Improving Adherence to Smoking Cessation Treatment: Smoking Outcomes in a Web-based Randomized Trial

Abstract

Background Partial adherence in Internet smoking cessation interventions presents treatment and evaluation challenges. Increasing adherence may improve outcomes. Purpose To present smoking outcomes from an Internet randomized trial of two strategies to encourage adherence to tobacco dependence treatment components (i.e., skills training, social support, medication): 1) a social network (SN) strategy to integrate smokers into an online community, and 2) free nicotine replacement therapy (NRT). In addition to intent-to-treat analyses, we used novel statistical methods to distinguish the impact of treatment assignment from treatment utilization. Methods N=5,290 current smokers on a cessation website (WEB) were randomized to WEB, WEB+SN, WEB+NRT, or WEB+SN+NRT. The main outcome was 30-day point prevalence abstinence at 3 and 9 months post-randomization. Adherence measures included website metrics of skills training and community use, and self-reported medication use. Inverse Probability of Retention Weighting and Inverse Probability of Treatment Weighting jointly addressed dropout and treatment selection. Propensity weights were used to calculate Average Treatment effects on the Treated. Results Treatment assignment analyses showed no effects on abstinence for either adherence strategy. Treatment utilization analyses revealed that joint community and medication use yielded large percentage point increases in 3-month abstinence rates over skills training alone across arms: WEB=20.6 (95% CI=10.8, 30.4), WEB+SN=19.2 (95% CI=11.1, 27.3), WEB+NRT=13.1 (95% CI=4.1, 22.0), and WEB+SN+NRT=20.0 (95% CI=12.2, 27.7). Abstinence rates among participants that used all treatment components were 25.7%-32.2%. Conclusions Novel propensity weighting approaches can serve as a model for establishing efficacy of Internet interventions and yield important insights about mechanisms. Clinical Trials.gov NCT01544153

A Digital Smoking Cessation Program for Heavy Drinkers: Pilot Randomized Controlled Trial

“…For all smoking outcomes, effect sizes were small, and none approached significance. Smoking outcomes were similar to those reported in a recent trial of smokers from EX [ 48 ], with abstinence rates ranging from 15% to 19% at 1- and 6-month follow-ups. The relatively small effect sizes align with those found in a study of quitline counseling with HD smokers, which showed a statistically significant 3.2% difference that favored the HD intervention [ 15 ].…”

Comparing the Efficacy of an Identical, Tailored Smoking Cessation Intervention Delivered by Mobile Text Messaging Versus Email: Randomized Controlled Trial

“…Several of these RCTs had different incentives to increase participation and decrease the loss to follow-up. This could be multiple follow-ups using the internet, email, or mobile phone if users did not respond [14-17], by payment for mobile phone use [18,19], by free Nicotine Replacement Therapy [15,20], by gift certificates [14], and by internet-based counseling from nurses [21] or tobacco treatment specialists [22]. However, as pointed out by Eysenbach, electronic health (eHealth) research studies with a high dropout or high loss to follow-up should not be looked upon as failures but rather a natural and typical feature of eHealth interventions that should be expected [23].…”

Optimizing Text Messages to Promote Engagement With Internet Smoking Cessation Treatment: Results From a Factorial Screening Experiment

“…A full factorial design was used in the study design phase [ 34 ]. The primary outcome used for sample size calculations was a composite engagement score, with weights given by the regression coefficients of a logistic regression model developed to measure the effects of website engagement on 3-month abstinence rates in the control arm of a previous randomized trial by our group [ 7 ]. This composite engagement score had the advantage of being continuously distributed, even if some of the individual engagement metrics were binary or count data.…”

“…Indeed, low levels of engagement with internet interventions have been documented across a range of health behaviors [ 6 ]. Given the evidence of a dose-response association between engagement and behavior change outcomes [ 7 - 9 ], identifying strategies to promote engagement has been noted as a priority across digital health interventions [ 10 - 12 ].…”