit date and report 6- and/or 12-month abstinence rates. For these trials, we recommend an initial 2-week grace period for prolonged abstinence definitions; however, the period may vary, depending on the presumed mechanism of the treatment. Trials of smokers who may not be currently trying to quit should tie follow-up to the initiation of the intervention and should report a prolonged abstinence measure of > or = 6-month duration and point prevalence rates at 6- and 12-month follow-ups. The grace period for these trials will depend on the time necessary for treatment dissemination, which will vary depending on the treatment, setting, and population. Trials that use short-term follow-ups (< or = 3 months) to demonstrate possible efficacy should report a prolonged abstinence measure of > or = 4 weeks. We again recommend a 2-week grace period; however, that period can vary.