Cognitive Behavioral Therapy for Psychosis: Predictors of Disengagement

Cognitive behavioral therapy (CBT) for psychosis has a high disengagement rate, and certain factors related to behavior, substance misuse, problems with occupation and activities, and the complexity of problems may predict disengagement, according to a study published in British Journal of Clinical Psychology.

This retrospective study included 103 individuals of varying gender, age, and ethnicity. All participants had a psychosis diagnosis and were referred for high- or low-intensity CBT therapy. The Health of the National Outcome Scales (HoNOS) was used to assess complexity and severity of mental health. The association of demographic and mental health variables with disengagement from CBT was examined using χ² or multivariate analysis of variance tests.

Overall, patients with psychosis had a 42.7% disengagement rate (n=44). HoNOS scores were higher in individuals who disengaged (14.5 vs 11.7; P <.05). When individual HoNOS items were evaluated, individuals with overactive, aggressive, disruptive, or agitated behavior were found to disengage at a rate of 54.9% vs 30.8% of patients without these symptoms (P <.01).

Other HoNOS items predicting disengagement included substance misuse (61.1% vs 32.8%; P <.01), depressed mood (56% vs 30.2%; P <.01), and greater problems with occupation and activities (53.3% vs 34.5%: P <.05). Gender and ethnicity were not found to be significant predictors.

The mean age of individuals who disengaged from CBT trended lower compared with those who completed therapy (33.8 years vs 40.5 years; P <.05). Association between individual clinical diagnosis and disengagement was not examined due to small sample size. Psychotic symptoms, including hallucinations and delusions, were not found to be significant predictors.

Limitations of this study include small sample size, a pragmatic definition of disengagement, as well as the retrospective nature of the study. Additionally, some of the participants had primary affective disorder diagnoses, leading to a potential focus on mood along with psychotic symptoms. 

The results of this study suggest behavioral and emotional factors are predictive of CBT disengagement rather than psychotic ones. These findings are in line with other research linking depression, alcohol, and drug abuse with disengagement of therapy. The association between disturbed behaviors and disengagement is novel and requires further study due to the small sample size. 

The investigators recommend these predictive variables be routinely screened at referral or assessment and targeted prior to or early in therapy to increase the likelihood of engagement.

Reference
Richardson T, Dasyam B, Courtney H, et al. Predictors of disengagement from cognitive behavioural therapy for psychosis in a National Health Service setting: A retrospective evaluation [published online June 10, 2019]. Br J Clin Psychol. doi: 10.1111/bjc.12222