Precision in progress: a scoping review of SMART designs in adaptive behavioral interventions for individuals with a...

Traditional randomized controlled trials are limited in their ability to inform the development of adaptive interventions tailored to the needs of individuals with a chronic physical and/or mental illness. Sequential Multiple Assignment Randomized Trials (SMARTs) provide a rigorous design for developing and evaluating adaptive interventions.

This scoping review examined the application of SMARTs to develop and evaluate adaptive behavioral interventions for individuals with a chronic physical and/or mental illness.

A search was conducted in 4 electronic databases, using a three-step process following Joanna Briggs Institute recommendations. Records were screened using Rayyan. Narrative synthesis and deductive content analysis were performed.

Fifteen unique SMARTs were found, all examining stage-specific main effects; few evaluated embedded adaptive interventions. Two-stage SMARTs with re-randomization of non-responders only were the most common design. SMARTs typically spanned 12-13 weeks, used a single tailoring variable (often the same as the primary outcome), and defined response using threshold-based criteria. Adaptations for non-responders involved continuing with or adding a new component to the stage 1 intervention, whereas responders continued with the stage 1 intervention or received boosters. Responder rates ranged from 6.1% to 87%. Based on the 12 large SMARTs (excluding pilots), stage 1 and 2 effects were often nonsignificant, and adaptive interventions did not seem to result in the intended improvements in outcomes. Reported challenges included complex logistics, large sample size requirements, and high costs, but SMARTs enabled real-world, dynamic comparisons.

Insights from this review can guide the design of SMARTs to meet diverse population needs.