Examining Longitudinal Stimulant Use and Treatment Attendance as Parallel Outcomes in Two Contingency Management Randomized Clinical Trials.

The primary aim of this study was to examine stimulant use and longitudinal treatment attendance in one 'parallel outcomes' model in order to determine how these two outcomes are related to one another during treatment, and to quantify how the intervention impacts these two on- and off-target outcomes differently. Data came from two multi-site randomized clinical trials (RCTs) of contingency management (CM) that targeted stimulant use. We used parallel multilevel modeling to examine the impact of multiple pre-specified covariates, including selected Addiction Severity Index (ASI) scores, age and sex, in addition to CM on concurrent attendance and stimulant use in two separate analyses, i.e., one per trial. In one trial, CM was positively associated with attending treatment throughout the trial (β=0.060, p<0.05). In the second trial, CM predicted negative urinalysis ((-)UA) over the 12-week treatment period (β=0.069, p<0.05). In both trials, there was a significant, positive relationship between attendance and (-)UA submission, but in the first trial a (-)UA at both baseline and over time was related to attendance over time (r=0.117; r=0.013, respectively) and in the second trial, a (-)UA submission at baseline was associated with increased attendance over time (r=0.055). These findings indicate that stimulant use and treatment attendance over time are related but distinct outcomes that, when analyzed simultaneously, portray a more informative picture of their predictors and the separate trajectories of each. This 'indirect reinforcement' between two clinically meaningful on-target (directly reinforced behavior) and off-target (indirectly reinforced behavior) outcomes is in need of further examination in order to fully exploit the potential clinical benefits that could be realized in substance use disorder treatment trials.