Moderators of treatment response to exercise in participants with stimulant use disorder: Exploratory results from the Stimulant Reduction using Dosed Exercise (STRIDE)CTN-0037 study

Abstract Exercise is a promising treatment for stimulant use disorder. However, efficacy has not been clearly demonstrated in a general stimulant using population where response to exercise is expected to be heterogeneous. Thus, examination of response heterogeneity to identify subgroups for whom exercise is either clearly indicated or not indicated is of considerable interest as findings will support more effective tailoring of patient treatments in practice and guide future research in stimulant use disorder. A secondary analysis of the Stimulant Reduction Intervention using Dosed Exercise (STRIDE) randomized controlled trial of 302 stimulant using or dependent participants was conducted to identify baseline clinical and demographic characteristics associated with differential response between participants in the exercise and health education control groups. Characteristics (i.e., moderators of treatment response) were identified using an established Best Approximating Modeling (BAM) method. Six moderators of treatment response were identified: Quick Inventory of Depressive Symptomatology-Clinician (QIDS-C) rated total score, exercise test maximum systolic blood pressure, number of lifetime drug treatments, Stimulant Craving Questionnaire (STCQ) total score, Addiction Severity Index (ASI) Family subscale score, and Cognitive and Physical Functioning Questionnaire (CPFQ) total score. For all moderators, the odds ratio of response to exercise vs. health education ranged from 0.32 to 2.52 or more depending on the level of the moderator. These results demonstrate that it is possible to identify pre-treatment patient characteristics that predict statistically and clinically meaningful differential treatment response to exercise.