Mediators of the Effect of Opt-In Versus Opt-Out Treatment on Smoking Cessation in a Randomized Clinical Trial

Smoking cessation interventions are effective. Most smokers do not receive them, in part because smokers are required to "opt-in". Changing the default (CTD) randomized clinical trial, conducted at a large academic medical center, demonstrated that CTD to providing treatment unless refused (opt-out) improved quit rates. The objective of this study was to evaluate if treatment participation and default-theory based variables explained the effect of opt-out on cessation.

We conducted a secondary analysis of CTD data (N = 739). Descriptive statistics and correlations of the baseline variables were provided. Structural equation models were employed to investigate whether the treatment assignment was causally mediated by a priori proposed mediators including perceived endorsement, counseling participation, and medication utilization.

The mean age of the participants was 51.39 (SD = 14.71) years, among whom 349 (47.23%) were females, and the majority was White (428, 57.92%). The indirect effects from the opt-out treatment arm to counseling participation to smoking cessation (b = 0.201, 95% CI = 0.105 to 0.303), and from opt-out to perceived endorsement to counseling participation to smoking cessation (b = 0.010, 95% CI = 0.002 to 0.022) were significant. The indirect effect from opt-out to perceived endorsement to counseling participation to medication utilization to smoking cessation was marginally significant (b = .001, 95% CI = 0.000 to 0.003).

The opt-out approach enhanced patients' sense that their counselor endorsed and/or recommended treatment, as well as counseling participation. The results partially support our hypothesis that the advantage of Opt-out care would increase treatment participation and would do so by psychological variables thought to underpin the efficacy of defaults.

Our results strengthen evidence for the value of delivering smoking cessation as a default (opt-out rather than opt-in) by illuminating likely mechanisms of action. These include increasing participation in counseling and two default-based theoretical mechanisms, perceived endorsement and implied endorsement.