Skip to main content

19-02-2014 | Mental health | Article

Psychotherapies have specific effects on psychosis symptoms


Free abstract

medwireNews: There are small but robust differences in the effects of different psychotherapies on psychosis symptoms, say the authors of a meta-analysis.

“The differences shown between interventions are small in terms of clinical significance,” say lead researcher David Turner (VU University, Amsterdam, the Netherlands) and team. “This may suggest that the major therapeutic effects of interventions occur through common factors”, in keeping with the theory that all psychotherapies have similar effects.

“However, the pattern of differences in efficacy is consistent with the specific aims of the interventions,” they report in The American Journal of Psychiatry.

The found that cognitive behavioural therapy (CBT), which focuses on reducing positive symptoms through cognitive restructuring, had a larger effect than other therapies on positive symptoms. The difference was small but statistically robust.

CBT was superior to all other interventions (pooled) for all symptoms and positive symptoms, better than befriending for all symptoms and better than cognitive remediation for positive symptoms.

Conversely, social skills training, which aims to improve patients’ social functioning to help them cope with social situations, had the largest effect on negative symptoms.

The researchers identified 48 studies, with a total of 3295 participants, that directly compared psychotherapies in patients with psychosis. Other interventions besides CBT and social skills training were supportive counselling, befriending, cognitive remediation and psychoeducation.

Befriending was the least effective intervention, being inferior to the pooled other interventions for all symptoms. The superior effects of CBT and social skills for positive and negative symptoms, respectively, although small, were significant and sustained in most sensitivity analyses. Analyses in which the associations were lost generally had low statistical power.

However, the team notes that the studies were of very variable quality, especially regarding the risk of bias. Studies involving CBT were most often rated as having no risk of bias (59%), whereas those assessing social skills training were the least often free of bias risk (12.5%), which limited the researchers’ ability to perform sensitivity analyses.

“It is important that future studies on the relative efficacy of social skills training address these issues,” say Turner et al.

Nonetheless, they conclude that their findings “have implications for the continued clinical implementation, design, and improvement of psychosocial therapies for psychosis.”

medwireNews ( is an independent clinical news service provided by Springer Healthcare Limited. © Springer Healthcare Ltd; 2014

By Eleanor McDermid, Senior medwireNews Reporter

Related topics