CBT and Early Intervention for treatment & relapse prevention in psychosis
Philippa Garety PhD, Professor of Clinical Psychology, Institute of Psychiatry, King’s College, University of London, South London and Maudsley NHS Foundation Trust
Treatments for Psychosis
- Since 1960s, primary treatment for psychosis has been anti-psychotic medication
- Demonstrated effectiveness in symptom reduction and relapse prevention
- But many people have sub-optimal response:
- Persisting symptoms (40-50%)
- Relapses (40-50% in two years; 80% within five years)
- Disabling side effects
- Resultant poor adherence
Cognitive behavioural therapy
- CBT originally developed in 1970s to treat depression and anxiety by Dr Aaron T Beck in Philadelphia, USA
- Focuses on ‘negative thoughts’ which drive negative emotions: the meaning given to events and experiences is what causes distress
- CBT involves collaboratively exploring and testing beliefs about self, world and future – a joint process of enquiry to test out new ways of thinking and behaving
Cognitive behavioural therapy for schizophrenia
- Developed from CBT for emotional disorders in the UK in the early 1990s
- Schizophrenia is experienced as cognitive disorder – a disruption of thinking and perception: these are raw material of CBT
- CBT draws on stress-vulnerability and cognitive models of psychosis
A Cognitive Model of the Positive Symptoms of Psychosis
It is the faulty appraisal of anomalous experiences (e.g. heightened perception; co-incidence of car numberplates; voice saying ‘got you’) and emotional distress, influenced by the cognitive/-emotional biases acquired through social adversity (e.g. people as threatening and hostile; self as bad and vulnerable) and the cognitive biases and impairments arising from biological vulnerability (JTC) , interacting, which result in the psychotic symptom –e.g, they are following me by satellite and planning to kill me
CBT for schizophrenia: main focus
- Collaborative development of an understanding of distressing psychotic experiences
- Re-evaluate appraisal of experiences
- Work on maintaining factors (e.g. reasoning style, self-concept, social isolation, appraisals of psychosis, emotional processes)
Methods of CBT for Psychosis
- Weekly or fortnightly sessions
- Sessions typically lasting 50 minutes
- Nine months’ duration (average of 20 sessions) but may continue for longer or end sooner
- Trained therapists
- Alongside medication and services
CBT for Schizophrenia What evidence for effectiveness?
- During the past 15 years, over 30 RCTs conducted
- Most trials with people with a diagnosis of schizophrenia and persistent positive symptoms, unresponsive to medication;
- Growing number of studies with early psychosis; acute episodes; and ‘prodromal’ or ‘at risk’ patients
- Main outcome measure: standardised assessments of psychiatric symptoms
Psychiatric symptom scores (BPRS)
(Kuipers, Garety, Fowler et al, 1998)
Change in delusions
(Kuipers et al, 1998)
Change in hallucinations
(Kuipers et al, 1998)
A Meta-Analysis of CBT trials
(Pilling et al, 2002)
- Selection of only ‘high quality’ RCTs (e.g. concealed allocation; info on recruitment and attrition; therapy and control clearly defined)
- Use of published data and additional data
- Results extracted meeting certain criteria - e.g. means and standard deviations available; excluding outcomes where more than 50% drop-out/missing
- Comparable data pooled
- Only schizophrenia-spectrum diagnoses
- 13 trials and 1293 patients
- Studies differed in stage – early/acute (2); acute mixed (2); persistent symptoms (6); community (2); day (1)
- Country - UK (10); USA (2); Israel (1)
Number needed to treat (NNT)
- The NNT is the number of people who need to be treated in order to prevent one additional bad outcome
- An NNT of 10 or less is usually regarded as clinically worthwhile, depending on the importance of the outcome and the cost and feasibility of the treatment.
- In CBT for psychosis, NNTs are calculated for (absence of) clinically significant symptom reductions (e.g. people with 20% or 50% reduction in total score).
CBT Numbers Needed to Treat (NNT) – Symptom Reduction
| Compared to standard care | |||
| N | NNT | CI | |
| At end of treatment | 121 | 5 | 4-13 |
| At follow-up | 121 | 4 | 3-10 |
| Compared to ‘other psychological interventions’ | |||
| N | NNT | CI | |
| At end of treatment | 149 | 5 | 3-15 |
| At follow-up | 149 | 7 | 4-100 |
Recent meta-analyses
Zimmerman et al, 2005; Wykes et al, 2007
- Inclusion of more trials (over 30)
- More studies of acute psychosis
- Results confirm conclusions of earlier meta-analysis:
- There is convincing evidence that CBT reduces persistent symptoms and distress which is sustained up to one year post treatment; medium effect size (0.4);
- There is more limited evidence for improvements in mood, relapse, social functioning and self-esteem and that improvements might last five years
- Less evidence for benefits in acute psychosis or for relapse prevention
PRP trial
(Garety, Freeman, Fowler, Bebbington, Dunn and Kuipers, in press; BJPsych)
- To investigate effects of CBT and of FI on relapse prevention, symptoms and social functioning in acute psychosis
- Participants: patients who had recently relapsed and had positive symptoms, and their consenting carers
- 683 satisfied inclusion criteria
- 301 patients and 83 carers participated
- 2 pathways – patients without and with carers
Patient recruitment and randomisation
Demographic Characteristics Patients
Clinical Characteristics and History
Primary Outcomes: Remission, Relapse and Re-admission
- Primary outcome data on 96% of the sample
- There were no differences between the groups, in either pathway, in patterns of recovery and relapse, and in total days in hospital at 12 and 24 months
- Fewer than 50% made a full recovery
- No treatment effects on full or partial recovery
- There were some limited benefits on secondary outcomes
Total Symptoms (PANSS)
(Treatment effect estimates -Difference in Means between treatment group and TAU (95% CI)
Depression (BDI)
(Treatment effect estimates - Difference in Means between treatment group and TAU (95% CI))
Total Symptoms (PANSS)
Treatment effect estimates (difference in means between CBT and FI and TAU) for those participants with carers, N=115
Delusions (Distress)
Treatment effect estimates for those participants with carers
Social Functioning (SOFAS)
Treatment effect estimates for those participants with carers
Summary of PRP findings
- Primary outcomes: no effects on recovery, relapse and readmission;
- Secondary outcomes: only one significant effect, of CBT compared to TAU – reduced depression (BDI) at 24 months; no significant effects of FI.
- In separate analysis of all those with carers, significant effects for CBT compared to TAU on two measures: delusion distress and social functioning at 12 months; no significant effects of FI
Conclusions
- The PRP trial findings do not change the earlier evidence that CBT is less likely to be helpful for acute psychosis and for relapse prevention.
- CBT might confer benefits for depression and social functioning, as well as symptoms, for some patients
Early intervention in psychosis
(McGorry et al 2007)
Three distinct elements:- Prevention
- Early recognition
- Early assertive multi-modal treatment
1. Prevention
- Treatment of ‘at risk’ populations before the onset of psychosis
- A very small number of trials (e.g McGorry et al, 2002; Morrison et al, 2004)
- Treatment involves medication, CBT or both
- Some evidence for reducing hospitalisation or delaying the onset of psychosis, but not yet firm evidence for prevention of onset (Phillips et al, 2007)
2. Early recognition
- Many people with psychosis not detected for many months or even years
- As a result, long delays before treatment is commenced
- ‘Duration of untreated psychosis’ (DUP) robustly associated with poorer outcome
- Strategies to improve detection and decrease DUP – community education; GP (family) doctor training; rapid assessment
- As yet, little evidence of sustained benefits, in terms of long-term outcomes, from these strategies; though providing treatment without delay reduces patient and carer distress
3. Early assertive community treatment – the LEO trial
(Craig et al, 2004; Garety et al, 2006)
- Early assertive multi-modal treatment: applying best practice in treatment and care, aimed at early and effective engagement with services, with goal of improving outcomes
- The LEO trial – RCT of early intervention
- First or second episode of psychosis
LEO Community Team
- Commenced in January 2000
- Multidisciplinary team of 10 clinicians
- Assertive Outreach model
- Provides intensive & assertive follow-up, outreach and crisis intervention of clients
- Integrated treatments with emphasis on functional as well as symptomatic recovery
LEO: Clinical Interventions
- Medication management:
- Atypical antipsychotics; low dose
- Emphasis on adherence
- Family work
- Individual CBT
LEO: Psychosocial Programmes
- Employment & Education
- Access to local colleges e.g. Lambeth College
- Support programs with employment agencies
- Vocational worker & pre-employment programmes
<
- Housing support
- Links & access to local supported housing options
- Social programs
- Weekly evening LEO ‘Drop-in’ for clients
LEO RCT
LEO: Patient Characteristics
| Factor | LEO | Control |
| Age M (sd) | 25.9 (6.0) | 26.6 (6.4) |
| Sex: Male | 55% | 74% |
| Single | 63% | 59% |
| First episode | 83% | 71% |
| White British | 38% | 26% |
| Unemployed | 54% | 51% |
LEO: Provision of Interventions
LEO: Adherence to medication
LEO: Service Engagement at 18 Months
LEO: Engagement by ethnic group at 18 Months
LEO Recovery From Index Episode, (N=123)
LEO Relapse at 18 Months (N=122)
LEO Positive & Negative Symptoms at 18 Months: PANSS (N=99)
LEO Social Functioning at 18 Months: GAF (N=98)
LEO Employment
- LEO participants spent more of the follow up in occupation:
- Leo 7.1 (6.6) vs. Control 3.7 (5.4) mths p < .01.
- Fewer were unemployed throughout.
- 23% LEO vs. 41% control p < .05.
- More LEO participants were in work at the point of follow up (32% vs. 20% n.s.)
LEO: Quality of Life (MANSA)
LEO Dissatisfaction with services at 18 Months (N=67)
Positive Findings for Early intervention - LEO
- Service engagement and satisfaction are higher and the service is better at retaining patients from ethnic minority groups
- Relapse is significantly reduced
- Reported quality of life is better
- Social and vocational functioning improved
- Symptoms: Modest effect only on negative symptoms
Summary
- Early intervention studies are few, but the LEO study suggests that early assertive community treatment, involving multi-modal bio-psycho-social interventions can improve outcomes, especially relapse and social functioning.
- CBT has a large RCT evidence base which shows consistent benefits for symptom reduction, in medication unresponsive patients, and some evidence of improvements in depression and social function, but little evidence for relapse reduction
Conclusions
- Early Intervention for people with newly diagnosed psychosis enhances engagement and relapse reduction; and CBT for people with persistent symptoms reduces distressing symptoms
- Both approaches should be added to medication in the treatment of people with psychosis