The effect of the alliance on social recovery outcomes and usage in a moderated online social therapy for first-episode psychosis.

Objectives: We investigated the effect of the therapeutic alliance on both change in social recovery outcomes and usage of a moderated online social therapy platform for first-episode psychosis (FEP), Horyzons.

Design: Secondary analysis of a single group pilot trial.

Methods: Clients completed an alliance measure adapted for guided digital interventions at mid-treatment. A series of multi-level models evaluated change in outcomes by mid- and post-treatment assessments (relative to baseline) as a function of the overall alliance. Quasi-Poisson models evaluated the effect of the overall alliance on aggregated counts of platform usage. Exploratory analyses repeated these models in terms of the bond (human-human) or the task/goal (human-program) alliance.

Results: Stronger overall alliance at mid-treatment predicted lower loneliness at mid-treatment and lower social anxiety at mid- and post-treatment. It was also associated with higher completion of therapy activities and authoring of comments and reactions. A strong bond with an online therapist was associated with lower loneliness and higher perceived social support at mid-treatment, lower social anxiety at post-treatment as well as a higher number of reactions made on the social network. Stronger alliance with the platform's tasks and goals facilitated lower social anxiety at both follow-up assessments and was further associated with higher completion of therapy activities and reactions in the social network.

Conclusions: The alliance may impact aspects of social recovery and usage in digital interventions for FEP. Specific aspects of the alliance (human-human and human-program relationships) should be considered in future research.