Moderators and mediators of a Web-based computer-tailored smoking cessation program among nicotine patch users

The aim of this study was to examine moderating and mediating factors of the efficacy of World Wide Web-based tailored behavioral smoking cessation materials. The design was a two-group randomized controlled trial in England and the Republic of Ireland. Participants were 3971 subjects who purchased a particular brand of nicotine patch and logged on to use a free Web-based behavioral support program. The intervention was Web-based tailored behavioral smoking cessation materials or Web-based nontailored materials. The 10-week continuous abstinence rate was assessed by Internet-based survey at 12-week follow-up. Potential treatment moderators were examined using subgroups of established or possible predictors of smoking cessation. Treatment mediators examined included 6-week follow-up measures of program relevance and amount of the Web-based materials read. Within all subgroups examined, subjects in the Web-based tailored intervention were more likely to report 10-week continuous abstinence at 12-week follow-up. Significant moderators, indicating a significant difference in program efficacy between subgroups, included presence of a tobacco-related illness (larger treatment-control differences among subjects with a tobacco-related illness), presence of nonsmoking children in the household (larger treatment-control differences among subjects with nonsmoking children in the household), and frequent alcohol consumption (larger treatment-control differences among subjects with higher alcohol consumption). Perceived program relevance at 6-week follow-up was a mediator of cessation at 12-week follow-up. Robust results of the tailored program may be explained by the tailoring strategies utilized in the treatment conditions. Moderating variables may be particularly useful to address in tailored messaging. The mediating factor of perceived message relevance may provide a partial mechanism of effective program tailoring.