Therapeutic alliance predicts mood but not alcohol outcome in a comorbid treatment setting

The pan-theoretical variable of alliance has been consistently reported to have a moderate yet robust effect on psychotherapy treatment outcome. However, the relationship is less clear in the addiction field where there is more limited research. The current study investigated the relationship between alliance and treatment outcome in an alcohol dependent and depressed population.

The Treatment Evaluation of Alcohol and Mood (TEAM) study was a randomized controlled pharmacotherapy trial with supportive clinical case management. Therapist and client alliance ratings were assessed using the Working Alliance Inventory (WAI) for 123 client-therapist dyads at 3 weeks. Outcome data was obtained at 3 and 12 weeks (end of treatment). Drinking-related measures included Percent Days Abstinent (PDA) and Drinks per Drinking Day (DDD). Mood outcomes were scores on the Montgomery Asberg Depression Rating Scale (MADRS).

Clients rated alliance significantly higher than did therapists and client and therapist ratings were not associated with each other. Baseline motivation was the only pre-treatment client variable associated with alliance, the higher the client's Readiness to Change Questionnaire-Treatment Version (RCQ-TV) score, the higher the therapist-rated alliance. Higher therapists' ratings of alliance were significantly associated with improved mood outcomes at the end of treatment but, with one minor exception, were not related to drinking outcomes. Therapist alliance was also significantly associated with treatment completion. In contrast, client-rated alliance was not related to mood or drinking outcomes, possibly due to a ceiling effect. Subscale analysis found that of the different components that comprise the alliance concept, the task component was most important for drinking outcomes whereas the task and goal components were equally important for mood outcomes. Controlling for early symptom change did not meaningfully alter associations between therapist alliance and mood. In contrast, the strength of associations between therapist alliance and drinking outcomes was reduced for PDA and DDD 12-week change scores, whereas the association between the therapist alliance and 12-week PDA became significant when previously this had not been the case.

Therapeutic alliance was associated with improved mood outcomes, which is consistent with other research. However, alliance, as measured by the WAI, and drinking outcomes, were not related. Findings from these investigations signal the need to re-examine the concept and measurement of alliance in substance-using treatment populations, particularly with regard to drinking outcomes. Within this re-examination, findings support a greater focus on the therapists' role in the alliance-outcome relationship.