• Charisse

MOST/SMART Intervention Designs

I've recently been exposed to the multiphase optimization strategy (MOST) and the sequential multiple assignment randomized trial (SMART), which are methods for building and evaluating interventions see more here.

My personal concerns about MOST are with regards to the high burden on researchers and the high N number you would need to get adequate power, especially if using a full factorial design during the optimization phase (although SMART would help ameliorate this issue to a large extent).

Adaptive designs can be very interesting for an psychotherapy intervention like Emotion Regulation Therapy. In psychotherapy interventions, it is possible that some participants may already be proficient at 1-2 of the skill components, but perhaps not the others, and therefore it might reduce participant and researcher burden if one or two of the therapy components could be skipped over. Using a SMART design would also allow researchers to better understand the interactions between different components of the treatment.

eHealth certainly lends itself well to these trial designs, particularly as the different components can be programmed beforehand so as to reduce burden on the research team -- using these designs for non-internet-based therapies with many components sounds like an absolute nightmare! I'm definitely interested in being involved in designing a MOST/SMART intervention.