A dyadic approach of assessing the therapeutic alliance in youth mental health and addiction treatment.

When studying therapeutic alliance, researchers typically use data with a 'one-with-many' (OWM) nested structure in which each therapist ('one') treats several clients ('many'). The OWM structure offers the opportunity to assess the interdependence in dyadic therapeutic alliance data by disentangling three components: (1) the client's tendency to generally rate the alliance with his therapists ('client perceiver' component); (2) the client's rating of the unique alliance with his therapist ('client relationship' component); and (3) the therapist's characteristics that elicit a consistent alliance rating in his clients ('therapist partner' component). Similarly, three such components can be distinguished for the therapists' alliance ratings. We used the reciprocal OWM design to distinguish these alliance components, pertaining to the first therapeutic session of two hundred and three youths aged 13 to 23 years who entered a youth mental health and youth addiction treatment facility. We found that most of the alliance variance pertained to the relationship component (client-rated alliance: 99.2%; therapist-rated alliance: 56.0%), with virtually no evidence of therapist's characteristics that elicited a consistent (high or low) alliance rating in his clients. Clients who reported a uniquely strong alliance with their therapist were more likely to have a favorable 4-months treatment outcome (Odds ratio 1.10; p < 0.001). None of the other (client/therapist) alliance components were related to treatment outcome, nor did we find evidence of dyadic reciprocity, which would occur when a therapist would report a particularly strong alliance with the client if that client reported a particularly strong alliance with the therapist, and vice versa.