A randomised controlled trial of a theory-based interactive internet-based smoking cessation intervention ('StopAdvisor'): Study protocol

Background: Internet-based interventions can help smokers to quit compared with brief written materials or no intervention. However, they are not widely used, particularly by more disadvantaged smokers, and there is significant variation in their effectiveness. A new smoking cessation website (‘StopAdvisor’) has been systematically developed on the basis of PRIME theory, empirical evidence, web-design expertise and user-testing with socio-economically disadvantaged smokers. This paper reports the protocol of a randomised controlled trial to evaluate the efficacy of StopAdvisor and determine whether it translates across the social spectrum. Methods: The trial has two arms with participants randomised to the offer of the interactive ‘StopAdvisor’ website (intervention condition) or a non-interactive, static website (control condition). Participants are adults from the UK, who smoke every day and are willing to make a serious quit attempt within a month of enrolment. At least 4260 participants will be recruited with a minimum of 2130 in each of two socio-economic sub-groups. The intervention comprises a structured quit plan and a variety of theory- and evidence-based behaviour change techniques for smoking cessation. Tailored support is offered in the form of a series of tunnelled sessions and a variety of interactive menus for use up to a month before, and then for one month after quitting (http://www.lifeguideonline.org/player/play/stopadvisordemonstration). The control is a static website that presents brief and standard advice on smoking cessation. Assessments are at baseline and 2-, 4- and 7-months post-enrolment. The primary outcome measure will be Russell Standard 6-months sustained abstinence, defined as self-reported continuous abstinence verified by saliva cotinine or anabasine at 7-month follow-up. Secondary outcome measures will include 7-day point-prevalence abstinence at 7-month follow-up, self-reported abstinence at 2- and 4-month follow-ups, satisfaction ratings of the website and quantitative indices of website interaction. All analyses will be by intention to treat and the main analysis will compare the two conditions on the primary outcome measure using a logistic regression model, adjusted for baseline characteristics. The efficacy of the intervention across the social spectrum will be assessed by a logistic regression focusing on the interaction between condition and socio-economic disadvantage.