x x x x x x

Room 12
Koco Building
Arches Industrial Estate
Spon End
Coventry CV1 3JQ

Tel: 02476 670606

LATEST NEWS! AIMHS will be hosting a MUS Reconvened Workshop with commissioner Eleanor Cappell on the 13th of June, the workshop will commence at 11am at Vicroft Court. AIMHS will be holding is usual co-production workshop at 1pm on the 13th of June in the same venue - guest speakers will be attending from Cov City Council to discuss STP and a new STP website.

Has cognitive behavioural therapy for psychosis been oversold?

22 April, 2014 by Andy

CBT is a recommended treatment for schizophrenia in the UK, but how strong is the evidence that it works? Keith Laws suggests it's not as strong as we might hope.

In 1952, the same year that chlorpromazine was introduced as the first effective drug treatment for schizophrenia, Aaron Beck first employed a form of cognitive-based talk therapy to treat delusional thinking. While chlorpromazine inaugurated the era of drug treatments in psychiatry, Beck’s psychological alternative was a slow-burner. Cognitive Therapy or Cognitive Behavioural Therapy (CBT), as we now know it, became part of the mainstream treatment for psychosis in the UK in 2002 when the National Institute of Clinical Excellence (NICE) endorsed it; and again in 2009, they further recommended that CBT be offered “to all people with psychosis or schizophrenia." Despite repeated endorsements by this Government agency, is it possible that CBT for psychosis has been oversold?

NICE, along with almost everyone else nowadays, 'takes stock' of the results of clinical trials by using the quantitative technique of meta-analysis. This allows findings of different trials to be mathematically summed, giving more weight to larger (and usually methodologically more rigorous) trials. Notably the dozen meta-analyses of trials investigating CBT for schizophrenia document the shrinking evidence for its effectiveness on symptoms. Although initial meta-analyses optimistically suggested that around 50% of patients showed significant symptom reduction following CBT, more recent meta-analyses estimate that just 5% benefit significantly when compared to controls.

A recent meta-analysis, published in the British Journal of Psychiatry, is the largest ever undertaken, examining CBT in 3000 individuals diagnosed with schizophrenia. It revealed a ‘small’ therapeutic effect on key "positive" symptoms such as delusions and hallucinations. The central finding, however, is that even this marginal effect vanishes in studies where symptoms are assessed blind i.e. assessors are unaware if patients were in the therapy or control condition. Indeed, when researchers know that the person received CBT, they are biased to record effects that are up to seven times larger. Another 2014 meta-analysis also documented a small effect of CBT on positive symptoms, but this time it evaporated when the authors controlled for ‘researcher allegiance’. Researchers themselves often display a conscious or unconscious bias that favours and inflates the apparent efficacy of their preferred therapy. In the case of CBT, controlling for such allegiance bias removes any efficacy.

Perhaps the most remarkable fact is that over 75% of published studies document no significant reductions in positive symptoms, negative symptoms or in the key symptom of hallucinations. How can an intervention survive and continue to be promoted if it largely documents no effect? One possibility is that proponents of cognitive therapy simply ignore negative findings. Last month NICE produced a ‘new’ guide re-endorsing CBT for “all people with psychosis and schizophrenia”, although in reality it simply re-packaged their 2009 guide. This means that the 2014 NICE 'update' ignores 11 RCTs published since 2009, 10 of which find no support for CBT as well as 5 newer meta-analyses which estimate CBT efficacy to be less than half of that actually claimed by NICE.

Many see talk therapy not just as an adjunct to medication, but as a potential alternative. This is typified by a paper in last month's Lancet, claiming to show that CBT reduces symptoms in those who have chosen not to take antipsychotic medication. The Guardian ran with "At last a promising alternative to antipsychotics", where Daniel and Jason Freeman critique antipsychotics on the basis of cost, side-effects, and the poor adherence to medication and claim CBT's “efficacy in reducing the symptoms of psychosis was comparable to what medication can achieve”. The BBC website ran similarly with “Talking therapies: effective as drugs” only to later be forced to change their headline to “Talking therapies: moderately effective”; while Science ran with the provocative “Schizophrenia: time to flush the meds?”

The hyperbole in some of these claims warrants examination and the Lancet study provides a perfect opportunity. First, drop out for CBT recipients in the Lancet study was high at 40% during 9 months of treatment and 50% by the follow-up. Second, CBT - and psychotherapy generally - has its own problematic and under-explored side effects; as does doing nothing, which partly represents the Lancet control condition, where many refusing medication were simply discharged. Third, the incidence of serious adverse events reported in the Lancet paper was extremely high at 10% (including 2 deaths in the "treatment as usual" group - which compares poorly with approximately 6 deaths in the 3,000 individuals during 50 CBT + medication trials). Finally, claims that CBT has as great an impact on symptoms as antipsychotic medication is a spurious and unfounded argument. No head-to-head comparison of drugs versus CBT have ever been made and as we have seen, serious doubts exist about CBT having any impact on the symptoms of psychosis.

In this context it is worth asking not whether CBT may become an alternative to medication, but what evidence, if any, would lead CBT advocates to reject their therapy? To admit that CBT has little or no impact on the symptoms of psychosis would have much wider resonance for politicians who have backed it, the many health workers who practice it and would be portrayed as reducing service-user choice. Within the realm of science, issues of expedience, economics and emotion should not be key drivers of mental health care.

Keith R Laws is a Professor of Cognitive Neuropsychology at the School of Life and Medical Sciences, University of Hertfordshire. His research focusses on cognitive problems in Schizophrenia, Alzheimer’s disease and Obsessive Compulsive Disorder.

On Twitter he is @keith_laws.