A new report has been published today by the British Psychological Society Division of Clinical Psychology entitled ‘Understanding psychosis and schizophrenia: why people sometimes hear voices, believe things that others find strange, or appear out of touch with reality, and what can help’.

The report is being formally launched at a conference today in London.

The 180-page document has been written by a “working party mainly comprised of clinical psychologists drawn from the NHS and universities”. It has been welcomed by a number of key mental health organisations, including Mind.

This blog offers a critique of three key elements within the report, each of which has been tackled by an individual with special interest in the area:

1. CBT for psychosis: Keith Laws
2. Medication: Alex Langford
3. The psychological and biomedical models: Samei Huda

The report reviewed by our bloggers was the version available on the BPS website on 20th November 2014.

Methods

Before we get into the critique, here is a brief overview of how the Understanding Psychosis and Schizophrenia document was put together.

The report is an update of a previous document published by the same body in June 2000 entitled Recent Advances in Understanding Mental Illness and Psychotic Experiences (PDF).

The working party who put together the report (over a period of 2 years) is made up of 25 people, the majority of whom are clinical psychologists. The introduction tells us that more than a quarter of the contributors are people with lived experience, i.e. “people who have themselves heard voices, experienced paranoia or received diagnoses such as psychosis or schizophrenia”. The report is edited by Anne Cooke.

It is unclear from the report what specific methodology was used to put the document together. As far as we can tell, no methods are described for literature searching, inclusion criteria, assessment of studies, critical appraisal etc. We would welcome a response from the team who produced the report to clarify this important issue.

1. CBT for psychosis: Keith Laws

Meta-analyses

Understanding Psychosis & Schizophrenia (UPS) includes a section entitled ‘Talking – Psychological Help’, which refers broadly to talking therapies, but focusses on Cognitive Behavioural Therapy (CBT) as the jewel-in-the-crown of talk therapy- given that other talk therapies remain largely un-evidenced for schizophrenia and psychosis.  The report begins by attempting to establish the ‘scientific’ credibility of CBT, highlighting findings from several meta-analyses of – what are now – over 50 published Randomised Controlled Trials (RCTs). The authors state that:

Although they each yield slightly different estimates, there is general consensus that on average, people gain around as much benefit from CBT as they do from taking psychiatric medication.^{19, 20, 21, 22, 23, 24, 25, 26}

The extraordinary claims in this sentence deserve closer scrutiny. The only real consensus is that the average effect size of CBT for psychosis has been unremittingly shrinking; now down to 0.15 for total symptoms (this is averaged across 5 recent large meta-analyses). In plain language 94% of those receiving CBT will overlap in symptom outcome with controls who receive only treatment as usual; and for clinicians to attain one more favourable outcome from CBT (compared to controls), therapists would need to treat almost 22 people. CBT for psychosis is an intervention that is not statistically significant, not clinically significant and therefore not cost-effective.

Although those CBT advocates conducting RCTs have almost exclusively used CBT as a quasi-neuroleptic i.e. focussing on whether CBT reduces the symptoms of psychosis, UPS talks much more about ‘distress’ and reducing distress. While ‘distress reduction’ is a laudable target for therapists, the very few trials that have measured ‘distress’ as an outcome provide no support for CBT being efficacious in this regard either.

But potentially the most dangerous and fallacious claim is that:

people gain around as much benefit from CBT as they do from taking psychiatric medication

In this context, the authors cite a selection of CBT meta-analyses (refs 19 to 26; including Jauhar et al 2014 twice [21 & 25], where I am a co-author). However, none of the cited work compares effect sizes for CBT and medication and for one good reason: not one RCT has ever compared the efficacy of CBT and antipsychotic medication. It is disconcerting that a document endorsed by the British Psychological Society so boldly misrepresents the findings and misinforms patients, with potentially serious implications for decisions that they make regarding their health care.

National Institute of Clinical Excellence (NICE)

The UPS authors repeatedly refer to the guidance from NICE on CBT, for example, stating that:

[NICE] considers the evidence strong enough to recommend that everyone with a diagnosis of schizophrenia should be offered CBT

Despite such appeals to authority (and the UPS authorship contains current and past members of the NICE committee advising on CBT for psychosis, including the Chair), the current NICE recommendations are based on possibly the least reliable meta-analysis ever published on this topic. Strictly speaking, it is not a meta-analysis but a trawl of ‘110 meta analyses’ that fail to directly assess study quality and only examine RCTs up until 2008. Since 2008, at least 14 better quality RCTs have been published and only one reports a significant effect of CBT on psychosis symptoms (and that was a non-blind study). In a remarkable critique (Perera & Taylor 2014) of the recent NICE (CG178) 2014 update (sic), the Chair of the SIGN committee (the Scottish version of NICE) Professor Mark Taylor says:

In our view NICE CG178 promotes some psychosocial interventions, especially CBT, beyond the evidence. NICE CG178 also makes strong nonevidence-based recommendations

[And that]

…it is unfortunate that the guideline appears at times to reflect the interests of its authors rather than impartial up-to-date evidence

Schizophrenia Commission and the ‘One-in-ten’ claim

Finally, the UPS document repeatedly refers to a recent report by the Schizophrenia Commission, when they reiterate the claim that the SC:

…found that only one in ten people who could benefit from it have access to good CBTp ³⁰. We view this with grave concern – indeed, it has been described as scandalous³¹

A 10% figure would be scandalous (as so-described in a blog by some UPS co-authors); however, UPS surprisingly fails to elaborate at all on this figure, which is so dramatic that we ought to look closer. A glance at the SC report reveals that the 10% claim comes from a presentation to the Schizophrenia Commission by ‘Haddock’; and more astonishingly perhaps, the SC also give no further detail). The unreliability of this 10% figure is a case study in how myths emerge in mental health when we consider data on CBT for psychosis provision in the recent National Audit of Schizophrenia covering all 64 NHS trusts in England and Wales.

The NAS found that:

CBT was offered to 39% of service users, with a range of 14% to 67% across Trusts

A mean figure almost 4 times larger than quoted in UPS. Interestingly, of those offered CBT, NAS also report that more than half decline this talk therapy. When ‘service-users’ across all Trusts in England and Wales themselves were asked, the majority said they had either been offered or had declined CBT for psychosis, leaving a minority (46%) who “had not received CBT and had not indicated that they did not want to receive it”) – which, of course, is not the same as this minority saying that they do want CBT.

2. Medication: Alex Langford

The medication section of Understanding Psychosis and Schizophrenia leaves me with mixed feelings.

The benefits of antipsychotic medication

There are certainly some positive aspects. Most importantly, the report notes that medication can be helpful for some people, even sometimes on a long-term basis. A series of poignant vignettes are given to illustrate those facts. This is a welcome message, because medication can sometimes feel like a dirty word when in conversation with psychologists less supportive of the medical model.

The lengthy discussion about making sure that medication decisions are collaborative, informed and focussed on making sure the patient takes as little as they need is surely something we can all agree on. Additionally, the space dedicated to side effects is probably justified in being very comprehensive.

Little evidence for the dopamine hypothesis?

However, despite the scattered strong points, the space dedicated to medication reads very much as a step-by-step critique of it. For example, more than half of the free text in the ‘how medication can be helpful’ segment is spent undermining how medications are commonly thought to work – with highly scientifically controversial statements.

The ‘key point’ that there is ‘little evidence that [medications correct] an underlying abnormality’ is bizarrely unfounded. An excellent summary by Kapur & Howes (referenced earlier in the report itself) and further imaging studies by Howes and others provide solid evidence for elevated presynaptic dopamine levels being a key abnormality in psychosis, and there is copious evidence that inhibiting the action of this excess dopamine using antipsychotics leads to clinical improvement in psychosis.

Psychiatrists adopting a ‘drug-centred’ approach?

Another assertion that ‘some psychiatrists are suggesting adopting a ‘drug-centred’ rather than a ‘disease-centred’ approach’, whereby medications are thought of merely as tranquillisers not as specific agents to treat specific symptoms, also completely fails to match my experience of clinical reality. On not one ward, in not one clinic, in not one discussion with another psychiatrist, have I come across that ethos. The only place I’ve seen it is in critical psychiatrist Joanna Moncrieff’s book The Myth of the Chemical Cure, which is referenced three times in the medication section, (as is her personal blog and a scientific paper by her) but let’s just say her theories aren’t widely accepted.

Underemphasising the benefits of antipsychotic medication

As well as their underlying mechanisms of action, the benefits of medication seemed to be grossly underemphasised. The report spares only a few vague sentences on their advantages, but the commonly used antipsychotics have effect sizes of around 0.4 to 0.66, which is huge. Clozapine, the one antipsychotic which is more effective than all the others, has an effect size for symptom reduction of 0.88 – immense. As for preventing relapses, on average only 2 to 3 patients need to take an antipsychotic for one of them to avoid a relapse; a number the rest of medicine is envious of and which no other invention for psychosis comes close to.

None of this is to say that drugs should be the only treatment we offer – no psychiatrist believes that – but misrepresenting the facts isn’t very helpful for the people being offered the tablets. Of note, I didn’t see any psychiatrists named as contributors to the report, but I’m happy to be corrected.

A note on ‘terminology’

I was also interested in the constant use of inverted commas around ‘antipsychotic’, and indeed the sporadic use of ‘psychosis’ itself in the same state. Clearly the authors, and those who they feel they represent, choose to see the terms not as ‘real’ but as ‘constructed’ or ‘disputed’. I’m fine with that – but I wonder if it will serve the interests of the majority of people for whom this report is meant. For those who have come to form their narrative of suffering and healing using these terms, which is plenty, I can see the inverted commas being quite devaluing. Again, I’m happy to be challenged.

Justifying forced medication

One final sticking point for me was the short section, found later in the report, on ‘whether forced medication can ever be justified’. The report chose only to mention that although ‘some’ psychologists agree that detention can be justified, forced medication is becoming less justifiable, and a UN Special Rapporteur on torture has compiled a report against it. This is selective information in the extreme.

The report might have better informed its readers by stating that both detention and the giving of medication against the wishes of a patient is sadly commonplace, but also something that is widely accepted as necessary, occurring across the whole of medicine when patients are too unwell to make decisions for themselves. They could have mentioned that two legal frameworks (the Mental Capacity Act 2005 and the Mental Health Act 1983) support this practice, and given details of both, including how patients could seek help if they become involved in such a process. The relevance of the UN Special Rapporteur report, a dense and esoteric document which was compiled after visits to a collection of developing countries and bears relatively little consequence to current UK practice, should have been explained. Including it as the sole legal reference in this section seems tendentious.

3. The psychological and biomedical models: Samei Huda

Psychological and biomedical models are complementary

Understanding Psychosis and Schizophrenia (UPS) aims to outline the psychological model of psychosis. It does a reasonable job of this whilst failing to recognise the limitations of this model. It is also less successful when discussing aspects of the biomedical model which the authors don’t understand very well, despite their apparent belief that they do.

Of course both models illuminate important aspects of the psychosis phenomenon. UPS talks about the continuum of human experience that overlaps with psychosis. This is important to recognise whilst also pointing out that many medical disorders exist on continua with “normality” (e.g. high blood pressure, type II diabetes etc). It’s important to state that many people who hear voices don’t have mental illness. In clinical practice, I do use the continuum model with people I see who don’t have a serious mental illness.

Explaining psychosis

The biomedical model often doesn’t explain why these events occur (except in cases where clear aetiology exists such as illegal substances or delirium) apart from stating they are phenomenon as part of a disorder. The biomedical model often doesn’t explain the content of delusions or hallucinations well. The psychological model explains this aspect of content better. People tend to be more interested in the content of their experiences or thoughts than the form they take.

Where a psychological model is less successful is in explaining why these unusual experiences take the particular form they do in individual people. Why do different patients have PTSD or depression or anxiety or voices or delusions or OCD (or different combinations of these) in response to trauma? The biomedical model doesn’t help here either, just using the explanation of underlying unseen disorder, but it doesn’t claim to make everything understandable and explicable, as the psychological model does. Incorporating “constitutional” factors turns the psychological “understanding” model into a biopsychosocial medical model.

In terms of explaining the kind of delusions I see in people with schizophrenia, the psychological model is better at explaining some of the factors that maintain the belief, but less about why they formed that particular belief. It doesn’t explain the change in perception of reality that the delusion flows from. Even the example formulation of the bizarre delusion doesn’t explain why the bizarre belief was formed and accepted. The psychological model presented here doesn’t explain the experiences of thought insertion, withdrawal and broadcasting. Discussion of ego boundary disintegration might have been useful. There is a lack of convincing explanation of First Rank Symptoms.

Psychosis is more than just delusions and hallucinations

The reduction of psychosis to just hallucinations and delusions is flawed. To be fair, this can be blamed on DSM III onwards. Cognitive impairment is attributed to distraction and arousal, and negative symptoms to depression and arousal also. Whilst this may be true in some people, this doesn’t explain the wide range of cognitive impairments seen in many people with schizophrenia (and to a lesser degree in their first degree relatives) not caused by medication and why there are group differences with people with bipolar disorder. Some negative symptoms appear to be primary, not caused by depression or secondary to self-disorders as assessed by the EASE interview (see www.EASEnet.dk for details). Cognitive impairment and negative symptoms are important as they often have a bigger effect on social functioning than hallucinations or delusions.

Thought disorder is described as being caused by being flustered thus temporarily disorganised in speech. This trivialises and doesn’t explain the persistent thought disorder I see in some people.

The UPS report would have been stronger if it acknowledged it’s weaknesses, but this goes against its absolutist view that all psychosis phenomenon can be understood. It states Bentall’s view that once psychosis can be understood then there is no reason to posit an underlying illness. Hence any weaknesses in understanding are skated over, and phenomena not easily explained in this way are minimised.

Failure to acknowledge limitations of the model

It’s unclear why the authors felt the need to discuss medication given their lack of expertise. Some of the recommendations are unrealistic. In many cases where people have psychotic experiences that cause them to be at risk to themselves or others, their views should be respected but they should still have help in acknowledging an illness viewpoint so they can make good treatment choices. In some cases, in order to reduce imminent risk to themselves and others, compulsory treatment is necessary, though the threshold should be high.

In short, useful in parts, but undermined by the authors unwillingness to admit limitations of their model.

Summary

Our thanks to Keith Laws, Alex Langford and Samei Huda for working to such a tight deadline in putting this blog together. We hope that you find their critique interesting and useful.

We are keen to hear your views of the Understanding Psychosis and Schizophrenia report. What does it add to our existing knowledge? How will it impact on patient care? What changes (if any) should be made to mental health services as a result? Please comment in the thread below to let us know what you think.

Link