Between-client and within-client engagement and outcome in a residential wilderness treatment group: An actor partner...

We examine aspects of engagement (MacKenzie, 1983) as predictors of longitudinal change in Outcome Questionnaire-45.2 scores (Lambert, Kahler, Harmon, Burlingame, & Shimokawa, 2011) for 68, 18-24-year-old male residents in a 10-bed, open enrollment 90-day residential, substance use treatment program. Engagement was partitioned into within-member, between-member, within-other members, and between-other members' effects. Within-member engagement represented how a group member's score for a week deviated from that member's average engagement score (across all weeks), whereas between-member engagement was the member's average engagement score. Similarly, within-other member engagement represented how the other group members' scores for a week deviated from the other group members' average engagement score (across all weeks), whereas between-other member engagement was the other group members' average engagement score. A 2-level hierarchical linear model showed the interaction of between-member engagement and between-other member engagement was related to decreasing OQ-45 scores. When other group members generally saw the group as more engaged, higher group member average engagement ratings were related to improvement. There was a significant interaction between within-member engagement and between-member engagement in predicting OQ-45 scores. When clients generally saw the group as more engaged, weeks with relatively more member engagement, compared with other weeks, were associated with improvement in OQ-45 scores. However, when clients generally saw the group as less engaged, weeks with relatively more group member engagement, compared with other weeks, were associated with greater deterioration in OQ scores. We recommend tracking week-to-week changes in member and other member engagement to identify group members who are not getting optimal program benefits. (PsycINFO Database Record