This study focuses on comparing reports of cigarette consumption using an interactive voice response (IVR) system and the timeline follow-back (TLFB) method. As part of a smoking cessation trial, daily reports of cigarette use were collected using an IVR system and TLFB for 7 days after participants attempted to quit smoking. Reports of cigarette consumption were assessed for the entire sample (N = 381) and for the subset of nonabstinent participants (n = 129 smokers). The results revealed high correlations between reports of smoking behavior, as assessed by IVR and TLFB. Compliance with the TLFB was superior to compliance with the IVR system. This problem should be addressed if researchers wish to use an IVR system in future smoking cessation clinical trials.
Keywordsinteractive voice response; timeline follow-back; smoking; tobacco Long-term abstinence, as measured by point prevalence assessments of smoking behavior, has traditionally been the most common outcome variable used in smoking cessation clinical trials (Fiore et al., 2000;Piasecki & Baker, 2001;Velicer, Norman, Fava, & Prochaska, 1999;Velicer, Prochaska, Rossi, & Snow, 1992). Although point prevalence smoking abstinence is a useful outcome measure, several alternatives have recently been proposed. Reduction in the rate of smoking behavior as an alternative outcome in harm reduction smoking studies (see, e.g., Hughes, 2000, for a review) appears to be a promising approach that will likely receive further study, and in such studies an assessment of daily smoking behavior for the duration of study participation would be preferable to point prevalence measures. The Society for Research on Nicotine and Tobacco (SRNT) formed a work group to review the literature on measures of abstinence in smoking cessation clinical trials (Hughes et al., 2003). In this comprehensive review, the work group's final recommendation was that researchers conducting clinical trials should report multiple measures of abstinence. Clearly, these researchers are advocating that several measures of abstinence be reported in lieu of solely point prevalence, and in these instances (e.g., prolonged abstinence) it would be preferable to measure daily reports of smoking behavior. Daily reporting of smoking may provide a more valid assessment of behavior than longer retrospective reports because it involves reporting over a shorter time period. Shiffman et al.(1997) conducted a study focusing on assessments of cigarette consumption and smoking lapse behavior (e.g., mood ratings) in which comparisons were made between real-time reports on hand-held computers and retrospective accounts provided after 12 weeks. In this study, participants' recall of lapses was poor, and they overestimated their negative affect and the number of cigarettes smoked during a lapse. Shiffman et al. (1997) suggested that caution be used with retrospective recall in research.Interactive voice response (IVR) technology has been shown to reliably assess a variety of daily behaviors (Piette, 2000). With regar...