Cognitive Behavioural Therapy for Psychosis: A Health Technology Assessment.

Abstract

Background: Cognitive behavioural therapy (CBT) for psychosis is a distinct type of psychotherapy that has been recommended together with antipsychotic drugs and comprehensive usual care in the management of schizophrenia, a complex mental health disorder associated with a high economic and societal burden. The objectives of this report were to assess the effectiveness, harms, cost-effectiveness, and lived experience of CBT for psychosis in improving outcomes for adults with a primary diagnosis of schizophrenia.

Methods: We performed literature searches on March 28 and April 5, 2017, and undertook a qualitative synthesis of systematic reviews of the clinical and economic literature comparing CBT for psychosis with any comparator interventions (e.g., usual care, waitlist control, or pharmacotherapy) in adults with a diagnosis of schizophrenia as defined by any criteria (including related disorders such as schizoaffective disorder).We developed an individual-level state-transition probabilistic model for a hypothetical cohort of adults aged 18 years and older starting with first-episode psychosis. We compared three strategies: usual care, CBT for psychosis by physicians, and CBT for psychosis by regulated nonphysician therapists. The CBT was provided in person together with usual care including pharmacotherapy: 16 structured sessions (individual or group) for first-episode psychosis and 24 individual sessions for relapse or treatment-resistant disease. We calculated incremental cost-effectiveness ratios (ICERs) over 5 years using the Ontario Ministry of Health and Long-Term Care perspective and a discount rate of 1.5%. We also estimated the 5-year budget impact of publicly funding CBT for psychosis in Ontario.In addition, we interviewed 13 people with lived experience of schizophrenia and psychosis about their values and preferences surrounding CBT and other treatments.

Results: CBT for psychosis compared with usual care significantly improved overall psychotic symptoms (standard mean difference [SMD] -0.33, 95% confidence interval [CI] -0.45 to -0.21), positive symptoms (e.g., hallucinations) (SMD -0.34, 95% CI -0.58 to -0.10), auditory symptoms (SMD 0.39, 95% Cl not reported, P < .005), delusions (SMD 0.33, 95% CI not reported, P < .05) and negative symptoms (e.g., blunt affect) (SMD -0.32, 95% CI -0.59 to -0.04) at end of treatment. No significant differences were observed for social function, distress associated with psychosis, relapse, or quality of life.Compared with any control, CBT for psychosis significantly improved overall psychotic symptoms, positive symptoms, auditory hallucinations, delusions, and negative symptoms. Compared with other forms of therapy, CBT for psychosis showed inconsistent results at end of treatment for overall psychotic symptoms, positive symptoms, auditory hallucinations, and delusions. In people with first-episode psychosis, CBT for psychosis was not significantly more effective for the prevention of relapse when compared with other forms of therapy or usual care (odds ratio [OR] 1.11, 95% CI 0.63-1.95 and OR 1.15, 95% CI 0.65-2.04, respectively).Low-intensity CBT for psychosis (fewer than 16 face-to-face sessions) compared with any type of treatment significantly improved overall psychotic symptoms and social function at follow-up (SMD -0.40, 95% CI -0.74 to -0.06 and SMD -0.57, 95% CI -0.81 to -0.33, respectively).In the cost-utility analysis, CBT for psychosis provided by nonphysician therapists compared with usual care was associated with increases in both quality-adjusted life-years (mean 0.1159 QALYs, 95% credible interval [CrI] 0.09-0.14) and costs (mean $2,494, 95% Crl $1,472-$3,544), yielding an ICER of $21,520 per QALY gained. CBT for psychosis provided by physicians was dominated because it was equally effective but more expensive (mean $2,976, 95% CrI $2,822-$3,129; ICER of CBT for psychosis vs. usual care: $47,196/QALY gained).Assuming a 20% increase in access per year (from 0% at baseline to 100% in year 5), we estimated the total 5-year net budget impact of publicly funding CBT for psychosis would be about $15.2 million for nonphysician providers and about $35.4 million if provided by psychiatrists. It is estimated that by the year 2021, approximately 110 nonphysician therapists or 150 physicians would be needed to provide CBT for psychosis to more than 12,000 adults with schizophrenia (including about 8,500 incident cases) in Ontario.People with schizophrenia and their family members reported positive experiences with CBT for psychosis. They felt it provided effective tools to help manage their schizophrenia but stressed that it was only effective in conjunction with medication to control psychotic episodes and overcome a patient's denial of illness. Geographic and financial barriers have restricted access to this psychotherapy.

Conclusions: Compared with usual care or any control, CBT for psychosis significantly improved psychotic symptoms, based on evidence of moderate to adequate quality; no significant improvements were observed for social function, relapse, or quality of life outcomes. People affected by schizophrenia reported that CBT for psychosis was valuable in conjunction with antipsychotic medication but that access to this type of psychotherapy is limited. Adding CBT for psychosis to usual care in the management of adult schizophrenia probably represents good value for money in Ontario. Depending on the type of provider, therapy format, and rate of access, the net budget impact to Ontario's publicly funded health system would likely be between $15 million to $35 million over the next 5 years.