This study aimed to evaluate the efficacy of cognitive bias modification (CBM) for smoking behaviors using a pre-registered two-stage Bayesian individual participant data meta-analysis. Overall, 19 randomized controlled trials (N = 2,727) were included. Outcomes included CBM’s effect on smoking-related cognitive biases, cigarette consumption, and relapse prevention among smokers motivated to reduce or quit. Pre-specified moderators examined whether CBM effects varied by training type, training congruency, control condition (passive, sham, or active), delivery context (face-to-face vs. online), and additional therapy. Results provided moderate-to-strong evidence against an overall effect of CBM across outcomes compared to control training. Moderator analyses indicated that effects were context-dependent: small favorable outcomes emerged when CBM was combined with minimal psychological support, and face-to-face delivery was associated with greater smoking reduction than online formats. Among CBM variants, approach bias modification (ApBM) showed the most favorable effects on consumption, although evidence for moderation attenuated after covariate adjustment. Overall, CBM did not demonstrate reliable standalone benefits for smoking cessation. Observed effects were weak and dependent on implementation context, offering limited support for CBM as a stand-alone intervention. Future research should prioritize integration with structured psychological treatment and clarify the conditions under which CBM may provide incremental benefit.
Pan et al. (2026) studied this question.