Cognitive behavioural therapy, or CBT, has in the past decade become a standard tool for helping people with schizophrenia deal with their symptoms. Recent developments, however, raise serious questions about how effective this talking therapy really is for this illness.
Despite strongly advocating that all people with schizophrenia should be offered CBT, the National Institute for Health and Care Excellence (NICE) – the organisation that evaluates treatments for the NHS – has inexplicably neglected to update its evidence base since 2008. In the intervening years the number of randomised controlled trials – the gold standard for clinical research – assessing symptom reduction through CBT has doubled, and many of them cast doubt on the institute’s recommendations.
Very few trials had reported on the impact of CBT on social and professional functionality, or quality of life, so with this in mind we recently published the first meta-analysis assessing the impact of CBT on these aspects.
The results were not positive. While we found that CBT initially improved the ability of patients to function, the benefit was shortlived. CBT failed to alleviate the distress associated with the symptoms of psychosis. CBT also failed to improve quality of life, as measured in a patient’s sense of self, hope, well-being, relationships and so on. Indeed, not one CBT trial has ever reported a rise in quality of life for people diagnosed with schizophrenia.
A growing weight of evidence
This was just the latest study to raise question marks, and a 2014 meta-analysis by our research group had already concluded that claims that CBT alleviates classic symptoms such as delusions are “no longer tenable”.
A 2014 meta-analysis from the Netherlands found that almost none of the latest studies reported reductions in schizophrenia symptoms, such as social withdrawal, apathy, or “emotional blunting” (having no positive or negative emotions). The authors concluded: “CBT studies focused on psychotic symptoms might not work as well in reducing negative symptoms as previously thought.”
Furthermore, a Cochrane review, an authoritative voice on evidence-based healthcare, concluded that CBT showed “no clear and convincing advantage” over other, sometimes much less sophisticated, therapies, or even simple, non-technical approaches such as befriending. This involves talking with the patient about neutral topics of interest, such as music, sport, books, pets and so on.
A smaller amount and lower quality of evidence was required to establish CBT as an intervention, than now exists for it to be rejected. Earlier trials of CBT for schizophrenia – such as those reported by NICE – were much less rigorous than their modern counterparts. For example, many early researchers did not use blind outcome assessment – that is to say, they knew which of their subjects had received CBT and which had not, potentially leading to confirmation bias. These earlier trials spuriously inflated the apparent benefits of CBT five or six times over.
Risk and reward
One of the main factors that can lead to the withdrawal of an intervention is if harm is seen to outweigh benefit. Psychological interventions such as CBT are often assumed to cause no harm, but a recent study urged caution. It warned: “The measurement and reporting of adverse effects in trials of psychological interventions for psychosis (and other conditions), is extremely poor.”
The fact that harm is not routinely assessed, or is poorly assessed in psychotherapy trials should raise a red flag over recent calls for CBT to be an alternative to antipsychotic medication. The single recent, relevant study comparing the two treatments showed that adding CBT to antipsychotic medication gave no significant additional benefit, while adding antipsychotic medication to CBT produced significant improvement in symptoms.
But even setting aside efficacy and potential harm, do the patients themselves actually want CBT? Evidence suggests not – according to the 2014 National Audit of Schizophrenia published by the Royal College of Psychiatrists, more than half of those offered CBT decline it.
The evidence used by NICE is low quality and outdated, and their endorsement of CBT is in dire need of reconsideration. There are serious doubts as to whether CBT actually reduces schizophrenia symptoms, and now also whether it improves key outcomes such as functionality, quality of life and symptom-related distress. If we want psychological interventions to evolve, then new research to be directed at developing and assessing alternative treatments.
And if we want an accessible, cost-effective and equally potent alternative in the meantime, why not listen to the Cochrane group. We might do just as well with befriending.