Assessing the effectiveness of CBTp across time : a systematic review and meta-analysis

Aim: Research to date has shown small effects of Cognitive Behavioural Therapy for Psychosis (CBTp) on reducing psychotic symptoms. Some have subsequently questioned whether CBTp should continue to be a recommended treatment for psychosis. There have been parallel advances in the understanding of psychosis and in the evolution of CBT that is specific to psychosis. The aim of this study is to examine whether these parallel advances have led to an improved effectiveness of CBTp across time. Method: The design of this study is a systematic review and meta-analysis. MEDLINE, EMBASE, PsycINFO, and CENTRAL were searched for randomised controlled trials examining CBTp interventions targeting positive and/or negative symptoms versus treatment as usual. Four meta-analyses were carried out to examine the effectiveness of CBTp for positive symptoms, delusions, hallucinations, and negative symptoms, with four meta-regressions to examine whether the effectiveness of CBTp for these symptoms improved across time. Result: A total of 28 studies (n = 2698) were included in the meta-analysis of positive symptoms which yielded a pooled g of -0.24 (95% CI -0.32, -0.16, P < .001) favouring CBTp, with non-significant heterogeneity (Q = 26.87, P = 0.47; I2 = 0%). A total of 13 studies (n=890) were included in the meta-analysis of delusions which yielded a pooled g of -0.36 (-0.59, -0.13, P = 0.002), with substantial heterogeneity (Q = 31.99, P = 0.001; I2 = 62%). A total of 16 studies (n = 849) were included in the meta-analysis of hallucinations which yielded a pooled g of -0.26 (95% CI -0.42, -0.11, P < 0.001), with non-significant heterogeneity (Q = 18.10, P = 0.26; I2 = 17%). A total of 19 studies (n = 1761) were included in the meta-analysis of negative symptoms which yielded a pooled g of -0.22 (95% CI -0.33, -0.12, P < 0.001), with non-significant heterogeneity (Q = 20.32, P = 0.32, I2 = 11%). Meta-regressions indicated a significant effect of year on the effectiveness of CBTp only for delusions (F(1, 11) = 5.99, p = 0.032; R2 = 0.594). Conclusion: The findings indicate small-to-medium effects of CBTp for positive symptoms, delusions, hallucinations, and negative symptoms, and that over time, there has been an improvement in the effectiveness of CBTp for delusional symptoms.