Aim
The moderate association between therapeutic alliance (TA) and psychological therapy outcome is well established. Historically, the field has not focused on people with a severe mental illness. This is the first review to conduct a meta‐analysis of associations between TA and therapeutic engagement as well as outcome in psychological therapy for psychosis.
Eligibility criteria
Eligible studies conducted a quantitative investigation of the relationship between TA during a psychological therapy and outcome at a subsequent time‐point.
Method
A systematic review examined the relationship between TA and engagement as well as outcome measures within psychological therapy for psychosis. Correlational meta‐analyses using an aggregate random effects model were conducted.
Results
Twenty‐four studies were eligible for inclusion (n = 1,656) of which 13 were included in the meta‐analyses. Client‐ and therapist‐rated TA were associated with engagement in therapy (rclient (c) = 0.36, p = .003; rtherapist (t) = 0.40, p = .0053). TA was also associated with reduction in global (rc = 0.29, p = .0005; rt = 0.24, p = .0015) and psychotic symptoms (rc = 0.17, p = .0115; rt = 0.30, p = .0003). The systematic review identified no evidence or limited evidence for a relationship between TA during therapy and depression, substance use, physical health behaviours, global as well as social functioning, overall mental health recovery, and self‐esteem at follow‐up. Although number of studies was small, TA was related to a reduced risk of subsequent hospitalization in 40% of analyses (across two studies) and improved cognitive outcome in 50% of analyses (across three studies).
Conclusions
The observed TA‐therapy engagement and TA‐outcome associations were broadly consistent with those identified across non‐psychotic diagnostic groups. Well‐powered studies are needed to investigate the relationship between TA and process as well as outcome in psychological therapy for psychosis specifically.
Practitioner points
This is the first review to conduct a meta‐analytic synthesis of the association between therapeutic alliance (TA) and both engagement and change in outcome in psychological therapies for psychosis.
TA (as rated by therapist and client) was associated with the extent of therapeutic engagement as well as reduction in global mental health symptoms and psychotic symptoms.
The significant associations between TA and engagement as well as change in outcome identified in the current review are broadly consistent with those observed across non‐psychotic diagnostic groups.
We consider factors that could impact upon the dynamic and potentially interdependent relationships between TA and therapeutic techniques, including attachment security and severity of paranoid ideation.