Transporting the effect of the ASSIST school-based smoking prevention intervention to the Smoking, Drinking and Drug Use Among Young People in England Survey (2004-2021): A secondary analysis of a randomized controlled trial

AimsTo conduct exploratory analyses into the transported effect of the ASSIST (A Stop Smoking in Schools Trial) school-based smoking prevention intervention on weekly smoking in young people between 2004 and 2021.DesignSecondary analysis of a cluster randomized control trial (cRCT).SettingEngland and Wales.ParticipantsASSIST trial participants comprised 8756 students aged 12-13 years in 59 schools assigned using stratified block randomization to the control (29 schools, 4193 students) or intervention (30 schools, 4563 students) condition. The target population was represented by 12-13-year-old participants in the Smoking, Drinking and Drug Use Among Young People in England Survey (SDDU) in 2004 (n = 3958), 2006 (n = 3377), 2014 (n = 3145), 2016 (n = 4874) and 2021 (n = 3587), which are randomly sampled school-based surveys with student response rates varying between 85% and 93%.Intervention and comparatorThe ASSIST intervention involved 2 days of off-site training of influential students to encourage their peers not to smoke over a 10-week period. The control group continued with their usual education.MeasurementsThe outcome was the proportion of students who self-reported weekly smoking 2 years post-intervention.FindingsThe prevalence of weekly smoking at the 2-year follow-up in the ASSIST trial in 2004 was 4.1%, 49.5% of students were girls, and 7.8% ethnic minorities. In the SDDU in 2004, the prevalence of weekly smoking was 3.6%, 47.6% students were girls and 14.4% ethnic minorities and in 2021 0.2% were weekly smokers, 48.6% girls and 27.8% ethnic minorities. The odds ratio of weekly smoking in the ASSIST trial in 2004 was 0.85 [95% confidence interval (95% CI) = 0.71-1.02]. The estimated odds ratio in the SDDU target population in 2004 was 0.90 (95% CI = 0.72-1.13), in 2014 was 0.89 (95% CI = 0.70-1.14), and by 2021 was 0.88 (95% CI = 0.60-1.28). The confidence interval ratio was used to estimate precision in the transported estimates in the target population and was 1.57 in 2004, 1.63 in 2014 and 2.13 in 2021, reflecting increasing uncertainty in the effect of ASSIST over time. Subgroup analyses showed effects were comparable when restricted to only English schools in the ASSIST trial.ConclusionsThese exploratory analyses indicate the effect of the ASSIST school-based smoking prevention intervention reported in the original trial may not have been replicated in the target population over the 17-year period of its licensing and roll out.