Discursive of Tunbridge Wells

A national scandal: psychological therapies for psychosis are helpful, but unavailable.

Home

A national scandal: psychological therapies for psychosis are helpful, but unavailable.

For years, drugs were it. If you felt paranoid, heard voices or were diagnosed with schizophrenia, the only thing likely to be on offer was ‘antipsychotic’ medication. Like all drugs, these have a number of different effects on our nervous system. Some of the effects can be helpful, for example calming us down or making our experiences less intense or distressing. Others may be less desirable. The unwanted effects – euphemistically called ‘side’ effects – of these particular drugs can be seriously distressing. For some people, they can be more disabling than the original problem. Despite the drug industry hype, it’s been a fine balance for many people, and worrying evidence is now emerging that some drugs can cause serious and permanent problems such as brain shrinkage if taken long-term.

In view of the downsides of antipsychotics it comes as something of a relief that there is a possible alternative. Psychological approaches such as cognitive behaviour therapy (or CBTp, the ‘p’ standing for psychosis) have become increasingly popular. NICE (the National Institute for Care Excellence) is sufficiently convinced of the effectiveness of these approaches to recommend that they should be offered to everyone with a diagnosis of schizophrenia.  Traditionally they have been offered in addition to drugs, but a recent trial suggests that they might also be promising as an alternative.

Last week saw a flurry of debate about this issue. On Wednesday, the Guardian published an article somewhat sceptical of the value of psychological therapies for people experiencing psychosis. The same day, a packed house at London’s famous Institute of Psychiatry debated the motion ‘This House Believes that CBT for Psychosis has been Oversold’.  We were there: Peter was one of the speakers against the motion.

The proposers of the motion suggested that CBTp should ‘live or die by meta-analysis’. A meta-analysis, for the uninitiated, is a statistical technique that allows findings from various trials to be averaged out.  As in the Guardian article, the argument was that some of the claims made by researchers in this field have been overblown.

Psychiatrist David Kingdon, himself a leading researcher in the field and responsible for some of the trials, countered that only certain studies get selected for meta-analyses. He also gave some of the background behind the figures – what actually happened, and what the (often positive and grateful) participants in the studies actually said. The yardstick for overselling in the mental health field has to be the efforts of the pharmaceutical industry.   Professor Kingdon asked the audience – many of them NHS clinicians with offices overflowing with drug company freebies – if they could remember ever seeing a mug or post-it note with ‘Psychological Therapy for Psychosis’ or ‘CBTp’ printed on it in a fancy typeface?  There was slight shuffling while members of the audience discreetly covered up the drug company pads and pens they were using…

As often happens, all four debaters were white, male, middle class academics, but we did hear from people who’d been offered therapy, including this man, David: “CBT has helped me remain aloof from this voice and I no longer believe what it says. I now think of it as a petty bully and don’t let it bother me…. The only thing I regret is that I didn’t have access to CBT sooner – it could have prevented a lot of suicide attempts and I wouldn’t have felt so awful for so long.”

The problem with many trials, and therefore with meta-analyses too, is that professionals decide in advance what they are going to measure and what counts as a ‘good’ outcome.  Those may or may not be the things that are actually important to people.  For example, trials frequently measure ‘symptom intensity’ – how loud someone’s voices are, say.  That may of course be irrelevant to how upsetting the person finds them, how satisfied they are with their life more generally, or what their expectations of therapy are.  Even if the intensity of the ‘symptoms’ doesn’t change much, therapy may help with other things, as David’s account illustrates.

We also heard evidence that CBT can actually lead to changes in the ‘wiring’ of the brain – it’s not just a sticking plaster solution.  And unlike drugs, talking therapies are only ever offered to people, not foisted or forced on them.  The influential Schizophrenia Commission recently found that despite CBT for psychosis being recommended by NICE, only one in ten people who could benefit are actually offered it.  The other nine presumably get drugs, willingly or in some cases under duress.

Then it was time for the audience to join in. And they sure did.  One person was concerned that CBT might squeeze out other psychological approaches.  A psychiatrist thought that was irrelevant given how few people are offered any talking therapy – he said that in his whole career he had never come across anyoneexperiencing psychosis who had been offered CBT.  Others pointed out that things are moving apace as we learn what people find helpful, and rather than splitting hairs over the relative merits of different approaches, we should celebrate that we now have the science to back up what service users have been telling us for years.  When we’re in great distress, be that because of life events or so-called ‘psychiatric symptoms’ like voices, we need the opportunity to talk through our experiences and – where relevant – how they have affected our view of the world.  We need a calm, supportive and non-judgmental atmosphere to do that in, with someone familiar with the territory.  If that was what mental health services provided, rather than – as too often happens – just insisting that people accept that they are ‘ill’ and take drugs, the outlook for those of us who need to use them might be very different.

And the result of the debate?  At the show of hands beforehand, the audience was evenly split (83 for,  87 against, 61 abstainers). After the arguments had been heard, around half of those who had previously voted for the motion went over to the other side, together with more than half of the abstainers.  Only 47 voted for the motion and 25 abstained. The vast majority, 132, decided that CBTp has NOT been oversold.

The issue is not one of overselling, it’s that psychological therapies are shamefully underprovided.

About the authors

Peter Kinderman is Professor of Clinical Psychology at the University of Liverpool and author of ‘A Prescription for Psychiatry’ (due for publication in September 2014). Follow him on Twitter @peterkinderman.

Anne Cooke is Joint Clinical Director of the Doctoral Programme in Clinical Psychology, Salomons Centre for Applied Psychology, Canterbury Christ Church University.  She is editor of the forthcoming report from the British Psychological Society Division of Clinical Psychology, ‘Understanding Psychosis and Schizophrenia’. You can follow her on Twitter @AnneCooke14

Share this page:

159 comments on “A national scandal: psychological therapies for psychosis are helpful, but unavailable

  1. I'll nip in first with a couple of thoughts. I wonder how well served the public are by the polarisation of such debates. Following the recent criticisms of CBT for psychosis I can't help but agree (up to a point) with the critics. These approaches are oversold. That fox has been shot and well before the recent kerfuffle. I think this is partly to do with taking a stance of saying 'this works for that' and kind of leaving it there. On the other hand I'm not sure how productive the recent criticism has been either. The primary focus has been on looking at symptom reduction and this may place insufficient emphasis on the point you make about distress reduction as an outcome criterion. This work also seems to be a little undermined by over statement and a determination to damn such approaches.Saying such therapies are oversold does not however mean they are not helpful or useful and the truth may be somewhere in between the positions taken in the recent debate. During the course of a literature review of CBT for psychosis I conducted a few years back three things emerged clearly. One is that the estimates of the number of people who benefit varies hugely from trial to trial and depends on outcome criteria. The second point was the the number of people who can benefit from these approaches is relatively small (not comparable to depression of the main anxiety presentations). The third key point is that the benefit seemed mainly concentrated in changed views of distress. The kind of thing that David's quote above addressed.Perhaps a more useful question to ask might be about how to represent the literature thus far in the NICE guidelines and clinical services. I'm not convinced (from both clinical experience and the literature) that everyone is suitable for these treatments. How do we decide quickly and cost-effectively who is?Of course there are many, many criticisms to be made of medication too and I think that it is important to look further at claims of harm caused by antipsychotic medication. If you are taking these meds, and relying on them, it must be very difficult to encounter the claim an counter claim when you go onto Twitter etc. The final thing that bothers me is the presentation of this as either/or. My own clinical experience is that, if it is done thoughtfully, medication and psychological therapies can be used in concert. Certainly I've encountered situations where medication has opened a path up for psychological approaches. I'm not sure how well professionals serve the users of services if we close down such possibilities.BTW for a discussion of who does and doesn't seem to benefit from CBT for psychosis here is a short qualitative study I published a few years back. https://docs.google.com/viewer?a=v&pid=sites&srcid=ZGVmYXVsdGRvbWFpbnxqb2hubWNnYXJjaGl2ZXxneDo4NGU0NzA0YTAxZTNiZjU

  2. There are at least two meta-analyses that "go against" CBTp–McKenna et al (2010) and Jauhar et al (2014). This piece re-capitulates a claim made elsewhere by Richard Bentall that both those meta analyses are both flawed by dint of the studies they (fail to) include. Bentall called this "rotten cherry picking" and suggests we should focus on other meta-analyses that come out in favour of CBTp. This constitutes something of an impasse; it is nigh on impossible for the non-expert methodologist to take an informed view. Which studies do we believe? At this stage evidence about how an audience voted, or the fact that CBTp changes the brain are less interesting to me than getting through this impasse. Those questions are not necessarily in contention; the efficacy of CBTp is. To my eyes, the criteria outlined in the Jauhar et al meta-analysis in BJP in January were clear and exemplary. The authors make clear what they define as bias (bias being a widely acknowledged problem in RCTs) and we can see that studies were excluded on those grounds. This view is evidently not shared by everyone in the field, but I would like to be told explicitly if it's wrong and how. I would like to see a more substantive outline of the ways in which the two meta-analyses' are flawed and why we should believe "pro" meta-analyses rather than "anti" ones.On a broader note-my attitude to the CBTp vs. meds debate is "a plague on both your houses". CBTp seems to have middling effects, and we know that meds are over-used (and "over-sold") given the limited (but real) role they have in treatment. We should be asking how to move to a position where this is not the full extent of the care that people are given. I recently read an old RCT of Batman & Fonagy's, investigating the effectiveness of their partial hospitalisation programme for BPD. The results were very encouraging to me. While it's tempting to attribute this to the psychoanalytic "active ingredient" (MBT), we should remember that the control was "standard psychiatric care"-which amounted to hospitalisation as needed (when someone's in crisis) and periodic psychiatric "check-in". The treatment was a structured, intensive period of contact with services, keeping people steady through a difficult period, and with a theoretically well developed psychotherapy. This sort of structural innovation seems much more likely to be the future of a fruitful involvement of clinical psychology in psychosis. Rather than defend CBTp tooth and nail, we should remain open to acknowledging its limitations and keep on thinking.

  3. Throughout this debate, what has been so interesting is that there has been endless critique of CBT for psychosis but precious little has been said about the research methods used to evaluate it. There are also huge assumptions being made about what constitutes 'effectiveness' in psychotherapy. It seems a great pity that the CBTp researchers are so focused on symptom reduction when there are clearly a number of possible outcomes worthy of scientific scrutiny. One of the biggest problems is that we are so dominated by a medical discourse that it is as if we can employ no other language in our critique of psychotherapeutic endeavour. It may be the case that some researchers believe that it is only by 'medicalising' research and attempting to think of psychotherapy as being akin to a drug, that they think their efforts will be taken seriously. Doubtless, they are perhaps more likely to gain research monies if they do this. Like it or not, psychotherapy is not a medical treatment and should not be evaluated as such. It represents a complex human interaction unique to the two people undertaking it, even when a standardised protocol is utilised. RCT's, while useful in telling us whether an intervention has had a significant effect, whether a therapy is better than another, or better than treatment as usual; can tell us nothing about the experience of psychotherapy and what service-users might find useful about it. What might be more useful is to begin the process with qualitative research which examines service-user experiences of CBTp. From this, we might well refine pertinent research questions and hypotheses that can then be evaluated quantitatively.

  4. Reading this blog post, i am left concerned whether British clinical psychologists lack basic training in the evaluation of evidence, not antidote. The statement in the blog "problem with many trials, and therefore with meta-analyses too, is that professionals decide in advance what they are going to measure and what counts as a ‘good’ outcome." is truly embarrassing. Peter Kinderman and Anne Cooke (and others), I challenge you to offer evidence refuting my critique of the miserably flawed Lancet CBTp study at http://blogs.plos.org/mindthebrain/2014/03/11/much-ado-modest-misrepresented-trial-cbt-schizophrenia-part-2/http://blogs.plos.org/mindthebrain/2014/02/25/much-ado-little-lancet-study-cognitive-therapy-persons-unmedicated-schizophrenia/Your shuck and jivin' is irresponsible and borders unethical in how it is intended to mislead patients and families.

  5. I'm a bit tired of this debate, which is being pursued by a lot of people who don't understand the pragmatics of RCTs, or who have never themselves tried to do anything to improve the well-being of psychiatric patients, or who are not clinicians and have never been in the position of not knowing what to do when faced with a patient who seems to be doing badly on standard treatments (if Keith Laws and I arrive together at St Peter's Gate and there's only room for one more entrant I won't be too anxious). In a such a situation, a sensible clinician will ask herself 'is CBT worth a shot?' and the answer will be 'yes'. Even on the Jahuar meta-analysis there was a small effect for CBT, so the questions we should be debating are not 'does CBT work?' but 'who does it work for?' and 'in what circumstances?'.With regard to the Lancet trial, the published analysis shows an effect. There may be other analyses that Laws and Coyne prefer that don't show an effect, but that's post hoc reasoning and rotten cherry picking. Many of Jame Coyne's methodological objections seem to be based on a fundamental difference of opinion about the nature of psychosis/schizophrenia, and about a the ethics and pragmatics of RCTs. For example, he goes on about the sample being too heterogeneous and schizophrenia being a chronic condition. Many people in the UK clinical psychology community including myself would dispute whether the concept of 'schizophrenia'is coherent (there is almost zero evidence that it is a reliable or scientifically valid diagnosis) or that the outcome of people diagnosed as schizophrenia is always chronic. Moreover, for a pragmatic RCT it's important to select patients who are representative of the services that the treatment will be based in – that's why the trials are called pragmatic. In UK services (and I'm pretty sure elsewhere) patients turn up with psychotic symptoms but only a small number of them meet the criteria for schizophrenia as laid out in that book of psychiatric fairy-tales called DSM-5.Coyne is upset that the Lancet trial used a TAU control but TAU is just about always used as a control in UK trials, for sound ethical reasons – anything else involves prohibiting patients from accessing treatments which they might want to access as the trial progresses. Hence, ethics committees and funding agencies such as the HTA will almost always demand TAU. Moreover, of course, TAU as a control tends to be conservative – it reduces the chances of finding a significant difference between groups.Does the Lancet trial prove CBT works with people not taking meds? Probably not, although the answer will no doubt depend on what standard you take for proof. Does it show that a clinician with a patient who refuses medication (which many patients have good reason for doing) might be sensible to refer the patient to a CBT therapist – yes, probably. Does it mean that further research on CBT for unmediated patients is warranted – absolutely.This is my last word on this topic. I've got much better things to do – like carrying out research into the social determinants of psychosis, figuring out how to use that knowledge to promote better public mental health and – yes – developing better treatments for some of our most disadvantged citizens (many of who would prefer me to spend my time this way than to waste time debating with people who have very little positive to offer).

  6. Sorry, Richard Bentall, do you uncritically accept what you read in Lancet? The reporting of outcomes by Morrison et al was thoroughly dishonest. You don't really understand my criticism of TAU in that trial. It was set of very different circumstances that, when combined, do not generalize to any one of those circumstances. Many of the patients left in TAU in the conventional settings were left with no treatment. That is not ethical. The Lancet interview with Morrison reveals he was aware that if minimal support were made available to patients, it might wash out any effects of CBT. He did not provide that because he desperately wanted to assure an effect, but he still did not get one.

  7. If we are to take "first do no harm" seriously, and to understand the history of science and medicine, and how knowledge progresses over time, then we see that it's often better/safer to do nothing than to attempt unproven and potentially harmful interventions such as lobotomies and psychotherapy.

  8. Doh! It's you who doesn't understand a very basic aspect of RCT design. Treatment as usual (TAU) means that patients in the control group could have any treatment they wanted that their clinician was willing to make available. The only reason why some control patients were left with no treatment was because they actively refused it. There was no attempt to minimise the support provided to them as quite milseadingly suggested in your post, indeed no attempt to influence the treatment available to them in any way..That's precisely what TAU means.Hence,TAU is the most ethical control. It's also highly conservative and so reduces the likelihood of a significant difference between the groups. Which is why UK research ethics committees and funding agencies usually insist on it.

  9. If we were to take the 'first do no harm' principle seriously, there would be no psychiatric drugs. For example, the evidence that antipsychotics have severe, sometimes life-threatening side effects (metabolic syndrome, agranularcytosis, increased risk of cardiac effects, diabetes, extrapyramidal effects etc etc) is not even controversial.There is no evidence that psychotherapy has comparable adverse consequences – that's one of its advantages. UK NIHR or MRC-funded psychotherapy trials are required to include detailed monitoring of adverse reactions by an independent data monitoring and ethics committee (DMECs, I chair one such committee and am a member of another). DMECs have the power to close down trials if there is evidence that the intervention is causing harm.But, frankly, anyone who thinks the side effects of psychotherapy are comparable to those of lobotomy is probably not interested in a rational argument. You might accuse psychotherapists of being useless, but you must surely conceed that they don't actually cut holes in their patients.

  10. PS While we're on the Lancet trial, there's a piece of Lawsian silliness that I've resisted getting off my chest, but I might as well go for it now. In his attempt to reanalyse the data using the figures in the Morrison tables, he says, "I would argue that their data show, for the first time, how patients who choose to be unmedicated display fluctuations in symptomatology".The figures in the tables are group means etc. You can't work out fluctuations from group data = you need individual data, because one patient may go up while another goes down. Even if you could work out fluctuations, you'd need a control group of medicated patients to show, "for the first time, how patients who choose to be unmedicated display fluctuations in symptomatology".You can't make this up (although some people appear to be trying)!

  11. Throughout this debate, I have felt embarrassed by Morrison et al trial data – which only serves to further clinical psychology. Leaving this aside and consider the 'gold standard' of trial data by Cochrane collaborative, the concept of CBT for psychosis does not and will not hold up as an evidence based intervention. On balance, the only real merit for a patient, (who are often forgotten) is an opportunity to talk about their experience of psychosis – which by implication is not a clinical intervention. A question that needs to be debated is – why does clinical psychology feel the need to promote any intervention that serves its desperate need to be a medical intervention? Which only further damages patients mental health and further isolates them due to a Dunning/Kruger effect within writing such as the one posited here. In closing, a simplified version of my comment for those suffering from the above effect is patient first and leave your making up of evidence based intervention to your own delusion.

  12. This seems an important conversation and I’m glad that sites like ‘Discursive’ (which I’m proud is hosted by my own institution) are opening up debates that have hitherto often only been held within the ‘ivory towers’ of academia. Interestingly, this comments thread is very different from the one going on the other website which published our piece, Mad in America (http://www.madinamerica.com). Many of the comments there are from current or former service users who feel they have been harmed by the absence of an opportunity to talk, and particularly by clinicians failing to listen. Maybe the issue of listening is also key to the debate about the potential harm in therapy. I’m not sure what concerns underlie the ferociousness of recent criticisms of CBTp, but from some of the comments it seems that the issue of potential harm might be one. What seems important to remember here is that therapy is essentially a conversation between (usually) two people. Like all conversations, there is the potential for unhelpful as well as helpful exchanges. The difference is often in the extent to which the people involved really listen to each other. I know from listening to recordings of trainee therapists that like all of us CBT therapists, particularly inexperienced ones, sometimes make the mistake of suggesting ways forward before they truly understand the person’s situation. Where the therapist really listens, the conversation continually changes as a result of the pair’s evolving understanding of the situation, therapy is a truly collaborative effort, and the potential for harm is much reduced. Perhaps we need to stop thinking of therapy as akin to a drug – a discrete ‘intervention’ that we ‘deliver’ to people, but as what it is – English Conversation. Unlike the previous commenter, I think that the ability to do that well, to have an on-going, collaborative, helpful conversation with someone who is in great distress, perhaps because of persecutory voices or paranoia, needs a high degree of both knowledge and skill and is often the most important ‘clinical intervention’ that we can offer.

  13. Richard Bentall, If you take the time to check Google Sholar, you will see that I have almost twice the citations that you do. I bring up this awkward fact only to suggest that the gap did not develop because I am the one who does not understand RCTs.Offering CBTp to persons with unmedicated schizophrenia is not necessarily "doing no harm". Offering an ineffectual treatments does harm when it discourages or delay seeking more effective treatment. That is the argument that all of us and science based medicine accept in rejecting homeopathy.CBTp is not at a stage of developing as an evidence-based treatment that it benefits from pragmatic comparisons with treatment as usual. Comparisons with TAU are less useful and ruling out the important hypothesis that CBTp is no more effective than a placebo condition, i.e., nonspecific supportive therapy. Again, I urge you to listen to the podcast of the interview that Morrison did for Lancet. He clearly indicates that the group was avoiding a comparison with supportive therapy because it might not show an effect for CBT.Pragmatic trials more justified at this stage of demonstrating effectiveness in delivery, not efficacy under more control conditions.Although quite small, Morrison's Lancet study of CBTp had a greater proportion of deaths than trials of CBT in conjunction with antipsychotics. I think ethics committees should pay attention to that.Richard, please check the Lancet article again and, even better, the trial paper In Psychosis. Only some of the participants relegated to treatment as usual had many options. Others were in traditional settings where they were likely to be expelled for not accepting medication. Blithely combining these two types of settings into one comparison/control condition is bad science and inhibits substantive interpretation of results.Richard, I see that you are now getting into the actual danger of the trial. I think the you have to concede that at the last assessment, if imputation is done has Morrison did, the data for most patients are made up. Elsewhere, Morrison has referred to studies of antipsychotic medication as not producing usable data if so large of the proportions of the patients are missing. But that is exactly what he does. So, if you going to be discussing the actual data, let us get into this inappropriate sleight-of-hand by Morrison.

  14. No, @Ann Cooke, The most pressing issue is that CBTp has been promoted by Morrison et al as being an alternative to antipsychotic medication, matching its efficacy. That has not yet been established and is irresponsible to tell consumers.

  15. Oh hello,is anyone actually speaking to 'the consumers'? None of these endless debates appear to be. Some of us just want protected talk time and who cares whether that's structured or unstructured talk, just to have the option, and the option to not take meds. NHS 'Recovery' approaches and mindfulness don't have an RCT 'evidence base' but are widely used within 'recovery colleges', but no one's staking their academic prowess on that.

  16. James, is this what it comes down to – how many citations we have? (begin ironic tone of voice) I guess that must make you automatically right about everything then (end ironic tone of voice).Having tons of citations does not make you right about the TAU issue. You're just plain wrong because you've misunderstood the concept.And as for the ethics of witholding meds: (i) nobody is witholding meds; these are patients who don't want them and have a right not to want them (respect for autonomy being one of the pillars of medical ethics); (ii) you might want to check out the paper by Wunderink, et al. (Recovery in remitted first-episode psychosis at 7 years of follow-up of an early dose reduction/discontinuation or maintenance treatment strategy: long-term follow-up of a 2-year randomized clinical trial.JAMA psychiatry, 70, 913-920, 2013, doi: 10.1001/jamapsychiatry.2013.19) which shows that patients kept on meds have worse long-term outcome than those taken off them.We urgently need to develop alternatives to medication-based psychiatry. And we have an ethical duty to provide patients with choice about treatment. That's why this kind of research is so important, and in the best ethical tradition.

  17. "We urgently need to develop alternatives to medication-based psychiatry".That's what is getting missed. I know I'm not the only service user who is frankly sick to death of all this

  18. Richard Bentall said: "…TAU as a control tends to be conservative – it reduces the chances of finding a significant difference between groups."May I ask, what evidence supports your assertion that TAU tends to be conservative when compared to a well-designed placebo control?Also, what effects are expected to be seen in a TAU arm of a trial for schizophrenia?And what effects are expected to be seen in a well-designed placebo control in a trial for schizophrenia?

  19. As the previous commenter, a point that is often lost in these conversations between clinical psychology and patients is – the hubris of many therapist, that in part has engendered a high degree of people leaving therapy and never engaging again. As such many of these patients find it helpful to have a collaborative helpful conversation in relation to their experience with a family doctor or in extreme situations staff that support their local A&E such as paramedics and nurses. Which by a medicinal definition is not an intervention or at best offers a placebo to the real issues. So is that your point made above or have I got that wrong But the on-going collaborative relationship is not the real issue, because on thinking about it – the pressing fact is CBT for psychosis, cannot engage with a degree of efficacy given the high drop out rate within the Morrison et al data. Further to this, often in the desperate need to promote such interventions, clinical psychology forgets this simple fact – within the Morrison et al study people have attempted suicide which can be directly correlated to the lack of efficacy. Given this condition, the study is being investigated for ethical concerns about design and presentation of results. So my question is a simple one – does the discursive of Tunbridge well want to continue promoting an intervention that is being investigated for ethical concerns?

  20. Richard Bentall said: "I'm a bit tired of this debate, which is being pursued by a lot of people who […] are not clinicians and have never been in the position of not knowing what to do when faced with a patient who seems to be doing badly on standard treatments […]. In a such a situation, a sensible clinician will ask herself 'is CBT worth a shot?' and the answer will be 'yes'."I think this might explain the strong differences of opinion in this discussion.Richard C Coyne is objectively discussing the scientific evidence-base for a specific therapy, whereas Richard Bentall is subjectively discussing the needs and the anecdotal experiences of a clinician.So I suppose a question that needs to be addressed, and is central to this discussion is: "If a clinician has run out of options when treating a patient, is an unproven therapy, better than nothing?"Homeopaths also report anecdotal success for homeopathy, however, when scrutinised in robustly designed trials, with a robust placebo control group, it is demonstrated that homeopathy fails to treat illnesses. Even so, some patients may find homeopathy helpful because of factors such as the 'therapeutic alliance' and the placebo effect.However, unproven therapies can also be harmful for a variety of reasons.Harms for CBT in clinical trials are notoriously under-reported.

  21. Thank you to everyone for a stimulating debate (with a single caveat).I’m slightly confused by the occasional calls to “show us the evidence” (of the effectiveness of cognitive behavioural therapy for psychosis; CBTp). While it’s admirable to subject that evidence to detailed forensic scrutiny, the evidence is there. There have (as we discussed in the Maudsley debate https://www.kcl.ac.uk/iop/news/debates/index.aspx) been several (more than 50, if memory serves me) trials of CBTp, and several meta-analyses. These have been reasonably consistent, with even the more critical (such as the Jauhar et al analysis http://bjp.rcpsych.org/content/204/1/20) concluding that “cognitive-behavioural therapy has a therapeutic effect on schizophrenic symptoms”. While it’s easy to say “show us the evidence”, the evidence is already there. But this debate touches on other areas. So as well as a detailed statistical critique, I think it is perfectly appropriate to widen the discussion a little.The issues of diagnosis and outcome are important. Don’t misunderstand me; I believe CBTp holds its own within the current paradigm, but these psychological approaches throw up wider issues.Current psychiatric diagnosis doesn’t map well onto any real-world entities (when, for instance, cluster analyses are conducted) and don’t appear to be very reliable in clinical practice. But perhaps more importantly from a psychological approach, traditional psychiatric diagnoses fail to predict response to treatment and, even more importantly, don’t map onto causal factors (for example the sequalae of childhood abuse), don’t map onto psychological processes (for example self-referent processing or source-attribution) and don’t even map onto biological mechanisms (as Thomas Ensel recently pointed out). This speaks to the need to re-think our approach to mental health issues in a fundamental way.We also, and therefore, need to ask questions about what ‘outcome’ means. So while I agree that, if a CBTp trial aims to reduce the severity of ‘symptoms’, then this claim should be tested, and likewise if a CBTp trial aims to reduce the severity of ‘distress’, then distress should fall, we also need to question these notions. As we improve our trials methodology – as a consequence of the critiques from critics – we should refine our definition of outcome, and that means questioning and challenging traditional paradigms.And then this debate speaks to other broader issues. Those of us interested in the effectiveness of CBT are also interested in fundamental questions about the psychological mechanisms that underpin psychotic phenomena. The research is too numerous to give hypertext links, but research into CBT also highlights the research into those underlying mechanisms (which we summarised a few years ago in a report currently being updated http://www.schizophrenia.com/research/Rep03.pdf).The point is that there are two debates running concurrently here – and both are valuable. The narrow debate is over the effectiveness of CBTp. In that debate, I think it’s great that the claims for CBTp are subjected to detailed, even forensic, scrutiny. But the meta-analyses point, convincingly in my opinion, to the conclusion that CBTp makes a valuable contribution to clinical care. The wider debate is over whether psychological perspectives challenge traditional thinking and enhance our understanding of mental health. I think they do, and I think it’s fantastic that we’re having this debate.My caveat, then, is that I hope we can remain civil. I am a little disturbed by the tone of some of the comments, and I think that personal invective is inappropriate and should be avoided. I also think that middle-aged men should avoid comparing the length of their … citation records. Let’s retain some dignity here.

  22. Thank you Peter, I was just thinking too much of this narrowly conducted debate over the last few months has looked an awful lot like academic willy waving

  23. Well said, Professor Kinderman. Clearly this is about much more than the actual issue under debate. If science is about a disinterested assessment of the evidence, then the relentless misinterpretations and jibes of some of the anti-CBT proponents – especially, but not only, those who have no clinical experience – suggest that this argument is primarily driven by non-scientific factors such as ego, threat and the defence of careers that are based on the biomedical paradigm. I agree with Anonymous, Richard Bentall, and no doubt many others that this is getting beyond tiresome. And I also agree with them and with Peter Kinderman and David Kingdon that the main casualties of this childish spat are service users themselves.Perhaps the most worrying aspect of the anti-CBTp camp is their apparent contempt for what service users actually say. In the aftermath of the resounding defeat at the Maudsley, service user voices have been variously dismissed as ‘anecdata’, ‘emotive anecdote’, ‘sentimental anecdote’ (Ben Goldacre – I am disappointed in you!) and so on. No one is denying that RCTs, meta-analyses etc have an important place, but how did some people reach a position where this means that all other forms of evidence are ruthlessly discounted? Indeed, that the SU perspective doesn’t count as ‘evidence’ at all, and that citing it can only mean that you have run out of arguments? I find this attitude deeply concerning, shocking and offensive.It’s time to challenge this narrow, elitist epistemological hierarchy. You don’t have to be a diehard CBT fan – indeed, I am not one – to believe that psychiatry needs more, not less, attention to what service users actually say, alongside its more traditional research endeavours. One of the earliest survivor campaigners, Viv Lindow, wanted the dismissive term ‘anecdotal evidence’ replaced by ‘personal testimony.’ If we had listened more to personal testimony and less to the superficially reassuring official evidence, we might have avoided all kinds of scandals in psychiatry – the widespread addiction to minor tranquillisers, the loss of memory due to ECT, the damage that neuroleptics can cause, and so on. What service users have overwhelmingly asked for, over many decades, is more chance to tell their stories, to be heard and listened to within trusting and respectful relationships. If CBT is one way of delivering this, and is found helpful by some, let’s not rudely dismiss their testimony so that we can continue to bash our opponents over the head with piles of statistics. Let’s offer it to them, and let them tell us how to make it even more effective.

  24. Anecdote =/= data & thus Not Evidence.Not hard to understand, is it?Stay away from terms like "epistemological"; you want to be taken on faith & that's metaphysics, not epistemology.Best regards from,A Service User*.* Stop trying to pretend you speak for patients. You don't & you shouldn't presume to. You CBT fanboys & girls are way too happy to dismiss us patients & exculpate yourselves when we express dissatisfaction with CBT, p or otherwise.The idea that CBT relationships are "trusting & respectful" is shark-jumpingly laughable. This comment is posted anonymously precisely because I no longer trust you, never have experienced anything like respect at the hands of CBT pushers & believe that you lot need not just an evidential wake-up call, but a major ethical overhaul. I'm not holding my breath on either count.

  25. Well said Richard Bentall as someone who's family through 3 generations has been affected negatively by psychiatric drugs, lives shortened, physical conditions. Just because we all experience psychoses, it's normal for us. But we can't get choices for treatment, it's drugs or nothing. That's the reality. And nothing isn't good enough.People's stories have to mean something, we are more than a statistic and not the same as transgenic mice, no matter what the neuroscientists say in their search for biomarkers for mental illness. I never believed the schizo labels and made my own escape, recovery they call it, getting off the drugs, doing it myself against the advice of psychiatry.

  26. Rory can I ask a simple question leaving aside all the narrowly constructed view and move towards a broader perspective informed by Cochrane review data. Given the vast number of RCT in relation to antipsychotics and no reported deaths. However within this study that you claim to be involved in, there has been a reported suicide and deaths. Do you feel any responsibility for not spotting the possibility of a vulnerability of this candidate? or as I suspect an un medicated schizophrenic patient is just a source of data. If so, how do you account for the death not affecting the efficacy of you intervention? Put simply Hundreds of recorded RCT in relation to antipsychotic medication and no deaths. Within your trail at least one death recorded. Further to this how do you account for a symptom profile that is lower than those used for antipsychotic medication trials and the drop out rate before the start of your trail. In closing, do you not think these are real ethical questions, that cannot be brushed aside by dunningKruger bias to blame other factors within the trail. Also any research understands an ethical committee with find any reason not to blame the research team. In-order not to open themselves or institution to legislative action. Please answer honestly and with integrity, otherwise don't bother giving the above excuses, you at least have that responsibility to the candidate that committed suicide

  27. I don't think you have grasped my point. Why do you use such a disparaging description of people's lived experiences, whoever you are addressing? And why, for you, does their personal testimony not count as a form of evidence?This increasing tendency to dismiss voices which have always been marginalised within a medical/psychiatric/positivist discourse is deeply concerning.

  28. I'm sorry you have had such bad experiences. Just to be clear, I am not a CBT therapist, don't use CBT with psychosis, and have no personal investment in defending it. My point is that we should take account of SU views when assessing its pros and cons – or the pros and cons of any other approach – and that this is a crucially important form of evidence, not to be dismissed by quoting statistics.

  29. Too right Clare, the problem with evidence is that SU testimony isn't even on the radar in the hierarchy of evidence, at best SU groups will only have small studies not RCT's, and can't access the funding. You sit on a guideline development group to be told "that's just your opinion"

  30. I agree.Since I (and David) used personal accounts in our presentations for the Maudsley debate, I think I should also add:..I don't regard my use of personal histories as an alternative to evidence. Firstly, they are a valuable form of evidence (not an alternative to it). But, for me equally importantly, they illustrate and make personal the account of the statistics.Let's be clear. Jauhar and colleagues found, in their meta-analysis that "cognitive-behavioural therapy has a therapeutic effect on schizophrenic symptoms". That's been replicated by others – as we've discussed quite a lot. But to discover what that means in personal, as well as statistical terms, we turn to the accounts of people like David. You can hear me reading out his words on the podcast of the Maudsley debate; "the difference CBT has made is amazing – it has really transformed my life". David's words aren't an alternative to the statistical evidence – they illustrate and personalise that evidence.

  31. Rory, please refer to your comment on 13 April at 16:22 you stated that a single suicide took place. Yet if we place this alongside the above statement at 20:31 there now appears to be two deaths. In friendly terms 'hardly a credible reporting of the data'. Which opens you to further questioning such as how many suicide attempts that took place during the trial' and not recorded. Which of course, such events would not have been related to participation in the trial. In closing, please answer the following question honestly and integrity ' how many death occurred during a RCT for antipsychotic medication'? The answer from Cochrane collaborative is none reported. Further to this, posting a URL is hardly a credible reply from a professional researcher. As such, I agree with the view above that you maybe an embarrassment to the research community

  32. Hi Richard,Just regarding your point about the "Lawsian silliness". I think you are incorrect. I agree that from a stable group mean you cannot deduce whether or not any fluctuations have occurred (for the reasons you mention). However if the group mean itself fluctuates then you most definitely can say that fluctuations have occured at both individual and group level – you can't get the group mean to fluctuate without the individual data also fluctuating!And regarding your second point – you only need a control group if you're wanting to compare it to that control group. It is perfectly reasonable to say "patients who choose to be unmedicated display fluctuations in symptomatology"; he's not saying they show more or less flucatuations than medicated patients.

  33. Professor Peter Kinderman said: "Jauhar and colleagues found, in their meta-analysis that "cognitive-behavioural therapy has a therapeutic effect on schizophrenic symptoms". That's been replicated by others – as we've discussed quite a lot."Your interpretation of the Jauhar meta-analysis startles me. Have you actually read it? Do you understand it?"Cognitive-behavioural therapy has a therapeutic effect on schizophrenic symptoms in the ‘small’ range. This reduces further when sources of bias, particularly masking, are controlled for."[http://bjp.rcpsych.org/content/204/1/20]Do you understand the bit about bias, and the bit about masking, and do you understand how to interpret a 'small' effect size in medical trials?This meta-analysis does not support CBT for schizophrenia, even if you ignore the significant methodological weaknesses of the reviewed trials.Professor Peter Kinderman said: "You can hear me reading out his words on the podcast of the Maudsley debate; "the difference CBT has made is amazing – it has really transformed my life". David's words aren't an alternative to the statistical evidence – they illustrate and personalise that evidence."The evidence from trials demonstrates that the majority of schizophrenia patients do not benefit from CBT, so do you also quote the words of the patients who have not benefited from CBT or who have been harmed by it? Or are you selective in the anecdotes that you choose in order to support your personal bias?I expect it is difficult to keep track of those patients who are directly or indirectly harmed by CBT, because they will lose trust in the clinician/therapist and disappear from the clinic's books.

  34. Actually, you can't say that individuals fluctuate on the basis of group mean fluctuations, unless you believe the highly unlikely scenario that all of the patients (who were presumably recruited at different times) fluctuated in syncrony! Looking at the group data I can't see much in the way of fluctuations in the group means anyway. And presumably, "I would argue that their data show, for the first time, how patients who choose to be unmedicated display fluctuations in symptomatology" means to make the claim that the fluctuations are somehow connected to the patients' refusal of meds, and you do need a medicated control group to show that. Otherwise, why make the claim (we all know that patients' symptoms fluctuate even when they are on meds).Nope, this is definitely a silly claim.

  35. Anonymous does not seem to be aware of the intense ethical scrutiny that a trial receives both before it begins (it has to pass a NHS Ethics committee and then be approved by the research governance committees of all of the participating trusts) or afterwards (monitoring by an independant trial steering committee and an equally independent data monitoring an ethics committee which looks for evidence of adverse outcomes and has the power to close down a trial if risks become evident).That said, it's obviously important to keep an eye on suicide risk. Obviously, its not possible to infer anything about the dangers of CBT from the one suicide in the Lancet trial, which might have occurred for any number of other reasons. The only thing we can do is look to other trials, and I'm not aware of any evidence from them that CBT is linked to suicide (in SoCRATES, for example, 1 patient in the CBT condition commited suicide whereas 2 commited suicide in the control supportive counselling group).

  36. TAU is conservative because the responsible clinicians are free to provide their patients with whatever treatment they think is optimal. Hence, it's meant to be the best care within the existing system of care. Bizarrely you might think (from a scientific point of view) this means that the responsible clinician can refer the patient for some kind of psychotherapy (it is considered unethical to withold a treatment from the TAU group if the treatment is available) which will obviously reduce the likelihood of detecting a difference in favour of the active treatment. (Trialists usually hope that this doesn't happen very much and it probably doesn't, especially when the active treatment is not widely available.)The effects of the TUA arm should therefore be exactly the effects of standard NHS treatment provided under optimal conditions. If you assume that NHS treatment is beneficial you should assume that TAU is more effective than doing nothing (and doing nothing is never used as a control because that clearly would be unethical).The issue of placebo controls in psychotherapy trials is a complex one, and is addressed at length in my book Doctoring the Mind (Penguin, 2009). Briefly, in drug trials a placebo is meant to control for the clinician's tendency to impart non-specific psychological benefits such as optimism and a feeling of being cared for. But these are essential ingredients of any psychotherapy so can't be controlled for in the same way. Hence, many psychotherapy trials simply compare psychotherapy+TAU with TAU.The alternative is to use another kind of psychotherapy as a control. For example, in the SoCRATES trial of psychotherapy for first episode schizophrenia (Tarrier, N., et al.. 18-month follow-up of a randomized controlled trial of cognitive-behaviour therapy in first episode and early schizophrenia. British Journal of Psychiatry, 184, 231-239, 2004) some patients got CBT+TAU, some got Rogerian counselling+TAU, and some got just TAU. Both psychotherapy groups did better than TAU at 18-month follow-up – which is pretty amazing given that the patients received quite a short period of psychotherapy when acutely ill.And this allows me to correct an apparently widely-held misunderstanding about my own position. If you look at Doctoring the Mind, you will see that I am not an uncritical advocate of CBTp. I think that it helps some but not all patients but then so do other forms of psychotherapy.Looking at the psychotherapy literature as a whole, it is difficult to reject what has become known as the Dodo bird conjecture (named after a character in Alice in Wonderland; see Bruce Wampold's excellent if somewhat old book on the topic, The great psychotherapy debate: Models, methods and findings. Mahwah, NJ: Laurence Erlbaum Associates, 2001). The conjecture states that all psychotherapies work (beat doing nothing) but no one kind of therapy beats all the others. I think we can see the conjecture receiving some support from those trials which have included not only CBTp but also some other kind of therapy.Does the Dodo bird conjecture matter? Well, maybe it does if you are a die-hard CBTp advocate who wants the NHS to provide this kind of therapy and no other. But, despite rumours to the contrary, that's not my position. I just want NHS patients to have access to talking therapies which they may or may not find useful. And I think CBTp falls into this category.

  37. Jon, I'm sorry you think I'm just discussing anecdotes. I can't see any anecdotes in my posts. I think I've been discussing data. My position is that there is evidence from trials that CBTp helps some but not all patients. I also can't see much evidence of harm. I therefore think patients should have this option available. But I was pointing out that, in practice, clinical judgments are uncertain. No treatment works for everyone. Hence, a clinician has to ask him/herself 'given what we know from the trial data, is this treatment worth a shot with this patient?' I think that, in the case case of CBTp, given the evidence available, for many but not all patients 'yes' will be a reasonable answer.

  38. Regarding "The idea that CBT relationships are "trusting & respectful" is shark-jumpingly laughable" from annonymous.Actually, I have data on this (from the SoCRATES trial of 316 first or second episode acutely ill patients randomised to CBTp+TAU, Rogerian therapy + TAU or just TAU (Tarrier, et al., 18-month follow-up of a randomized controlled trial of cognitive-behaviour therapy in first episode and early schizophrenia. British Journal of Psychiatry, 184, 231-239, 2004). Patients completed a confidential questionnaire about the therapeutic alliance at the end of their fourth session.The results are currently in submission to a journal and need to be peer-reviewed. However, in our analysis we claim that the therapeutic alliance was often but not always good, and that those patients with a good therapeutic alliance had a much better outcome than those with a poor therapeutic alliance. More controversially, perhaps, we will be claiming that the new statistical techniques which we developed especially for the purpose (Dunn, G, & Bentall, R.P. (2007). Modelling treatment-effect heterogeneity in randomized controlled trials of complex interventions (psychological treatments). Statistics in Medicine, 26, 4719-4745) show that the therapeutic alliance is causal in determining outcome, and that the effect is not merely a consequence of hidden covariates.

  39. Hello to all,I have attended the Maudsley debate and posted a pamphlet-blog-entry on the MadinAmerica blog. I am writing as a person using both her life time education – based on clinical and cultural anthropology (as psychosocial dynamic & dialectic approach to understand persons interacting in socio-cultural group realities) and as a survivor of extreme experiences, mentally, emotionally, bodily of challenging and disruptive nature (with trauma and dehumanization-related content that I was told to be delusional only). I will repeat here my main criticism- the concept and nature of CBT 'therapy' had not been discussed (in Maudsley debate) – it seems to depend on the humane experience of the therapist and his/her desire and capability to support the person with 'psychotic experience' to 'make sense of' her/his experience – the danger of 1rst/ mostly white middle-class role models given the therapeutic power to be guides using 'higher psy & CBT knowledge' for/over the 'clients' as subordinate and 'confused, deluded….' help-seeker- the 'truth' to use concepts and training (talks?) based on individualisation and mentalisation of the 'clients' 'problems' with de-contextualisation of real structural violence and injustice as background to actor-committed violence, oppression, menace, distorsion, threatening … in the lives of people with extreme 'distress' and 'mental survival' responses to terrifying real and unjust circumstances – the ongoing exclusion of the experience of people in recovery from providing valuable support that is more than self help (for those who know how existencially life saving it is to exchange with others who have gone through extreme states and are well able to contextualize, make understandable the 'altered experiences' and exchange on many possible ways to restaure valid human dignity with the difficulties of living in injust conditions as to voice 'righteous anger' in focused ways and to relate to historical injustice and social activism as accessible real world groups for the active restauration of personhood) – the lessons learned from the Hearing Voices Movement and the Paranoia Network with regards to elaborating a meaningful scaffolding though exchanges of social and biographic experiences of multiple participants which introduces validation and equality and safe spaces created by the honesty and diversity of peoples destinies and narratives as well as introduces different approaches for a mentally-socially embedded understanding of extreme experiencestbc

  40. continued:I agree with Peter Kinderman that the outcomes in the trials need be challenged and I say tey need be changed; positive and negative symptoms – the reifications and re-appropiations by researchers and psycho-doxa of people's experiences – are terms which express the distant ignorance of observers (sorry folks, I know some professionals get more or less of it, but then no more symptom talk) .- Positive symptoms are not less ideations, voices, feelings, associations, meanings even though verrrrrry exxxxxxtreme than the thoughts, feelings, strange ideas that any professional or other person experiences in a stressfull and menacing life situation (being exposed to the terror of the most powerful authorities or structural injustice, for the existencial terror of this) – as extremely meaningful responses and messangers from the past these outcomes should not be expected to decrease after 6 to 10 'talks' – even God understands this is an impossible demand.- Negative symptoms … I always ask how to withdraw from a world that does no longer exist (that merits to be understood) in a social trustful or bearable way and has immense impact of bodily exhaustion and perception… symptoms may delude the observer. But then who does not know how the world disappears after it has been shattered the x-time and the spirit just gives in… or perhaps starting 'talks' with a stranger with psy mental higher knowledge is really distressing and voices can louder and the soul doesn't want be agitated and tortured again… perhaps after all so many fears and horrors that one just hides in the nowhere world – but yes that can become more hopeful… with some time for trust to build ok, more than 10 sessions would be nice, cause finally, psy stranger, nobody knows what you really plan to 'talk' and train me, may be that is just another strategy of distorting my shattered memories and wild ideas, visions or voices.In my view significant changes should be offered to the distressed person who shall be given a choice of- with whom 'talk' (if talking is CBT would that be NICE) and- wether the assumptions of CBT technique feel right with the 'client' – talk with a experienced, trained, inter-vised peer – attend to co-facilitated (peer/professional) recovery groups, sit in and decide when to become active in groups of people struggling with their extreme experiences- groups offered for the humanizing processes which I. YALOM has described so very well- groups offered' to overcome the power and knowledge imbalance introduced by seemingly higher psycho-mechanism knowledge hold by psy professionals, which is not true, unless they had accumulated experiences with learning from extreme social and cultural vulnerabilisation/oppression/abuse/exclusion/threateningof being etc and of having their lives disrupted by extreme experiences and continuing to recover from the psychological, emotional and social impacts of these experiences which is what 'experts by experiences' and Hearing Voices/Paranoia group facilitators have to offer besides mostly just facilitating a safe and diverse group process, cause …as YALOM as so well written or Jacqui Dillon and Gail Hornstein have written about theirs long time facilitation of Hearing Voices support groups in the Journal ‘Psychosis’IMPORTANT I wish to add here what has positively troubled me after my MiA blog pamphlet. I had listened to Peter Kinderman the first time when I attended to the Maudsley debate on CBTp. I feel I have not responded to him. I got the impression that Peter Kinderman is a very respectful and warm and fine person and I think that for many distressed people it can be a supportive and helpful experience to talk with him. I trust my intuition cause it is a fine human 'instrument'. This does not disempower my below writing. It is an act of respect and honesty towards what and how Peter Kinderman shared in the Maudsley discussion from his (CBT) talks/being with people in distress.Kind regards.

  41. @Anonymous, your statement "Hundreds of recorded RCT in relation to antipsychotic medication and no deaths" is simply errant nonsense. Go and check you facts before commenting.

  42. Although I hesitate to feed the anonymous trolls, some facts might be helpful. According to the Lancet paper: "We recorded eight serious adverse events: two in patients in the cognitive therapy group (one attempted overdose and one patient presenting risk to others, both after therapy), and six in those in the treatment as usual group (two deaths, both of which were deemed unrelated to trial participation or mental health; three compulsory admissions to hospital for treatment under the mental health act; and one attempted overdose)." http://www.sciencedirect.com/science/article/pii/S0140673613622461

  43. What is this linking to clinical psychology about? It's simply bizzare. There are many other professions that use, advocate and research CBTp. (including my own one of nursing) Someone earlier said that the critique was and weaked by overstatement. This is a prime example.

  44. Respectful warmth, belief, attitude, and non-imposing are part of the human qualities which to my mind make the biggest difference in therapeutic relationships, that's more important than any intervention. What HVN etc stand for is most precious but they too need to be careful to not fall into the trap of being prescriptive

  45. Many thanks to the commenter on this post. I stand by my original comment (at the top of the thread) though that I wonder how well the public are served when debates of this sort get polarised and claims/criticism overstated. It seems clear from this thread that, while CBTp may not be the holy grail or, for some, a realistic alternative to medication there is cogent evidence that it may have some benefits for some (some!) experiencing such the kinds of difficulties we are discussing. That isn't to say that those delivering therapies are not sometimes unrealistic about the effect they are having and, for broad question of effectiveness vs treatment as usual, it is extremely important for therapy providers to show some humility. This either or debate however seems to me to short change the public and close down both potentially useful avenues and existing combinations and collaborations between treatment.

  46. While we're on it I do wonder about that recent meta analysis of CBTp. Seems to consider/conflate a whole bunch of different CBTp treatments and not just the ones in NICE guidelines. Seems a bit of an error given the kinds of conclusions that were drawn in Guardian etc.

  47. This point about other conditions is a good one I think. One of the issues in assessing psychotherapies via RCTs more generally is that it is not always clear what is producing the effect (specific or non specific factors). there are a few studies of CBTp which employ non-spcific counselling control groups. These don't do too badly actually.

  48. The comments by anonymous above are clearly designed to provoke anger. Mr Byrne did the right thing by withdrawing from the discussion. I admired his restrained and civilised responses to what was an overt attempt at trolling.

  49. An important difference between medication and psychotherapy is that the latter, if done well, takes the experiences of the person seriously, as a meaningful aspect of life. The conflicting results of trials of CBTp may relate to whether CBTp is performed as a technical, manualised treatment, or as a skilled and very individualised collaboration between two people. Both can be called CBTp and the papers published from trials very rarely give you much idea of what actually took place.

  50. Ben – If you question the inclusion of some studies, you can go to the Jauhar et al database website – http://www.cbtinschizophrenia.com/ and remove the studies that you disagree upon, and re-run the analyses in Excel to see if it makes a differenceThis seems to be an excellent idea with regard to any future meta analyses in this area or elsewhere – and would address your point

  51. Professor Coyne prefers animus to evidence. Most of his points wither under scrutiny, but he relies on most people being too lazy to go to source.

  52. As a user/survivor and a foreign student in Mental Health Services research methodology I am astonished as how little the designs of the CBT trials inlcuded in the Jauhar et al Syst. Review and Meta-Analysis are being discussed. Is there such a high agreement on the manualized components, the number and frequency of sessions, possible adapation to the diverse client samples (age, ethnicity, social network/support, type of residence, education, un/employment, SES, QoL, duration of 'psychosis'/schizphrenia/spectum/high-risk, MHA/CTO hospitalizations in the last year, etc.)?Is the end-of-study measurement 'clinically' AND meaningful to the clients (not just statistically easiest data comparison endpoint)?Are the outcome measures positive/negative/overall symptoms 'clinically' meaningful AND acceptable and meaningful to the clients?Apart from discussion on blinding of outcome assessors, have the clients been able to relate their experiences at different assessments/measurements to the 'general wording' of questionnaires/interviews? Have qualitative data been collected to assess the 'degree in sympomology' the scales/clinical interviews are supposed to measure? Have the clients been asked wether they understand the assessment procedures and do relate them to their answers?I insert this as it is known how often the experiences of clients are not reliably and validly assessed in measures based on general and abstract categories.There are many relevant questions still. The methodological rigour may want for masked assessment whereas for the client it may be hard to relate when being interviewed by unknown assessors, especially of standardized interviews/scales don't allow for meaning making processes to happen. So how is the quality of interviews for outcome assessment monitored?In a much more nuanced and academic language Rory Burne has highlighted essential issues concerning the CBT intervention- and trial-design as well as internal validity of what outcomes are assessed in his response to the Jahaur et al Syst. Review and Meta Analysis.Rory E. ByrneRE: Cognitive-behavioural therapy for the symptoms of schizophrenia BJP published online February 7, 2014 In future, and for ethical reasons, I suggest patient representatives be part of trials' steering group or user researchers with SMI experience part of the research teams so that they can assure the importance of understanding the clients perspectives, priorities, and personal recovery needs are reflected in the choice of outcomes; that user advocates or user researchers can facilitate clients experiences with and struggles/acceptability of the procedures of outcome assessments, and more.If mental health services need undergo some transformation to really support clients with psychotic experiences in their step-wise development of personal recovery (which may well prioritize coping with unusual experiences and beliefs, decreasing distress by increasing hope and empowerment through social group normalisation, etc) then the design of trials, the choice of outcomes and measures, and the careful interpretation of results based on both qualitative data and quantitative data should become standard. 'The wider debate is over whether psychological perspectives challenge traditional thinking and enhance our understanding of mental health', writes Peter Kinderman. It would be fantastic if more users/clients had a serious say in the decisions about relevant issues in mental health services and therefore in mental healt services/interventions research so that benefits and efficiency of therapies and support increase, negative experiences/adverse effects and unneccessary costs decrease.

  53. I think Ben has named the large mammal in the room here. The less helpful aspects of the debate are driven, at least partly, by a strident anti-clinical psychology rhetoric. This is a notable feature of both James Coyne’s and Keith Laws’ Twitter and blog accounts. The official justification is that the profession is ‘anti-science’, but the comments are framed in remarkably unscientific terms, and often consist simply and solely of attacks on individuals and the profession as a whole.CBTp is an odd battleground to choose, since it is a prime example of constructive collaboration between psychologists and psychiatrists – and indeed nurses, as Ben says. The fact that psychiatrist David Kingdon was also defending it at the Maudsley somehow seems to have faded into insignificance. It seems to be more important, for some people, to berate clinical psychologists for the crime of creating division between professions – while stereotyping their views and practices in a way that is surely unlikely to increase harmony. Why has the relatively small profession of clinical psychology become a target in this way? It cannot be unconnected to their recent position statement on classification, calling for alternatives to psychiatric diagnosis. By putting their heads above the parapet, they seem to have become fair game (although the Critical Psychiatry Network, which issued a supporting statement, seems to have escaped blame.) I guess this is a tribute, in a way, to the impact the profession is having. It is also, perhaps, a foretaste of what we can expect as these views become more mainstream. In the meantime I hope that clinical psychologists will continue to demonstrate their commitment to rigorous, clinically-relevant scientific investigation – impressively illustrated by Richard Bentall on this blog – and also be willing to promote other kinds of evidence that would doubtless fail the Coyne/Laws test: qualitative studies, practice-based evidence and service-user-led research.

  54. I end up with really mixed feelings about what James Coyne says. He makes some good points about evidence but they all too often seem to get lost in aggressive invective and a very unfortunate turn of phrase (that business about RCTs). We need people like him but he undermines his own points with this kind of stuff.

  55. Thank you. I thought I’d been civilised, so was surprised to learn today that to at least one observer I’d gone ‘ballistic’ http://i.imgur.com/9X884tQ.png and ‘unleashed a lot of vituperation’ http://i.imgur.com/7WRhC5R.png (meaning I’d been abusive, harsh, or angry). Those comments struck me as remarkably dishonest (‘retracted’? ‘quickly’?), so I asked the blog admin here if my deleted comments could be recovered to let anyone interested decide for themselves. They can’t (thank you though), but comments still appear in emails, screenshots here: http://i.imgur.com/zF2IDsl.pnghttp://i.imgur.com/NQY3y69.pnghttp://i.imgur.com/0hPFXlH.pnghttp://i.imgur.com/HVHBr6e.pnghttp://i.imgur.com/i5YqebV.pngI don’t think I ‘went ballistic’ or was ‘vituperative’, and I know I didn’t retract anything, or quickly. I’d like to ask James Coyne if he stands by the two personal comments above.

  56. Hi Ute, I think that’s a great comment. You very kindly mentioned my brief response to the Jauhar et al analysis; well in turn I’d like to recommend your comment as being far richer. I think you highlight exactly the factors that deserve the most attention. I’m in the very fortunate position of working with researchers who do as you’ve suggested by including members with experience of mental health difficulties in research teams and steering committees, so I know they agree with you too. I have a journal article from 2003 written by Prof Bentall and my boss Prof Morrison, where they say: “We suggest that one way of addressing the ethical issues is to involve service users in the design of trials… By consulting people who have experience of psychosis, or some of the high-risk participants in the trials that have been completed, it should be possible to agree on a set of outcome measures that are meaningful to those who are offered preventative interventions.” I can’t see a link to the article but is ‘Response to the commentaries’, here http://informahealthcare.com/toc/jmh/12/4 And I’m one of those people. I was a participant in this trial of theirs http://bjp.rcpsych.org/content/185/4/291.long and, a few years later, was a colleague of theirs in this one http://www.bmj.com/content/344/bmj.e2233 I also agree with your discussion of methodological factors to be addressed, we certainly have more to learn from participants and ongoing developments. I think CBTp researchers in most cases have had to measure primary ‘symptoms’ in trials because the longer-established and dominant psychiatric model has done so, almost exclusively. But CBTp studies have increasingly also measured ‘secondary’ or qualitative outcomes (the 'Choice' measure for example http://schizophreniabulletin.oxfordjournals.org/content/36/1/126.full), and hopefully will continue to evolve in this way, always learning from those we’re working with.

  57. Rory, as the person you was replying to, I never felt you had gone ballistic or even unleashed a lot of vituperation. Sadly, I had not taken this discussion as a serious one (as so many of them are not), but as more of a bar room banter or more accurately a bet to see if you would retract or any statements by me would be deleted – james is correct. Further to this I am aware that within at least sixteen RCT of antipsychotic treatments was a factor in the death of vulnerable patients such as the elderly. The idea behind an inclusion was to test if anybody would spot a posited argument by Keith laws. However, on issues of ethical responsibility/ scrutiny I do stand by my comments, with the following caveat I would like to see the same level of investigation assigned to this study generalised across all RCT that involves vulnerable groups Given this, I would like to offer my apology.

  58. "A posited argument by Keith Laws" – really?I have never said anything about a lack of deaths in RCTs of antispychotics – so please dont attribute your poor excuse -for your own words -to me"Bar room banter"? You should take the discussions about serious mental illness on a board such as this….seriously

  59. Richardwhy do you think that the individuals need to fluctuate "in synchrony" (as you say)And when you say "presumably recruited at different times" – how do you see that as relevant in that study?What I actualy said was "I would argue that their data show, for the first time, how patients who choose to be unmedicated display fluctuations in symptomatology (as we might expect given they are unmedicated) …but crucially, these fluctuations are as large as the changes seen in the CBT group. Hence, it is reasonable to ask…have Morrison et al simply documented 'normal fluctuation' in the symptomatology of unmedicated patients …and nothing to do with CBT"http://keithsneuroblog.blogspot.co.uk/2014/02/my-bloody-valentine-cbt-for-unmedicated.htmlSo, I am saying it is reasinabvle to ask the question… we have a control – a comparison of change in those who do and dont receive CBT. My point is that the 'control' group show an 'improvement' in symptoms by 9 months that is not significantly different from that of the CBT group by the end of the trial; and indeed, the two groups do not differ at the end of the intervention (9 months).The controls then show a further negative change between 9 and 18 months, while the CBT group then remain unchanged across that time period – so the difference at the end reflects the natural change occuring in the controls

  60. In denying the relative potential harmfulness of unproven psychological interventions (e.g., psychotherapy) versus unproven physical interventions (e.g., lobotomies), are you advocating for some version of the old "sticks an stones may break my bones but names will never hurt me" idea? If there is such a thing as an emotional wound or mental condition that can benefit from some kinds of treatment, then there are emotional wounds or mental conditions that can be worsened or iatrogenically created from some kinds of treatment. If you don't know on the basis of well-designed replicated scientific evidence what the short and long term effects of a treatment are likely to be, then you're taking an unjustified risk and may be hurting too many people. That is unethical. Such potentially harmful clinical interventions ought to be discontinued until there is valid replicable scientific evidence to support it. The burden of proof is on she who asserts a positive (i.e., she who asserts that an intervention if safe and effective), not on critics to prove a null hypothesis. The history of science and medicine tells us it's better to err on the conservative side in the absence of reliable evidence. Less is more.In the trial you discuss, what were the criteria for adverse reactions? How was it insured that all variations of decline were captured? What, if any, procedures and standards were applied by the DMEC to ensure blinding and rule out bias, allegiance effects, etc.? Do DMEC members have any conflicts of interest? What adverse reactions were reported, and what adverse reactions went unreported? How many people in the trial got worse or dropped out? How many trials has your committee closed down due to perceived harm to subjects? How does the adverse reaction rate for the interventions you've tested in trials compare to the adverse reaction rates for therapies that have been released to the public? What are the standards and measures for ensuring public safety, and how rigorous/scientific and transparent are they?

  61. Richardregarding "…TAU as a control tends to be conservative – it reduces the chances of finding a significant difference between groups."With what are you comparing this? Of course, the treatment may change for anyone in an RCT – and they often do (including medications) – but what alternative is there? And as you say, …it seems unlikley to happen, If something the clinician does, reduces the power to detect an effect, then we have another intervention worth considering – as it would be seem just as efficacious as CBT (in this example)Hence your point about the Dodo bird conjecture. As it stands, Jauhar et al documents no significant differences when comapring effect sizes of CBTp vs TAU trials and CBTp vs active controls (such as befriending/supportive counselling).So, whether the control arm is 'doing something' or 'TAU', there is no difference in effect size – so power does not seem impacted in this comparison

  62. PeterIt would be good to hear you expand your views on our meta analysis – you twice here quote (correctly… if somewhat selectively) from our abstract and have repeated the same quote on Twitter.As Jon Denberry rightly further quotes the next sentence – the already small effect size diminishes further when blinding is considered – and in fact becomes nonsignificant both for positive symptoms (ES = .08) and for negative (ES = .04) symptoms (and nonsignificant for hallucinations). In fact the effect sizes are up to 7 times larger in unblinded than blind trials. A very small effect remains for overall symptoms (ES=.15) although the confidence intervals do not inspire 'confidence' I would be interested to hear your thoughts on the importance of blinding in assessing CBT for psychosis – and if you think it important, then why not quote our relevant findings? – the findings do come from 20 blinded trials compared to 10 nonblinded – so they are not a spurious or small comparison by any means

  63. If symptoms of individuals fluctuate then I see two potential issues:1) When are measurements taken – If trial subjects are given a choice about when a measurement is taken (for example when a survey is filled in) will they choose to do so at a particular point in there fluctuation cycle for example on a good day. Then are measurement windows the same for each of the measurement points within the trial of could having different measurement windows lead to overall improvements being seen. Also could other factors in the design such as expectation affect when measurements are taken within the window.2) If there is serial correlation within the random fluctuations then this will affect the underlying assumptions used in statistical tests. I know in statistical modelling work I did many years ago serial correlation from sensor readings had quite an effect on the fit of a model. Given many diseases fluctuate I'm surprised this effect isn't discussed.

  64. Anon – you would need to ask authors of the RCTs (its not typically reported in papers), but I assume it 'should' be fairly random across both CBT and control groups – though of course, may not be random at all in unblinded trials

  65. Given many diseases fluctuate I'm surprised that I couldn't find stats papers on fluctuating random processes although I think economists may have studied them. It may be all groups in an RCT vary in the same way but if the underlying mechanisms invalidate the assumptions of the significance tests it could be significance if found when it shouldn't be found. I don't like to rely on intuition where there are potentially complex non linear combinations of factors. So I was thinking the issue of fluctuation could be easily explored using a stochastic petrinet with states based on a powerset of questionaire reports or symptoms then having different transitions based on different ideas of disease variation. This model could then be used to derive population statistics (through simulation) and this could be used to test the reliability of significance testing under different assumptions about disease variation and measurement factors. Given in many trials have small effects I think it is important to understand how much we can rely on the accuracy of any measurement system. Engineers realized many years ago that adding more significant figures to a measurement doesn't increase accuracy. If the error bounds of a measurement system are quite large and increased by measurement errors in fluctuation it makes it hard to draw reliable conclusions on small effects. I think any RCT should discuss operating characteristics of the measurement system they are using but I've never seen this done!

  66. What is important for the person who has psychosis?I know that the results to meta-analyses in elitist academic journals are not. I know that aggressive and personal arguments by high profile academics don't help them.I know that from clinical experience it is entirely dependent on the individual service user, what method is best for them. Whilst I am no proponent of CBTp as a blanket intervention (nor of CBT as a blanket intervention for anything) it has worked well for some service users, however I have generally found better outcomes (that are meaningful to the service user, not some irrelevant scoring system) using integrative and third wave interventions. Moreover, genuine systemic practices involving multidisciplinary teams, and crucially families and the service user's real worlds I have found even more effective. I would personally like to further my use and see a greater UK focus in research into the extensively evidenced and extremely effective [in comparison to CBT and medication] "open dialogue" methods used in Finland.There's no way of doing RCTs (and meta-analyses) of such sensitive and service user central interventions, they are all different. It is simply not possible to compare what should be completely individualised interventions. Hierarchy of evidence is a concept that can work well with some biological conditions (all tablets contain the same chemical, all therapies are not 100% the same), but it is well established now that psychosis has greater links to social and traumatic factors. The evidence that matters is the evidence for each individual service user, and if that means that experienced clinicians feels that part of that care should be CBT then what would be more useful would be the reporting of single case work that was successful (and crucially what was not).More integrated, regular, and "as standard" reporting of clinical work by those carrying out psychological/psychosocial interventions are what are needed in my view. And for the people who read them not to get caught up with wasting their time arguing over semantics of scientific papers, but actually drawing together what factors might be the ones they could sensitively consider when conducting their own clinical work, from the experiences of others.I would argue that that a true scientific approach is actually one where the clinician really lives the identity of a scientist practitioner as they do their job, rather than relying on comparisons of therapies, comparisons of RCTs and comparisons of meta-analyses by academics who appear to have lost touch with what is actually the point of the research in the first place: which is to help find the best methods for helping people experiencing psychosis to get the outcomes that they want and require.I would argue that the academic world needs to take a step back into the world of the service user and re-think what "hierarchy of evidence" would be more appropriate to be spending their time debating, arguing and in my opinion wasting valuable skills over.I would also argue that there needs to be more readily accessible channels for clinicians of all disciplines to submit and access single case work for example, than the time-intensive and political world of academic journal publication.I know as a clinical that I would value a review of such papers more so than an RCT or meta-analyses of RCTs.

  67. Benwhich studies did you think were conflated and not relevant to NICE guidelines?As anon says, we could remove and see what happens to the effects sizes – takes 30 seconds in our Excel program – as you say, it would be nice to see future meta analyses in this area do likewiseWere there some specific conclusions of mine n the Guardian piece that you object to? http://www.theguardian.com/science/sifting-the-evidence/2014/apr/02/has-cognitive-behavioural-therapy-for-psychosis-been-oversold

  68. For a clinician to ask this question they need to be presented with clear data in a very concise manner. This is where misleading abstracts are particularly bad. A clinician may read an abstract (not even having access to the full paper) then they get the wrong impression of the trials success and base clinical judgements on an abstract which has been written to over inflate the results of a trial – presumably to allow researchers to claim success. The clinician also needs to communicate the risk reward tradeoffs to their patient.I think people are generally quite poor and understanding risk reward tradeoffs and certainly make different decisions based on whether they are stated in a +ve or -ve manner. (see Kalmann and Tzversky). So again the presentation of results and ensuring they are stated in a neutral manner is important to clinical decision makers and patients.I think that as part of a trial protocol a results summary sheet should be prepared which will automatically be filled in from the trial database and published. (note this may include choice points around different stats tests based on the underlying distribution of the data). To my mind this should be reviewed by the ethics committee with the idea of this is how decision makers will be presented with results.I would go a little further and say there should be a third party service that would maintain the trial database locking access according to the protocol and automatically release such a results summary sheet the day the trial completes.Allowing researchers to wrap words around data often doesn't add to clarity but seems to influence judgement around the data. I've seem some very strange semantics applied to recovery and normal that require a very careful read of papers to understand.

  69. I would argue with the idea that anecdote is not data. It is data you may have issues with its quality and a set of anecdotes don't allow any notion of quantification but can be very important. With no theory of evidence behind evidence based medicine its hard to use them.Anecdotes around side effects can be particularly important. A trial of say 100 people has a small probability of picking up a side effect that occurs 1 in 10000 say. A number of anecdotes around a side effect suggest the existence (not quantification) of the side effect problem.

  70. I think it is comments like the one above that highlight the lack of understanding in general about how mental health difficulties originate and how they are experienced; and also the sarcastically implied rude dismissal of service user's perspectives for a condition that (most evidence suggests) is not medically originated (like prostate cancer is).

  71. Dear senior researchers and/or clinicians, dear fellow (ex)mh service users,It makes for an astonishing read how the service users experiences, needs and priorities are quoted by researchers and clinicians in some posts. I am curious about envisaged consequences.You may have noticed that Rory Byrne and me write as (ex) users/survivors, Rory Byrne as a user researcher, I am currently a MSc Mental Health Services Research student and enjoy my education at the Institute of Psychiatry, King’s College London, where the world’s unique Service User research section is located. I find there are promising developments with regards to complex intervention studies and the use of mixed methods. If some of you senior researchers had some reading suggestions for me to look further into the issues of evaluating complex interventions as process-models instead of component-models (I refer to Hawe et al (2004) Complex interventions: how “out of control” can a randomised controlled trial be? http://www.bmj.com/content/328/7455/1561 ) I would be grateful.Rory Byrne http://bjp.rcpsych.org/content/204/1/20/reply#bjprcpsych_el_55200 and me, both in posts here, have made some crucial conceptual points as well as methodological considerations taking into account the needs and experiences of a significant subgroup of users/consumers and which challenge the 'internal' validity of the trials/meta-analyses under debate.I suppose you are more cogent than me of the diversity of peer-reviewed and 'grey' research literature on human experiences like voice hearing/auditory hallucinations or paranoia, unusual and fixe beliefs. These are also different from 'high wired' psychotic experiences. I wonder wether these heterogeneous human experiences are all treated by similar CBTp therapies and all included in the same trials and meta-analyses? I read the qualitative study John McGowan posted and take away the impression of patients struggling with quite a range of 'psychosis'(?!) experiences. So I will focus on hearing voices. From rereading studies from the grey and from the academic literature yesterday, I again found agreement on the voice hearing experience as such changing very little (intensity, frequency, volume) be it after CBT or Hearing Voices peer group attendance. Qualitative and quantitative findings confirm that voice hearers experience other benefits from either group CBT our HV peer groups: normalisation, hope, new coping mechanisms, less stigma, socializing more and easier, some even taking up activities after years or planning to return to education. These are very significant outcomes related to recovery of personhood, more self-control, more social activities, sounds like approaching a ‘healthy lifestyle’. Therefore the traditional symptoms' outcomes need be challenged. Perhaps the clinical psychiatrists/psychologists and academic psychiatrists/psychologists majority 'belief' of the need to make 'voices' disappear can finally be questioned as part of an outdated concept of 'schizophrenia'. You will certainly know that an expert researcher as Marius Romme has been making this point too. You will know of the stimulating work Richard Bentall has been doing on more detailed understanding of ‘paranoia’ experiences, as well as other ‘continuum’-researchers. 'Negative symptoms' also need be deconstructed and differentially related to and explored as the actual problems and experiences in the heterogeneous group of patients (replacing the mixed schizophrenia cluster by the more modern sounding psychosis one brings zero advance in clarity and patient centeredness). tbc

  72. continued:There is a lot of knowledge in the academic literature and grey literature about what patients in respectful therapeutic alliances or in peer support groups can say about their changed understanding of unusual and extreme experiences, their struggle against devastating professional, social and self-stigma, their step-wise building of strategies for personal recovery. There are many unknown and few outstanding role models who have recovered from severe mental and/or affective disturbances, there is a growing number of recovery narratives from people with experience of SMI. This constitutes a wealth of existing or potential evidence which should, in my very strong ethical view, be scrutinized and conceptualized into group or individual therapies and peer support groups, the exciting Open Dialogue in social networks work, WRAP and recovery groups, or Recovery Colleges for mutual learning in groups. Different modalities of human interaction settings and exchanges may need be combined, as several authors rightly say, one size does not fit all. My point is that times have changed, we no longer live in Schneiderian or Bleulerian eras – todays clinicians and researchers have the historically unique opportunity to collaborate with people who are in recovery from severe breakdowns and disturbances. This demands important changes in the exploration and conceptualization of SMI experiences and appropriate therapeutic exchange settings. Isn’t that most researchers favorite challenge? Have you, dear senior researchers and clinicians ever embraced this historical opportunity: It is you who have eminent how-to knowledge and socio-professional executive power, as researchers and as clinicians, and it is you who can choose to invite people with lived experience of disturbance and being-in-recovery to together create what you like to call a paradigm shift. You may talk in cost-benefit-ratios, I may prefer deeply humane words like dignity, acknowledgement, mutual learning, healing of soul, mind, bodily functions, and subjective recovery. You can restaure the exploits in your professions by inviting new colleagues, you can also restaure much of the faith in the once admired and trusted professions of researchers and clinicians/therapists, you can help save lives which otherwise risk be wasted, you can even contribute to saving 'taxpayers' money in the long term. All it needs is a change of perspective: that co-production of therapies, support groups, reciprocal educational mental health recovery programmes, transformation of theories and changes of the constructs used in outcome measures are exciting and worthy endeavours. There are also exciting developments being made over the last years in complex intervention research and much can be gained by inviting a small group of well educated user/informal carer researchers for future developments. However, it seems, that the suggestions Rory Byrne and me have made in this blog seem uninteresting to the majority of experienced researchers and clinicians I enjoyed reading from here. I am truly interested in learning from you, read your opinions, and please be critical, I appreciate 'critical appraisal' and I enjoy discussions and, yes, even debates.Kind regards to all.

  73. AddendumA 'small group of user/ informal carer researchers' needs be annotated: I am proud as a surviver to see this small group growing in number, with some senior researchers being role models of excellent research, especially in England. Some most needed and relevant user research projects have been conducted by user researchs as independent researchers, often commissioned by charities and trusts. Using the quantitative qualifier 'small' only makes sense in comparision with concentional academic researchers. I may be trapped to use 'small' more easily than British user research professionals cause in my birth country, Germany, there are no user researchers in academia until today, there is only one independently working survivor researcher doing amazing work both nationally as internationally with the shamefully miniscule to non-existent funding by German charities.

  74. Richard,Sorry to be such a pedant here but it is quite simply a logical impossibility for a group mean to fluctuate without the individuals fluctuating. I wonder if we're meaning different things because this is simply basic arithmetic:(A+B)/2 = GroupMeanGroupMean cannot change unless A+B also change. This fluctuation does not have to be in synchrony. I apologise if I'm coming across as petty or stubborn but I really can't see how this is even a matter for debate.

  75. "Such potentially harmful clinical interventions ought to be discontinued until there is valid replicable scientific evidence to support it"I agree with your point that psychological treatments could potentially be harmful. I think as Richard Bentall says there is a growing appreciation of the need to monitor this. I'm not sure I've understood the rest of your point though. How would you evidence a therapy's safety if you discontinued it on the basis that it might be harmful? You can only prove a therapy's efficacy by doing it (in the safest possible way you know how at the time) and evaluating it, surely? I would agree that perhaps trials should publish data on the questions you go on to ask about but would it really be feasible to capture ALL variations of possible decline for clients? It seems to me that would be impossible to quantify when clients regularly state that they find it hard to put into words the ways in which therapy has made them feel (good and bad).

  76. Ute, another excellent contribution to this discussion, I think worth combining into a blog or article, I'll certainly refer back to these comments as my memory may miss something! And I agree with you wholeheartedly on all points. I agree most of all with a particular point. You say: “I wonder whether these heterogeneous human experiences are all treated by similar CBTp therapies and all included in the same trials and meta-analyses?” – I think this is a crucial question/s for therapists and analysts to consider. How much do particular ‘ingredients’ of CBTp help with different unusual and/or distressing experiences, and how sensitive are meta-analyses to measurement of the variety of CBTp ingredients delivered in each trial? As you say, there may also be significant benefits for people attending peer groups, and with such a wide range of potential therapeutic factors and personal outcomes to consider, how well does current meta-analysis explain the total picture?I hope it’s not true that suggestions such as these seem ‘uninteresting to the majority of experienced researchers and clinicians’. I think many of our colleagues will find your comments of real value, as I do, not least among your IoP group (indeed, world-leaders in learning from lived experience). Best wishes.

  77. Let's 'wait and see', Rory, wether our user RESEARCHER agruments, suggestions and challenges will be responded to. That we contribute as (junior) researchers needs be highligted. Isn't it peculiar that all throughout knowledge's history pivotal contributions were made by people using knowledge gained from margins and by opening up to new, essential aspects of phenomena. Not few of them may have been considered 'crazy' due to their 'out of the box' of normal discinpline's thinking and to being passionate about their subject. One of my anthropologist colleagues is intrigued by the role extreme or unusual experiences play in the history of scientific advances and artistic innovations. There may be some point as to how some consciousness challenging features of mental/affective disturbances can actually make some of us 'ex-crazys' become damned good scientists – obviously taking in the whole education every researcher undergoes and develops further during his/her productive career. So here's a thumps up between ourselves. Not to be over-played though, cause I think it needs a good degree of intellectual pathos & passion, as well as some mental grandiosity, to be a prolific, inspiring and innovative scientist -> we may all meet somewhere on the psychological continuum, with some coming for places afar. Most important, there is no reason, to not take us user researcher (for) serious. So let's keep being productive, supportive and open for discussion, reflection and debate. Controversy is as important as constructice collaboration. But hello, sounds of silence may only be enhancing the journey in musical spheres, lest so in and between sciences.

  78. It would be indeed unethical to offer a treatment with unknown long-term effects but that is not what is happening here. Trials of CBT typically have much longer and successful follow-ups (in SoCRATES we able to follow up 75% of the sample for 18 months) than drug trials (typically 6 week follow up with 50% lost to follow-up – see the numerous reviews of this by people like Ben Goldacre and David Healy). Psychotherapy trials (because they are almost invariably funded by government agencies such as the MRC or HTA) are usually much more methodologically sound and well-regualted than drug trials (which are usually funded by industry and which, of course, involves treatments which have a much greater capacity to harm). I think that what is being forgotten or ignored by Discussant on 15th of April is that (a) there is a huge literature on psychotherapy for non-psychotic problems – which, incidentally, many psychotic patients suffer from as well as their psychotic symptoms – and the relative benefits vs minimum risks associated with this kind of treatment are well documented; (b) there are noiw about 50 CBT trials with psychotic patients which provide no evidence of unacceptable risk associated with it – if CBT provoked suicide this would have been detected in the larger trials (e.g. SoCRATES) that have already been carried out. To suggest that CBT is a highly risky procedure is promoting a scare story that has no basis in reality. Safety (compared to medical intervrentions) is one of the advantages of psychological interventions (which, of course, is not say that patients are NEVER harmed by therapy, just that such harms seem to be very rare).Regarding the other issues raised:1. Trial teams are required to report to DMEC anything that might reasonably be thought of as potential adverse reaction, which might include sudden symptom exaserbations, dangerous behaviour, self-harm, suicide, hospital admission etc. 2. DMEC then looks at the surrounding documentation – and can seek further evidence if it likes – to make a determination of whether the adverse reaction could have been caused by the psychotherapy (very often there is some other explanation – the patient has suffered the sudden death of a close friend, say, or, as in a recent case I saw, a deterioration in functioning occurred between inception and the start of therapy, and the patient's participation in the trial facilitated early intervention).

  79. But for this to be a signal rather than noise in the data, they would have to fluctuate in syncrony! If the patients in group A have higher mean scores at time 1 than at time 2, and if the mean scores then 'fluctuate' upwards by time 3, they would have to all (or at least most of them) go up and down together. That's obviously nonesense. You're apparently a trainee psychiatrist – do your patients rhythmically oscillate in this way?Still silly!

  80. Ute and Rory and Anonymous 15.4.14 at 11.18 – thanks for taking the discussion in such a constructive direction. As a clinician working with psychosis, I wholeheartedly endorse what you say. Dismaying as some of the invective has been, I have been impressed and heartened by learning more about the close ethical scrutiny and inclusion of SU views that goes into the trials described by Richard Bentall. Although it’s not my own preferred approach, I can well believe the powerful personal testimonies/irrelevant and sentimental anecdotes (take your pick) about its helpfulness. And let’s face it, as Peter Kinderman says, we’re only arguing for offering it – not holding people down and injecting it into their buttocks.You are right about the exciting developments on the horizon for psychosis. If your main concern is solid evidence about outcomes, look at Open Dialogue, the extensively-evaluated approach based on communication and shared meaning-making which is currently achieving the best results in Western Europe. If you are more likely to be swayed by the personal testimony of ex service users who now lead full lives alongside their unusual experiences, listen to the inspiring stories of Ron Coleman, Rufus May, Jacqui Dillon, Rachel Waddingham, Eleanor Longden, Peter Bullimore, Louise Pembroke, Flo Bellamy (most of whom were originally diagnosed with schizophrenia), and many others. If you want the best of both worlds, we can escape the either/or battle and integrate service user perspectives into research in the creative ways you describe. My main reservation about (some of) the CBTp research is that it hasn’t entirely managed to escape biomedical model assumptions. I agree that we need to move from notions of diagnosis to narrative, curing ‘symptoms’ to increasing quality of life, and treating to healing. Thanks again for your thoughts.

  81. If the control group's average improved at the 9-month mark compared to baseline, then the individuals in the control group (from which the group mean was calculated) by definition improved as well.

  82. Hi,The Juahar et al 2014 metaanalysis is interesting, but it's only one of a number of analyses that have broadly concluded the same thing – that CBTp has either a 'moderate' or 'small' but positive and replicable effect size in heeling people with psychotic experiences (and we know that this is a complex issue, with many aspects of outcome assessment still to be resolved). The findings (as I read them) are as follows:First, with a set of inclusion and exclusion criteria that have themselves been criticised, the 'headline' effect size in the Jauhar paper is statistically significant and in the 'small' range. As I say, the inclusion and exclusion criteria for that study have been picked over to some extent – good, as I say, for forensic analysis.Richard Bentall has referred to the post-hoc fort her analysis as 'rotten cherry-picking' – by which he means that if certain studies are deleted, the overall effect size will change. That's why we prefer to rely on hypotheses set out in advance, and why I prefer the published results rather than post-hoc deconstruction in blogs. On the issue of blinding per se, as someone who has conducted RCTs in this area and has on-going clinical responsibilities, I think the jury is out. Blinding is a complex issue when your discussing a person not being aware of whether or not they've received a therapy; it is certainly methodologically difficult and could be considered ethically dubious. Moreover, as someone pointed out at the Mausdley debate, blinding could have untended consequences.So… given that these caveats (problems with post-hoc 'rotten cherry-picking' and difficulties with blinding) I note your (the authors') contention that the effect size in the Juahar paper reduces further beyond the 'headline' figure when certain questionable assumptions are made… but, clearly, I am not convinced.Overall, I am more convinced by a series of meta-analysis all broadly concluding that CBTp has a small or modest effect. That seems the most reasonable conclusion thus far, but I agree that we should continue to research the area and to use as much scientific probity as we can muster.And.. while the evidence points thus, I think it would be wrong to change the clinical guidelines – that CBTp should be available to those who could benefit.

  83. Hi1) Perhaps you could be specific and tell us which meta analyses have the ‘same conclusion’ as Jauhar et al – please give outcome measures and effect sizes – because this is simply untrue – perhaps you havnt read our meta analysis closely?2) The inclusion and exclusion criteria have not been criticised – again please be specific and reference your claim and say which inclusions and exclusions3) I don’t follow what you are talking about re cherry picking – please elaborate how its relevant to our meta analysis4)Your view on blinding is what? Is it important or not? What unintended consequences specifically? – I answered the comment at the Maudsley Debate – so please, what consequences specifically?5)You are clearly “not convinced” by ‘evidence’ and have provided no argument here 6)You conclude by saying “Overall, I am more convinced by a series of meta-analysis all broadly concluding that CBTp has a small or modest effect. Please say which meta analyses and why you prefer those meta analyses with seemingly more positive (small) outcomes, but which have fewer studies and less control over risk for bias? 7) Finally again, how difficult is it….it is not ‘Juahar’

  84. Looking back on the posts here, I am struck by how little critics of CBTp understand: (a) about the process of conducting RCTs (misunderstandings of TAU and trial entry criteria by Coyne; the false implication that many adverse reactions to CBTp could be systematically ignored by trialists and by DMECs); (b) about CBTp itself (Prof Law's claim on his Twitter feed that CBTp doesn't involve a model, and his worry that CBTp theory cannot predict the 'dose' of therapy required which, of course, no one who was knowledgeable of the field would expected it to do); (c) about the wider psychotherapy literature (e.g. on the therapeutic alliance); and (d) about the messy and challenging NHS environment in which both clinical trials and individual treatment decisions are made (including the misconception that CBTp is somehow an instrument for the professional advancement of clinical psychology whereas, in reality, it is practiced by other mental health professionals also). I suspect that this poor understanding reflects the fact that, in many cases, the critics are not clinicians, have no training or experience in any kind of psychotherapy, and have never tried to conduct a proper RCT themselves (which is an arduous, expensive and highly regulated business).To someone who has worked quite hard, over a long period of time, to promote better care for patients with psychosis, this is naturally frustating. I hope this frustration is understandable to others. It's particularly hard to hear all this labour dismissed by people who haven't themselves done anything tangible to improve NHS care (hence my St Peter quip earlier – which was a jokey way of making this point, but which seems to have provoked a complete sense-of-humour failure in some corners of the blogosphere).As I attempted to make clear in my 2009 Penguin book 'Doctoring the Mind', I am not an uncritical advocate of CBTp. It is not a pancea, that's for sure. I think it helps some people and not others, and that other types of psychotherapy can probably also sometimes be helpful. Regarding the overall effectiveness of CBTp, I agree with NICE's assesment (based on their own meta-analysis) that there are some benefits for positive symptoms, although these are not large on average (of course a small average response in all likelihood will translate to a large benefit for some patients). I am also not inclined to dismiss the large number of service users who have also argued for the greater availability of talking therapies (as evidenced by the recent Schizophrenia Commission Report). To dismiss these voices out of hand seems insulting to the lived experience of service users.As pointed out by some other posters, an important issue in all this is choice. In the not too distant past, and still in some services, patients are offered medication, with its known hazards, and absolutely nothing else. Very often, services still do not consult patients about what kind of help they want, or even inquire into their life stories (as Eleanor Longden has noted, patients are asked what is wrong with them rather than what has happened to them). I want services to be more human and patient-centred. Of course I do not want patients to be cajoled into any kind of therapy but I do want them to have the option of CBTp if they think it will be helpful. In short, I want to increase patient choice whereas, so far as I can see, Prof Laws and his followers want to limit it.

  85. Re 'dose' and models of CBTp, people can go and look at the Twitter discussion between Richard Bentall and myself from 2 nights back My point referred to whether there is a model for predicting how much CBT (sessions, duration, 'dose' has been used by some in this speculation – its not my choice of word)So, 'dose' refers to duration of intervention and how that might be relevant to CBT for psychosis – Perhaps Prof Bentall doesnt think duration is important, but others clearly seem to think it is e.g. Rory Byrne who has commented here amongst others such as Paul Hutton and Anthony Morrison – perhaps Rory could elaborate?see http://bjp.rcpsych.org/content/204/1/20/replyor http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42013003911orhttp://www.tandfonline.com/doi/abs/10.1080/17522439.2013.825005Also, NICE recommend a minimum of 16 sessions of CBT for psychosis – my question is why this duration ('dose') – seems a perfectly reasonable question to askPerhaps because RCTs have a accidentally arrived at a modal number of 16 sessions? But why do they? My question is what has led to this and why not explore other durations – is htere an underlying model to explain this duration?Indeed at the other extreme, Prof Kingdon has suggested using brief CBTp see http://summaries.cochrane.org/CD010646/brief-versus-standard-cognitive-behavioural-therapy-for-schizophreniaI am sorry that Professor Bentall often feels the need to avoid answering questions by being petty or trying to belittle (BTW the name is Laws)

  86. It could indeed be the (implausible case) that all of the patients fluctuated in their symptoms in the same direction at the same time (that they all went up between time 1 and time 2 and then all went down between time 2 and time 3). But it could also be that a single 'outlier' changed dramatically and the rest stayed the same, or that someone with a low score was missing (couldn't be followed-up) at one of the time points and that affected the mean, or it could all just be a fluke due to random noise and sampling error. You just don't know.To repeat, you can't look at fluctuations without individual data. You really can't. And, for that matter, you also can't, with any confidence, carry out stastical analyses of the data without the individual data points.This is simple stuff. Not rocket science.

  87. Its very simple – from Baseline to 9 months (both CBT and TAU improve and dont differ from each other) From 9 to 18 months (CBT doesnt change but unmedicated TAU change downwards)'Fluctuations of individual patients' is irrelevant – if the upward trajectory of CBT or TAU reflects one individual or the downward trajectory of TAU reflects one individual, then: a) Morrison et al would need to re-analyse their data using quite different statistics (with different assumptions) and it is also not consistent with any changes in the standard deviations across time

  88. Ha ha, Richard Bentall is back to attacking critics' "misunderstanding" without specifying exactly what they are misunderstanding. I do not know, maybe British professors can get away from that kind of authoritarianism. Would not fly in the US. Regardless, I welcome him to go to my blog posts and indicate what I am misunderstanding.http://blogs.plos.org/mindthebrain/2014/02/25/much-ado-little-lancet-study-cognitive-therapy-persons-unmedicated-schizophrenia/http://blogs.plos.org/mindthebrain/2014/03/11/much-ado-modest-misrepresented-trial-cbt-schizophrenia-part-2/I based what I put in the blogs on a careful reading of the trial, its registration which violated Lancet requirements by being filed after recruitment had started, as well as Anthony Morrison's interview with Lancet.

  89. The point is that 'dose' is a metaphor. Do you calculate the 'dose' of therapy in sessions, minutes, words said by the therapist or homework tasks prescribed and completed? You can't directly import the dose concept from drug trials into psychotherapy trials.The number of sessions will vary very considerably from patient to patient for a variety of reasons and, if you look at the relevant papers, you'll see that this happens in the clinical trials as well as 'real life' – typically trials specify an upper limit at the outset.NICE has to make an estimate of the number of sessions likely to be required for economic reasons. When they say 16, they mean, in their estimation, most patients will show a worthwhile benefit in this time. But, as David Kingdon points out, it's sometimes possible to get benefits from much shorter duration. The SoCRATES study provided just five weeks of therapy plus a few 'booster' sessions later, and obtained a significant effect on positive symptoms at 18 months.And I'm sorry for dropping the 's' from your name. Those pesky typos will keep on happening!

  90. Dear Ute and Rory,I agree with Clare – far from being uninteresting, I think that your comments are some of the most interesting and constructive so far. At points the discussion seemed to be veering dangerously close to unconstructive point-scoring. You have opened it up again by asking a different sort of question: not ‘is x better than y or ‘who is right and who is wrong?’ but ‘what are the different things that can be helpful and why?’ I look forward to continuing the conversation.

  91. This is a general comment, but also a response to James Coyne’s point from 13th April at 11.55. I guess it’s for readers of the Lancet paper to judge whether they feel it’s making claims beyond the data. My own reading is that the authors are presenting their findings as suggestive rather than definitive – for example the abstract states that ‘a larger, definitive trial is needed’. Peter and I were careful to use similar, tentative wording in our piece: ‘a recent trial suggests that (psychological approaches) might…be promising as an alternative’. Of course everyone is different, and nothing is going to help everyone. But surely the main point is that whilst not everyone finds CBT or other talking-based approaches helpful, many do, and for some they can be life-changing or even life-saving. Of course many people find ‘antipsychotic’ medication helpful too. However, in view of its frequent, well-documented adverse effects (‘side’ effects) and emerging evidence of the potential for harm if taken long-term (see e.g. http://www.joannamoncrieff.com or http://www.Madinamerica.com) there is a pressing need to explore possible alternatives. When looking at ‘effectiveness’ we need to remember that unlike the field of physical health with its blood tests and so on, there are few ‘independent’ tests in mental health. Most ‘data’ and ‘evidence’ is based on what service users tell researchers. And as Clare Fraser says, ‘what service users have overwhelmingly asked for, over many decades, is more chance to tell their stories, to be heard and listened to within trusting and respectful relationships. If CBT is one way of delivering this, and is found helpful by some….let’s offer it to them, and let them tell us how to make it even more effective’. Of course not all conversations are helpful, but we need to learn from service users over time how to make our therapy conversations as helpful as possible.

  92. So now you are indiacting that duration (dose) is impoirtant…contrary to your dismissal of it above and my raising of this issue on Twitter (with your customary politeness)As I said, it woudl be good to hear from Rory Byrne (and others) who have actually raised this issue – if you think its irrelevant, I would like to see a discussion between you two…that would be informativeI see no reason why you would want to shut down a discussion of this by saying I "lack understanding" or by calling it a "daft question" on TwitterIf economics are key, then David Kingdon's approach is important (though he currently has no evidence from his Cochrane review) – nonetheless, your approach seems confused and contradictory

  93. While I'm sympathetic to the desire to allow patients greater choice over interventions intended to improve their quality of life, there does seem to be a bias towards interventions which involve paying health professionals for things like 'Art Therapy', rather than just providing patients with greater resources to pursue their own interests. If psychosocial interventions do not provide clear health benefits over and above that gained from the patient's own use of greater resources, then I'm not sure why they should be prioritised over simply increasing welfare payments to those who are sick and disabled.Recently we have seen how the DWP's adoption of a biopsychosocial model of illness has been used to make dramatic cuts to welfare payments and to justify unwarranted assertions of authority over the lives of claimants: http://disability-studies.leeds.ac.uk/files/library/A-Tale-of-two-Models-Leeds1.pdf The DWP has already been caught forcing those on unemployment benefit through a positive psychology intervention: http://skwalker1964.wordpress.com/2013/06/19/dwps-top-psychologist-to-be-investigated-over-fake-test/Recent changes could also allow claimants to be penalised for having not made use of treatments which their assessor believes would increase their ability to work: http://diaryofabenefitscrounger.blogspot.co.uk/2014/01/dwp-limbo.html?spref=tw In the current social setting researchers claiming to be able to help patients may find that those claims are being used to disempower and manipulate those that they had wanted to help. It is important to be cautious and rigorous in the claims made about the efficacy for all treatments, not just so that patients can make informed decisions about their own lives, but also so that others cannot use unjustified claims against them.

  94. Annenobody ever presents their findings as definitive – I would not make that claim for our meta analysis (Jauhar et al). However one views CBT for psychosis, its clear that the Lancet study has numerous probloems that make it almost impossible to interpret as supportive of CBT for psychosis in unmedicated individuals.It would be a mistake for anyone to conflate our criticisms of CBT for psychosis as somehow an endorsement for medication – it is not and has never been. I am a psychologist and would of course like psychology to provide a successful adjunct to medication …but CBTp is a) not an alternative (as some have claimed) and b) not that successful in whichever meta analyses you look at….and given the lack of NICE updating their guidance from 2009 in the 2014 document – they prefer to currently stick with RCTs that stop at 2008 an extremely poor meta anlysis – or I should say 100+ meta analyses that they conduct in their cherry picking exercise – while this happens, alternatives find it hard to occupy any space alongside the weakly supported, but dominant culture of CBTp…In terms of where to go, it is useful to ask those who run the RCTs why they have taken a quasi-neuroleptic approach in over 50 RCTs over 20 years and why they focus on symptom reduction – that is their outcome measure of choice…not oursIn many respects, if you acknowledge what we are saying – it is a clear argument in line with what you are saying i.e. trying something different – moving away from the focus on symptom reduction – which is simply not well supported in well conducted trials!

  95. I agree, and I would point the 12.59 commenter to other articles on this site for a discussion of the problems with the 'illness like any other' approach in mental health, for example : 'Seduced by Biology', 'Is Life a Disease?' and ' I'm Sicker than I thought I was'.

  96. If changes in means could be driven by outliers or random noise, then that just highlights the fact that the sample size here is far too small to be scientifically valid. The world is full of bullshit studies like this — with small samples, anything can happen, and then only the sexy results get published, creating misinformation that arises only from sampling error.

  97. Richard, In the example I gave to clarify (although it seems to have done the opposite) A and B referred to individual patients not groups, apologies if that was not clear.In your response you stated:“If the patients in group A have higher mean scores at time 1 than at time 2, and if the mean scores then 'fluctuate' upwards by time 3, they would have to all (or at least most of them) go up and down together. That's obviously nonsense”In the Morrison trial the TAU group had mean(SD) PANSS of 73.88 (13.42) at baseline, 63.26 (13.21) at 9months and then at 18 months they go back to 71.24 (20.35). So this is exactly what happened in the trial – I am not sure if I have misunderstood you or you are implying that the trial is nonsense.Let us simplify and say there are only 2 individuals in the TAU group – one possibility is that both initially had scores of 74, then 63, then 71. Alternatively 1 patient could have stayed constant at 74, and the other started at 74 gone down to 52 then to 68. There are 100s of possible combinations that would give those means even with just a 2 person group, with the 25 included in the trial the possibilities are endless. However with those means and SDs given, one option is NOT possible and that is for the PANSS to stay stable for the majority of the group. There will have had to be significant variation in significant proportion of the group for the mean to fluctuate that much with those SDs. You do not need any rhythmic oscillation. Regarding my clinical experience (and I imagine yours)- yes patients show marked variation in their symptoms so I did not find those results particularly remarkable.

  98. Thank you Anne, this is a very interesting debate, which is made all the richer for Rory and Ute's contributions, Kieran.

  99. Thanks both for really encouraging comments! I'm glad to have contributed to what's become a fruitful discussion. I agree that Anonymous' comment beginning – 'What is important for the person who has psychosis?' is very constructive, which you echo Anne ('what can help and why'). Research into recovery and valued outcomes is one way I look at this, and to recommend colleagues' work, a new Delphi study presents a very rich picture of individuals' recovery priorities – http://schizophreniabulletin.oxfordjournals.org/content/early/2014/04/11/schbul.sbu047.abstract. Qualitative studies into CBTp may also be useful adding here. Dr McGowan's study is very helpful (at top of comments), and I'd also recommend two qualitative syntheses (McGowan et al in both) – http://www.ncbi.nlm.nih.gov/pubmed/21457606 and http://www.ncbi.nlm.nih.gov/pubmedhealth/PMH0062066/ (colleagues again, sorry). For anyone who can't access the journals I highlight this section of the first paper (image) – http://i.imgur.com/edApov5.png. Shamefully self-promoting, I also suggest these two studies as potentially useful – http://www.ncbi.nlm.nih.gov/pubmed/22874770 and http://onlinelibrary.wiley.com/doi/10.1111/papt.12013/abstract

  100. Prof Laws, I think Richard has concisely explained the exchange about CBTp model/s, and the question of ‘dose’/duration etc, in that he’s drawn a distinction between these, where you’re conflating them. I looked at your twitter page and you first said: “they [CBTp proponents] have no model only ad hoc hypotheses”. Richard explained that there are multiple well-established psychological models underpinning CBTp (referring to substantial IoP work, Freeman, et al). The question of ‘no model’ was resolved. Then you asked: “So, whats the appropriate 'dose' level in the model for CBT for psychosis?” This is a separate question, and as Richard says above, the models referred to aren’t expected to predict ‘dose’/duration required for efficacy. For my own use of the word ‘dose’ in my response to Jauhar et al, I used it the same way I may use the word ‘symptoms’, when ideally I’d describe the actual phenomena referred to (eg., hearing voices). For practical reasons we often have to use widely-shared language, which all too often is still biomedical. In the study I referenced (Dunn et al, 2012), duration differentiated those who attended the fewest sessions (<5) from the wider participant group; the significant findings came from comparison of ‘full’ and ‘partial’ therapy: “Consistent delivery of full therapy, including specific cognitive and behavioural techniques, was associated with clinically and statistically significant increases in months in remission, and decreases in psychotic and affective symptoms. Delivery of partial therapy involving engagement and assessment was not effective.” I think duration is still a useful dimension to explore, as in Prof Kingdon’s approach, as it could be very useful to understand if cost factors, for example, determine or restrict access to therapy.

  101. RoryThe model question was not resolved and wasnt changed – I raised it *specifically* because I want to discuss the avenues of possibility that CBT advocates have open to them rather than pouring our fruitless ad hoc hypotheses …and that is precisely why my next tweet referred to itand I took your, Paul Hutton & Anthony Morrison's suggestion of 'dose'/duration as one such possibility – my question is the same – is there a model to provide a basis for that assumption or why CBT is delievered as it currently is delivered and advoacted by NICE (as I stated above)Yes, the Dunn study is another example where dose/duration (whatever one calls it or measures it) may be important – unfortunaely not much else exists as Im unaware of any RCT that has compared different doses/durations of CBT. Is there a model that incorporates shorter duration? It would indeed be a money saver and so, its useful to know if a model exists for how it operates – is it longer or shorter? Does it interact with symptom type? other factors ?

  102. "is there a model to provide a basis for that assumption or why CBT is delievered as it currently is delivered and advoacted by NICE (as I stated above)" – Which assumption? That 'dose'/duration is important? I take it from Prof Kingdon's findings and the fact that Dunn et al's was a novel finding that this is an evolving area, and as I think others have said, NICE recommendations reflect considerations of cost-effectiveness.

  103. You misunderstand; I said you conflated the general question of 'no model [in CBTp]' with the specific question of 'model to predict 'dose' effect'.

  104. I dont think its "quite a statement" at all – I have given one major reason on my blog and there are many others reiterated by others (to do with 50%+ drop out, use of medication, CBT being offered to controls, and so on)These make it very hard to know what it says (if not impossible)I really dont think such endorsements mean much for 'any' study – it is, of course, an appeal to authority and most make little or indeed, no reference to the substance of the study at all. I find it a curious novelty (to me)

  105. No you conflated it – that was not my intention as I just said. My interest is in the pragmatics for delivery here vis a vis models (as I have immediately clarified to Richard Bentall on Twitter at that time)

  106. I think all parties have agreed – either here or on Twitter – that the argument about statistics and methodology has become unproductive. Let's agree to lay that one to rest. In order to proceed in a more constructive direction and put some of the bad feeling behind us, can we follow the new thread of debate helpfully opened up by Ute and Rory, and highlighted by Anne? I for one would be very interested to hear people's views – especially, though not only, those who come from a service user perspective.

  107. I am amazed by how this seems difficult to understand. Nothing here contradictions my observation about the numerous different ways in which these mean observations might have been generated from individual data. But let me try one more way of pointing out the improbability of Prof Law's claim to have seen fluctuations 'for the first time'.Fluctuations implies that people go both up and down. But if they're all recruited at different times, and at different points in their careers (as the TAU patients were), we would expect as many to be going up at any particular moment as would be going down. How could it possibly be, then, that the difference between the time points on TAU reflects genuine fluctuations? For this to happen, a randomly selected group of patients, entering the study no doubt at different times, would all (or mostly) have to decide to get better over the following nine months and then they'd all (or mostly) have to decide to get worse over the following nine months. What could possibly be driving these kinds of 'fluctuations'. The phase of the Moon?To repeat, to investigate fluctutions you have to look at individual data. To elaborate, you also have to have a lot more data points and use sophisticated multilevel models (because observations are clustered within individuals, who are clustered within groups). You can't use simple ANOVA or do it in SPSS (we use Stata).We have been investigating fluctuations for years and we use experience sampling to do this, with between 20 and 60 observations per individual at pseudorandom intervals, usually over a period of six days. (This technique was developed by Philippe Delespaul and Inez Myin-Germeys in Maastricht and there is a vast literature on it which they people blogging here seem to be completely unaware of. The technique has also spawned quite a lot of technological innovation involving PDSs and phone-apps.)Our first paper using ESM, which was published in the Journal of Abnormal Psychology in 2008, showed that extreme fluctuations in self-esteem are associated with paranoid delusions and can be found at http://psycnet.apa.org/journals/abn/117/1/143/. (The full reference is: Thewissen, V., Bentall, R. P., Lecomte, T., van Os, J., & Myin-Germeys, I. (2008). Fluctuations in self-esteem and paranoia in the context of everyday life. Journal of Abnormal Psychology, 117, 143-153.)We have since accumulated quite a lot of ESM data from psychotic patients, patients with a diagnosis of bipolar disorder, and children at genetic risk of bipolar disorder, and quite a few analyses have been published (you'll find them in Google Scholar if you look under my name or Inez Myin-Gemeys). We are now looking at measures of non-dynamic (chaotic) fluctuation, particularly approximate entropy, and we hope to publish something on that soon.I'm sorry but I'll repeat what I've already said many times. You can't look at fluctuations using mean data. You really can't! You need to use lots of data points and use sophisticated stats, which you can't just run in SPSS. It is not true, as someone above said, that there's not much research on fluctuations. I've certainly done a fair bit and if you look at my ESM papers you'll see how it should be done.Now please get over being wrong. It happens to us all once in a while.And with that I am going on a well-earned camping trip, out of contact of the internet or mad Twitter-feeds!

  108. Thanks Clare. So just to remind people, some of those questions were: What is important to the person with psychosis? What are the different things that can be helpful and why? How can we make our therapy conversations as helpful as possible?

  109. I am a service user and I am chary of therapy, especially CBT, partly because I had a bad experience with a psychologist, but mostly because I find the very idea of it inherently problematic. Recently I saw a support worker for 3 months. She helped me with practical stuff, we visited galleries and we had many, many cups of coffee. We did talk about my experiences but we also talked about her life/art/books/gardening/feminism/politics etc. I liked her and she genuinely liked me and I knew that in the way you do know these things regardless of what is actually said or not. This really helped me, but it is not, of course, something that can be prescribed.I think Harry's post above, which no one has yet picked up on, contains some very important points.

  110. Hi Richard,Thanks for your response I think it finally clarifies why we were disagreeing!Like many of the debates in psych a lot rests on semantics. You say: “Fluctuations implies that people go both up and down.” I had been using a less narrow definition and taking fluctuation to simply mean “not remaining stable”.I agree that from the group means you cannot tell that individuals had both gone up and down. What you can tell is that they have not remained stable – i.e. some individuals might have been going up the whole time when others would have been going down the whole time. To be clearer I guess what Keith could have said is “the study shows that the PANSS scores in the TAU group do not remain stable”. I’m also amazed it took that long for us to understand each other – part of the problem with communicating over the internet. I hope this clarifies the confusion.Enjoy the camping!

  111. Up-dating my reading of the posts, I was glad of the broadening of the discussion. I appreciate the questions about the models of CBTp and how the supposed cognitive change processes are expected to lead to changes in 'self-conception' and 'agency' (behaviour) – specifically for people with unusual perceptions and experiences (psychosis). Rory Byrne, Richard Bentall, Keith Laws have highlighted the importance of valid conception of these supposedly underpinning processes with regards to which 'dose/duration' efficiency may be expected. Questions relevant to practice as well as to the formulation of research questions, manualization of CBT, relative flexibility of adaptation to the participants, benefits and rejections, monitoring during trials, etc. In my view the experiences in trials can highlight many of the experiences in practice, so I do not agree to separating both, as they also concern research methodology.Richard Bentall's questioning of the expertise and knowledge of the discussants with regards to CBTp and other psychotherapies in practice and research has prompted me to re-think about the integrity of my thoughts as I am not implied in CBTp in practice and research.From my knowledge I have contributed what I estime addresses some of the issues in the design of trials on CBTp. My interests concern how people with severely changed experiences can work together with mh professionals. I mean collaboration in which diversity is used to improve interaction, learning, support on the basis of reciprocity welcoming difference. Important features are respect for socio-cultural, knowledge/belief differences, genuine curiosity and dignity towards each other. I am talking group contexts here where there is diversity and fluid travelling in roles, emotions, perspectives, suggestions – the safe space in peer support groups. Best groups I have been part in involve/d peers and professionals with the professionals and the peers using their personal life experiences and their social skills.I am a MSc mental health research methods student, so I can't share expertise from having worked in research teams conducting RCTs or observational studies. Nevertheless my learning of methodology and design issues for conducting research is constantly informed by introducing user/survivor views and knowledge which is quite something to bear with (the role of a user researcher including many challenges as frameworks for user research are under discussion).It is motivating to me that some academic researchers are interested in how to co-produce with user researchers. In my view this is a reciprocal endeavour and it is my standpoint that there is a lot of important expertise I learn from academic researchers without having to annihilate our differences in conceptions and perspectives about issues concerning the understanding of severe mental and emotional distress. I appreciate Anne Cook's suggestion to ask 'What is important to the person with psychosis? What are the different things that can be helpful and why? How can we make our therapy conversations as helpful as possible?' Here is where I find it helpful to work in groups cause newcomers can come sit in and listen and chose what they get a grasp on, find they're not alone and some real basic stuff of people being there and bringing parts of real lives with them. Coming back to CBTp the qualitative interviews and observations from studies may provide what people found helpful or what they felt unseen or offended by etc.tbc

  112. continued:For today I think I carry with meProfessor Keith R Laws16 April 2014 17:49In terms of where to go, it is useful to ask those who run the RCTs why they have taken a quasi-neuroleptic approach in over 50 RCTs over 20 years and why they focus on symptom reduction – that is their outcome measure of choice…not oursIn many respects, if you acknowledge what we are saying – it is a clear argument in line with what you are saying i.e. trying something different – moving away from the focus on symptom reduction – which is simply not well supported in well conducted trials!and Richard Bentall16 April 2014 14:50As pointed out by some other posters, an important issue in all this is choice. In the not too distant past, and still in some services, patients are offered medication, with its known hazards, and absolutely nothing else. Very often, services still do not consult patients about what kind of help they want, or even inquire into their life stories (as Eleanor Longden has noted, patients are asked what is wrong with them rather than what has happened to them). I want services to be more human and patient-centred.Using the 'St Peter's quip' may work considerably well with some users, it's not only Elyn Saks or Jacqui Dillon hinting at the capacity for irony and humor born out of having to handle immense distorsions of thoughts and interactions; so being in the world can be an exercise of black humor for many, and again, a way to meet on the famous continuum, respecting that we have all seen things in our lives that others ain't, so that intersections of viewpoints can be illuminating or just anti-desperately ironic with regards to research or the human condition.

  113. I agree, and agree with Angela's comment (10 April 21:03): "RCT's, while useful in telling us whether an intervention has had a significant effect, whether a therapy is better than another, or better than treatment as usual; can tell us nothing about the experience of psychotherapy and what service-users might find useful about it. What might be more useful is to begin the process with qualitative research which examines service-user experiences of CBTp." I also value your and Kieran's discussion of the 'open dialogue' approach, all I've seen is very promising, and I say that as someone who also supports RCTs. We could also do with more work on peer groups as Ute mentions below. I think all well-researched approaches can tell us something important (though facts are that RCTs run in the context of NHS provision have to prioritise 'symptom'-reduction, for now…). You also said: "What service users have overwhelmingly asked for, over many decades, is more chance to tell their stories, to be heard and listened to within trusting and respectful relationships.", and Anonymous (13 April, 12:22) concurs: "Some of us just want protected talk time and who cares whether that's structured or unstructured talk, just to have the option, and the option to not take meds." This is just one brief example and could be dismissed as 'anecdote', but it is something that's clearly and consistently echoed in the literature (and isn’t a comment in support of CBTp in particular). It seems to me as important to acknowledge in this discussion as any debate of stats or method, which have already had extensive air-time.

  114. This is a really important perspective I think. I conduct interviews to explore how people see therapy, and it’s really helpful to hear your experience. I hope if you want or get access to psychology again it’ll be more positive. Could I ask if you’d say anything about the CBT approach being problematic, that could be really useful (no need if you’d prefer not).It looks like the support worker you’ve been with did a great job, would love to see that approach available more.I think any treatment-related problems with benefits are serious, though I don't feel I know enough detailed info to add too much. I agree with Harry that people should have access to more general support, and not just formalised interventions if that’s what would help more. Though I would say I think there’s a valid place for both concerns overall (developing better social/general support, and better specific interventions).

  115. Hi all, Fascinating debate. I think that an issue that has not been discussed enough however is that CBTp has largely been evaluated in its effects on 'Schizophrenia' and psychosis as a syndrome. This was always confusing to me as a clinician as our primary aims as a CBT clinician were to reduce distress and disability associated with individual psychotic symptoms rather than 'schizophrenia' per se. It is therefore curious, and I would argue, invalid to evaluate CBTp using measures such as P-PANSS which were popularised during drug trials. Have we missed the point here? I know this is not a new point this was made by Birchwood and Trower (2006) but little has changed in the way that RCTs are designed in terms of outcome selection, and this point does not find its way enough into debates such as these. For those interested in a more thourough examination of these issues, and some preliminary evidence supporting my position I have recently had a relevant empirical paper accepted at Journal of Cognitive Psychotherapy which speaks to this. A pre-print PDF can be accessed here: https://www.dropbox.com/s/ren5pgqsx3pip2b/Brockman%20et%20al_preprint.pdfBestRob

  116. Thank you Rob, I think this is a vital point. We know that many, many people experience what might be seen as 'psychotic symptoms' -for example around one in ten have heard voices speaking when there is no-one there. Only those who are distressed by their experiences are likely to approach services and ask for help. As clinicians and researchers our focus needs to be on reducing distress, and only the person him or herself can say how helpful our efforts have been in this regard.

  117. Thanks for the link to your preprint paper, Robert Bockmann. It supports my criticism of overall/positive/negative symptomology outcomes as dismissing the reality of struggling with voice hearing and unusual beliefs.Alternative conceptualisations of psychological and emotional processes in experiencing and relating to specific extreme phenomena/symptoms have been discussed for more than 15 years (Chadwick & Birchwood, 1996; Romme & Escher, 2000; Birchwood & Trower, 2004). However, there seems to be little empirical evidence on adaptation of CPTp models nor outcomes to specific symptom domains.I find this rather disappointing and ask myself if it is really the issues of, potential benefits for clients with distress-related unusual beliefs or voices which guide research priorities.I am not writing from an academic standpoint only, in referring to academic and grey literature on psychological processes and social factors in unusual beliefs/paranoia experience, but I use and reflect upon the domains in the exchange processes in a paranoia peer support group I co-facilitate weekly. Given this background I am both sad, and also angry, about the narrow models I see discussed and which possibly underpin CBTp practice. I hope to be totally wrong in the latter and welcome practising CBTp practicioners and researchers to quote papers on a broader range of psychological and social procedures used in working with clients.I can't agree more with Anne Cooke emphasizing that 'As clinicians and researchers our focus needs to be on reducing distress, and only the person him or herself can say how helpful our efforts have been in this regard.' User-defined recovery is still missing (Gordon & Ellis (2013) Recovery of evidence-based practice. http://www.ncbi.nlm.nih.gov/pubmed/22830603) Reading even most recent academic papers on CBTp referred to makes me doubt that crucial and pivotal progress has been made to co-produce practice, generalized models and research together with clients/users. As long as the voices and experiences of people with psychotic experiences can be disqualified or re-interpreted by 'epistemic violence' of eminent researchers and clinicians alone, change can be suppressed. Is it really necessary to await more scandalous reports from the Schizophrenia Commission? Ii seems that a serious, shared need for renewal in perspectives and strategies, shared by clinician-researchers and clients/users, is needed for constructive change to happen. That is what I find – in accordance with Kuhn's theory of paradigmatic crisis and change – the most important lesson from the birth of the Hearing Voices Movement: Marius Romme and Patsy Hague together explored possibilities to leave the impasse of depressing, harassing voice-hearing; together with Sandra Escher they explored and investigated human strategies not yet used within the narrow framework of the discipline, using cross-sectional surveys as well as interviews which were detailed through iterative processes of nuancing relevant themes and subthemes from the experiences of voice hearers. Reading the latest posts I get the impression Clare Fraser and Anne Cooke are truly interested in working together with people with lived experience of distressing beliefs, voices or intrusive thoughts and eventually also with some who have worked out strategies to integrate them better in their lives and ascribe changed meanings and relations to their voives/beliefs. Several posts refer to the needs, priorities and voices of users/clients. This is supportive and motivating. However, one-sided declarations are not enough. I suggest a literature review (2013) Co-production in mental health: why everybody wins http://www.neweconomics.org/blog/entry/co-production-in-mental-health-why-everybody-wins

  118. I did not mention Richard Bentall's work, chapters on Paranoia (Doctoring the Mind, 2009) which the professionals but also research interested users appreciate. Here is The Psychology of Paranoid Delusions -in a pdf nutshell. Bentall's presentation at INTAR 2011 conference with research overview http://www.intar.org/files/INTAR2011-Toronto-RichardBentallPsychologyofParanoidDelusions.pdfFrom a user-researcher and peer support pactitioners perspective:Tamasin Knight, University of LiverpoolYou’d better believe it: Accepting and working within the client’s own realityhttp://www.dbdouble.freeuk.com/Clinpsych9.htm The comprehensive work of Tamasin Knight is no longer freely availableBeyond Belief – Alternative Ways of Working with Delusions, Obsessions and Unusual Experienceshttp://www.peter-lehmann-publishing.com/books/knight.htm

  119. I think there some really good points being made here and I’d like to respond to some of them. Just for the record, I’m a clinical psychologist who has personal experience of psychosis (diagnosis of schizophrenia and bipolar – patient choice in action!) and using mental health services. I also have clinical experience working in psychosis, specifically trained in CBTp, and research experience in clinical trials (although in pain not psychosis).Firstly, I think that some of the disagreements stem from the nature of psychotherapy itself. As Stephen Morley has pointed out, psychological therapies are “manufactured at the point of delivery”. Good psychological clinicians are often those who can respond flexibly and creatively to the client’s idiosyncratic need, both in session and over time. In many ways, psychotherapy is more akin to art than science. Thus it is simply not possible to do justice of the complexity of delivery without a flexible research design. Most research funding agencies however treat psy therapies like medication and require a degree of manulisation in the delivery that is not really reflected in real clinical practice. I don't think it’s necessarily fair to criticise the sceptical scientists here for their scepticism, although I do think some of the points betray unfamiliarity with the idiosyncrasies of psy therapy. Perhaps this an area where psychologists need to get better at explaining what psychotherapy involves and how it really does differ from medication? And, from this place, we can discuss together more appropriate ways of researching its effectiveness that does employs scientific rigour but does justice to the somewhat unique context of a relational approach?Secondly, I’ve become increasingly irate at arguments that dismiss any psychological intervention as “unproven therapies” and this risk-aversive rhetoric about “protecting the patient” and “harmful effects”. Absence of evidence is not evidence of absence. Just to illustrate, I was helped in a therapeutic community, about as “unproven” as it’s likely to get. However, following my time at the community after a psychotic break I was stabilised without medication, went back to education, got married and trained as a psychologist. I might be an N of 1 here, but when you are that N of 1 such things are really quite important. Psychiatric detention and ECT left me depressed and traumatised. A psychiatrist was willing to “bet 1 million pounds” that I wouldn't make it without medication (!). Give me the unproven therapy any time. I have been off medication and out of services for over 15 years. Probably time to cash that bet.And this brings me to my last point. We ignore service user experience of what works at our peril. Jon Dewberry’s point comparing the“anecdotal success” of CBTp to homeopathy, presumably to illustrate the dangers of relying solely on patient self-report, are frankly a bit offensive (sorry Jon if I’ve misunderstood your point here). Is Open Dialogue “unproven”? What about Hearing Voices groups? Or Therapeutic Communities such as Soteria? Service users prefer these approaches for many reasons: they are humane, supportive, relational and most of all, they HELP. Psychiatry in the main does not, and many service user accounts illustrate this (see the Schizophrenia Commission’s recent report as an example). There really is no way of quantifying the impact of something that literally saves your life. After all, a drowning person does not have the luxury of estimating the effect size of the length of rope thrown to them.

  120. Hi All, Just noticed a newly published meta-analyses which looked took a single symptom approach titled: 'The effects of individually tailored formulation-based cognitive behavioural therapy in auditory hallucinations and delusions: A meta-analysis' It can be found here:http://www.schres-journal.com/article/S0920-9964(14)00134-0/abstractI think this is a step in the right direction- moving away from syndrome outcomes to single symptom outcomes. I think however that this could be taken further to outcomes more relevant to the goals of CBTp – the reduction of distress and disability.BestRob

  121. I am a victim (I don't say survivor because the better part of me was murdered) 'successfully treated' on anti-psychotics who would give anything to go back to when he was 'more ill'. Why is there this obsession with reducing symptoms? Symptoms are irrelevant in the case of psychiatry, which doesn't identify real distress but only social and epistemological conflicts. Anybody who appeals primarily to symptom reduction demonstrates himself thereby callous and incompetent. Quality of life is all that matters, something psychiatrists couldn't give a hoot about. It should be obvious that the reasons psychiatrists have such a mortal hatred of talking therapies are not based on reason but (apart from the drug money) more to do with the fact that they view it as BENEATH THEM. They consider themselves higher than us in the food chain, you don't try to understand your dinner's innermost heart before you slaughter it. But alas, they are incapable of understanding anything anyway. Whenever they open their lying mouths they give away their total lack of insight into human nature and inability for any form of empathy. If they were able to look even into their own hearts, all their teaching in mischief would evaporate before the ever-glorious and resplendent Light of Love. If they were told to cure by their wise counsels or inspiring personality, their spiritually-malnourished, infantile psyches would be like babes before wolves. Max.

  122. As for the actual efficacy of talking cures…It should be perfectly obvious that counselling is absolutely necessary for anyone involved in psychiatry. Not to treat any so-called 'disease' they invent and spruce up with slick-sounding names, but to deal with the system itself which contradicts some of the most basic 'salutary illusions' told to us about the Freedom and Justice of our society and is apt to terrorize the bold and pervert the righteous. Human beings crave moral recognition and understanding, especially when they're being treated like something the dog produced.Max

  123. Delusions are not something that only 'psychotics' have; strictly speaking if one takes psychiatry seriously, everyone is psychotic because delusions are the norm for humanity. Only the Awakened can begin to penetrate the veil of Maya. In order to achieve that and attain bliss, meditation and other spiritual practices must be taught, such as a willingness to transgress societal norms – something that seems to come easier for those labelled 'psychotic'. This doesn't require research, it is testified to by the experience of sages throughout history. Why not start with teaching those labelled 'psychotic', if that's where the government is willing to spend its precious resources? The question is not whether they are labelled 'psychotic', it is whether they are willing to apply themselves. 'Psychotic' is only a term relative to the delusions of psychiatrists, not absolute, so it is unreasonable to assume that they should be any less equipped than anyone else for Enlightenment. Though it would have to be done before medications destroy their brains. Max.

  124. Dear Simon Clarke,Thank you very much for your very HELPful and instructive thoughts on the scandal that users/survivors voices and narratives are still being disqualified a/ by the hierarchy of what counts as good evidence, even though evaluation of complex interventions is expected to use mixed methods b/ by an as important disregard of a majority of conventional researchers to meaningfully integrate users/survivors in all phases of research projects development and conduct in non-tokenistic ways.I do wonder if basing a research projects rationale and objectives in pre-existing theories produced exclusively by conventional academic researchers is the best way to protect and maintain 'evidence making' in the hands and heads of conventional researchers at the dispense of helpful and meaningful 'paradigm change' and practice change. Weter we hold our breath or not, the unmet needs stay with the users/survivors.'Just to illustrate, I was helped in a therapeutic community, about as “unproven” as it’s likely to get. However, following my time at the community after a psychotic break I was stabilised without medication, went back to education, got married and trained as a psychologist.' Yours, Simon's, report relates to an increasing number of other survivors/users personal recovery stories which very often are based in being accepted as a respected adult and learn in adult ways from each other in peer support groups or peer run/peer & professional co-produced recovery-oriented programs. Now again, the question of why these kind of programs have been much to little researched and evaluated (some have been rigorously explored and evaluated over the last decade) is easily answered: they do not fit in with the established psychological and services research experts episteme with its focus on individualist psychological and mental processes and the position of knowledge and practice authority held by mh professionals. It needs be mentioned that neo-liberalist research policy has been very successful in exterminating research based in more socio-cultural paradigms, Open Dialogue much refers to socio-linguistic and socio-cultural-developmental research done by Bakhtin and Vygostky; only few weeks ago KIM HOPPER’S contract has been terminated, who has done lifelong work in elaborating innovative research methods into/with groups living in poverty and people with experiences of extreme distress and disturbances.Nevertheless I keep being hopeful because there is growing political support for the roles of peers (users/survivors) in the co-production and delivery of recovery oriented support in mh services and policy promoting the 'involvement' of users in research in the UK, in Scotland, the USA and in Canada, and some other countries.

  125. Part of my ongoing learning curve and dialogues is to explore the opportunities for mh professionals and researchers which arise from co-production with 'experts by experience'. Even though – with the discussion of DSM 5 – or the last report of the Schizophrenia Commission – a growing number of mh professionals and researchers call for a necessary change of paradigm – the experiential frameworks and insights of the people with severe distress/disturbance experiences and their explanations are not 'really' asked for or permitted to give crucial input. The only exception to this is part of the scientific and social history-in-the-making of the 'Maastricht approach' developed in collaboration of researcher-clinicians and voice-hearers. Sharing and shifting their questions and explorations through epidemiological surveys and hundreds of interviews re-build a new, psychosocial, biographical, interactional, context related understanding of the phenomenon of voice hearing. The expertise and experiences of all stakeholders involved were necessary for this ongoing transformation and elaboration of frameworks of voice-hearing to 'get real'. Similar progress is being made in the co-production of mh services delivery, commissioning, evaluation and improvement in England. However when it comes to a more valid understanding of the 'psychological mechanisms which underpin psychotic experiences' I wonder why there seems to be little endeavour to decisively engage with users/survivors knowledge. Possibly most of us do not see our experiences as being ruled by a handful of psychological mechanisms?I read some of the literature authors posting here referred to and the reductionism in the psychological mechanisms used for theorizing interventions begs many questions. This obviously is not a tension-less field of inquiry. I am saddened and frustrated that even user-friendly researchers seem (am I wrong?) to undertake little in change of questions, perspectives and frameworks. As there have been issues of cherry picking, there seems a lot of authorized cherry picking of 'psychological mechanisms' more useful for underpinning a specific CBTp approach than the next. And I know of the high importance of trustful relationships of people with SMI experiences with their psychiatrists or psychologists, given they have been able to find such a person. I think that a long-term trusting relationship with a humane, respectful, curious and supportive psychiatrist or psychologist is decisive and most often necessary for enabling and supporting user defined recovery. So wouldn't it been really helpful to include users in exploring how CBTp can be made more relevant and helpful, using the enabling as well as the critical experiences of users who felt not acknowledged and addressed as adults with their own stories, knowledge, suffering and thriving for values, roles, goals? I have now formulated several times my questions concerning the broader challenges of and opportunities for bringing about the called for changes in a/ formulating theoretical understanding, b/ conducting research and c/ supporting individually meaningful, recovery-promoting practice. I do not want to bore people with redundancy. I belief ‘psychosocial’ dialogical work to be of high relevance to research and practice and continue to scrutinize academic and grey literature for advances in research into complex interactional dialogical practices beyond individualist paradigms.

  126. I’ve really enjoyed the thought-provoking comments of the last few days. Reflecting on the whole debate, I think it is important to name some of the games that had been going on up to that point, both here and on Twitter. I have identified one called ‘Not Scientific Enough For Me’ (cf Berne: ‘Games people play.’)The gameProvoke, misrepresent and jibe at your target that s/he is ‘Not scientific’ and uses anecdote, ad hominem arguments etc.. (The catch: Only I can define what counts as ‘scientific’ – and nothing you do ever does.)When target eventually responds with irritation, pile in with fake outrage and attack them with anecdote, ad hominem arguments etc.. Repeat ad infinitum, inviting others to join in.Related games: ‘Mine Is Longer Than Yours’ (citation lists etc); ‘You’re Fighting Again’ (seeking tension between clin psychs & psychiatrists). Payoffs: Allows initiators’ Child ego states to vent fury, fight, name-call, etcUndermining SUs who might blow the whistle on the much bigger game of ‘Psychiatry’Undermining clinical psychologists who might blow the whistle on the much bigger game of ‘Psychiatry’Deflects from discussion of unscientific and damaging nature of psychiatric labels, drugs, ECT and so on.Antithesis:Name the game: ‘I think you are inviting me to play NSEFM’ and politely but firmly refuse to engage.Having avoided distraction, continue productive co-operation across professional groups in close liaison with SUsChallenge overly narrow and excluding definitions of ‘science’ and ‘evidence’Continue to promote research and practice that is inclusive of SU voices Routinely replace ‘anecdote’ with ‘personal testimony’ and treat it as a valid and extremely important form of evidence.What do others think?

  127. I think it is really the personal testimony of those who receive the 'treatment' that matters. I think the statements of actual 'patients' should be treated with exactly the same dignity as a first-hand witness statement in a court of law.They are the only ones who can accurately bear witness to how it has effected their quality of life, the soul-destroying consequences of the drugs, the only ones who have to live in their skin, something which is fiendishly ignored by the present system. Psychiatrists on the other hand would nearly all personally 'personally testify' that their patients got miraculously better on the poisons they forced them to take, since the real catastrophic consequences are infinitely distant from them. Since they have no direct access to the effects of their treatment – they are not first-hand witnesses – such opinions should be treated with only minor attention.Max.

  128. Thank you Clare. Liberating 'game of ideas'. Would it be 'child egos' states… 'I think discussants know it is about the epistemic daily business to hypothesize, operationalize, experimentally or RCT test, what is 'biopsychological' 'illness' 'mechanisms''reality. No, I don't see no games. I see who is the 'objectified' subject. If diseased, deluded and disturbed, the 'mentally ill' produce (not reliable) knowledge, (not valid) explanations, (not adequte) beliefs. [Even if no longer acutely psychotic], 'people with SMI' have no contribution to make to the serious, powerful, rigorously controlled procedures of scientific knowledge operationalization, measurement, interpretation, discussion, suggestion of need of further research and better pactice. Circle, cercles circle.It seems scientists over many hundreds of years have had minds with access to special states for times, or heard voices, or explored realms of mind-range in extreme moods. Oha, they're all dead and buried. That peronal stories and testimonies are informed by broader social realities, structured narratives and thematic-matrices, thereby gaining in relevancy and power, the objectivist and operationalist scientists have little heard of. Stories of recovery, wit their specific narrative patterns and thematic matrixes from outside the formal treatment system have gained knowledge power, social scientists got interested, rehabilitation researchers, social psychologists, even politicians. Something moves, something grows…So sad that the traditionally devided seem have no time to collaborate to create new frameworks, innovative mixed methods, new multi-stakeholder research using all diverse perspectives and expertise to transform and improve…. research in practice… even paradigms. In the meantime, what is of importance: the responsibility for project, publications and disciplines' expertise keep brilliant researchers fighting – neoliberalist research policies turn former colleagues into competitors, see citation lists and impact factors. No child ego games.An increasing number of survivors/users do no longer want to play the dependent contra-aggressive weak-naughty-kiddo-roles of blaming back. No child egos on our side. But knowledge, resources, education, and experiences and knowledge from both sides. Also, we agree to meet on this famous continuum. Foresight: we are ahead of ourselves, more important power and responsibility work and debates keep senior researchers and excellent clinicians occupied. So much pressure, so much daily excellence and monthly impact factor. Knowledge challenges by newcomers, what! users/peers with valid knowledge frameworks, untimely, check agendas, time flies, future 2 has planned us in already. No games.

  129. Looking at Coyne's publication list, it appears that he has not in fact done any research on CBTp or CBT of any kind, so aim not sure how drawing attention to his citations adds anything to the debate?

  130. Its worth adding the important bit of the abstract:The statistically significant effect-sizes were 0.36 with delusions and 0.44 with hallucinations, which are modest and in line with other recent meta-analyses. Contrasted with active treatment, CBT for delusions lost statistical significance (0.33), but the effect-size for CBT for hallucinations increased (0.49). Blinded studies reduced effect-size in delusions (0.24) and gained some in hallucinations (0.46). There was no heterogeneity in hallucinations and moderate heterogeneity in delusion trials. We conclude that CBT is effective in treating auditory hallucinations. CBT for delusions is also effective, but the results must be interpreted with caution, because of heterogeneity and the non-significant effect-sizes when compared with active treatment.

  131. It adds no more than your comment does – nobody decides who may or may not criticise research on CBT (or anything else), although Richard Bentall seems to pursue such an absurd argument – otherwise you would be committed to pushing service users out of the equation of commenting on research …on service users – absurd comments now evolving here

  132. JIM VAN OS calls for a new scientific paradigm of 'psychosis' as human variation through collaboration of people with experience of severe mental or emotional crises and critical scientists. I hope this ground-breaking article gets translated into English asap.Maybe it is more appealing to read an internationally renowned researcher's perspectives on the need to synthesize and transform evidence from experience and scientific investigation to the academic discussants… than only glancing at the calls for co-production of knowledge from survivor/user investigators, researchers and bloggers. Nevertheless I enjoyed that Jim van Os highlights many of the arguments me and other users/survivors have presented in this blog.Crisis als kans. Zijn ggz en psychiatrie zich van zichzelf aan het bevrijden?Deviant. maart 2014 nr. 80'Psychiater Jim van Os, optimist, ziet knsen in de kentering: meer samenwerking tussen ervaringsdeskundigen, ggz en kritische wetenschappers in onderzoek en een behandeling die een beter leven buiten de ggz belangrijker acht dan het stabiliseren van symptomen.''Het bewijs van psychose als menselijke variatie, en als uit-komst van menselijke tegenspoed moet urgent worden gedissemineerd. Niet als opstandige mening, maar als een neutrale wetenschappelijke bevinding, gevalideerd door ervaring van mensen. De krachten die samenzweren tegen de visie van psychose als menselijke variatie zijn formidabel, en vereisen een duurzame investering in een informatie-interventie binnen en buiten de ggz. Ervaringsdeskundigheid en kritische wetenschap kunnen hiertoe samenwerken buiten de DSM-gerichte gevestigde kaders, instituten en verenigingen, zonder te worden opgeslokt in de myriade van consensusbehoefte, bureaucratie en belangentegenstellingen.'Professor Jim van OsPsychiater verbonden aan de afdelingPsychiatrie en Psychologie, Maastricht UMChttp://www.tijdschriftdeviant.nl/images/stories/deviant80-06.pdf

  133. CBT or drugs…? I expect their are people who have been coerced into having some sort of psychological therapy….. I doubt anyone has ever been sat on by five people and had their clothes ripped off and had CBT sessions inflicted on the in a seclusion room.CBT is just one small part of starting think about how people can be helped in more humane ways….. and to me that must be a good thing. I think this point gets lost…

Leave a Reply

Your email address will not be published. Required fields are marked *