Sunday, February 23, 2020

Comments by KeithLaws

Showing 23 of 23 comments.

  • Just a few quick points:
    a) van Os may (like many others) have argued that bipolar and schizophrenia have ‘similarities’, but he also argues that they have ‘differences’ http://bjp.rcpsych.org/content/194/2/101 and http://www.ncbi.nlm.nih.gov/pubmed/15474912
    b) whether they are the same or not is not relevant here though – as interventions need not be the same nor necessarily similarly efficacious in bipolar and schizophrenia. For example, there is no evidence whatsoever that ‘any’ psychological intervention reduces the mania symptoms or indeed, depression symptoms in bipolar disorder (despite being effective in unipolar depression!). Similarities have no bearing on what interventions will be efficacious – more likely, its the differences that are key
    b) The book chapter you link to is very poor on many levels – first, it is 12 years out-of-date and so totally redundant. Second, it finds only a small impact of psychotherapy in trials that used a control group. Third, even where a tiny effect is found, most of the trials are open i.e. assessed by people who knew the group assignment – this was a problem of early trials – so even the small effect (r-.1) would disappear with blind outcome assessment

  • Not sure why you think I am skirting any issue. As a psychologist, my research interest is in the role of psychological therapies in psychosis and not in medication.
    But to answer your question, yes, I do know the Wunderink study and it seems sensible to me. The evidence suggests that psychological interventions might be potentially helpful at the earliest point (e.g. preventing transition to psychosis), but less apparently successful after that point (at least in terms of the outcomes that CBT researchers use in their RCTs)

  • Thanks Steve. The balance regarding how medication ‘helps’ and ‘harms’ has been quite well explored as has the interaction with duration of treatment. Certainly more investigated than whether psychological alternatives (such as CBT) are harmful or indeed, helpful. Interestingly, some evidence now suggests that longer durations of psychological therapy (esp CBT) may also be counter-productive for people with psychosis http://bjp.rcpsych.org/content/207/3/269.1.long

  • Hi Fiachra, Thanks – its an interesting Powerpoint presentation. If they haven’t written it up for publication , then they should.
    I cant tell too much from the slides, but it looks like it would fall foul of the kinds of methodological rigour required in a controlled trial (i.e. a control group, random assignment, blind assessment of outcomes). The paper I referenced above is an analysis of RCTs. I understand that not everybody is enamoured with RCTs but health bodies (NHS, NICE in UK) are when considering investing in interventions.
    On the other hand, as I said above, peer support is often desirable (for anyone in any distress) and I am surprised that the evidence is so universally negative on outcomes in peer-support trials (for people with schizophrenia and bipolar diagnoses) – it looks like it should be an effective and easy-to-administer source of help

  • The evidence on peer support making little or no difference in both bipolar disorder and schizophrenia is not my research, but comes from a meta analysis of all published studies. Interestingly it was conducted by the chief meta-analyst who is also responsible for the NICE guidance on bipolar disorder alluded to above.

    I am not aware of research showing that medication (in these disorders) has a ‘detrimental’ effect on the outcomes I mentioned (i.e. hospitalisation, employment, or self-reported outcomes such as symptoms of mental health problems, quality of life, recovery, hope, empowerment or satisfaction with services) but would be very interested to look if you can provide references

  • I think it is important to make a clear separation between the psychological precursors and the psychological interventions proposed for bipolar disorder and schizophrenia. Leaving aside psychological precursors, it is worth looking at the psychological interventions that you mention in your letter.
    You say “Given the diversity of these needs, this care might include peer support, information, psychological therapy, or (where the person finds it helpful) medication. So to simply promote medication without also presenting other approaches (such as psychological therapy, as recommended by NICE) is biased and irresponsible”
    If we turn first to the empirical evidence on peer support for bipolar disorder and schizophrenia, it is not positive. A recent meta-analysis found no evidence that peer support has a positive impact on hospitalisation, employment, or self-reported outcomes such as symptoms of mental health problems, quality of life, recovery, hope, empowerment or satisfaction with services http://bmcpsychiatry.biomedcentral.com/articles/10.1186/1471-244X-14-39 Although peer support is of course desirable, empirical evidence suggests it has little or no impact in the areas that we might hope and expect.
    If we turn to your reference of the NICE guide advocating CBT for schizophrenia and psychosis. This NICE guidance is now unacceptably outdated – their review and current recommendations cover studies published only up until 2008 – clearly in this instance, NICE are not meeting their own commitment to keeping guidelines current and evidence-based.
    In the case of the NICE guideline advocating psychological interventions for bipolar disorder, the evidence is very low quality and as with the NICE guidance on schizophrenia, has very recently been questioned https://theconversation.com/are-treatment-guidelines-on-schizophrenia-and-bipolar-disorder-just-plain-wrong-54369

  • Stephen
    You say “My eyes were opened in a very positive way. Things suddenly made sense; psychology, interactions, life, all became much clearer. I now understood that we are all unique, not only because of our experiences and views on life, but because of our neurocircuitry. I began to see how our unique cognitive profiles make us who we are.”

    I would suggest the opposite may well be true of neuropsychology – uniqueness and ‘content’ are not parts of the explanatory model(s). The discipline does not allow us to say anything about individuals – except in the ironic sense of saying we all possess precisely the ‘same’ cognitive and largely the same neural architecture.
    Contrary to your argument, neuropsychology assumes the opposite of uniqueness – it assumes ‘sameness’ i.e. that every person on the planet is ‘the same’ (cognitively… if not neurally) – bar the odd ‘alien’ amongst us (pace Caramazza). Again ironically, this ‘sameness’ is precisely why the individual case study has been so popular in neuropsychology – e.g. we learned lots about short-term memory from HM, but we never learned much/anything about ‘HM’.
    I would be interested to hear how – apart from the trivially true sense – you see neuropsychology reveal uniqueness through neurocircuitry or cognitive profiles? I can see how you might arrive at these views, but not via neuropsychology…

  • Actually the name is Laws -even if we don’t know yours.
    Even if there were 5 -and I will check – that is irrelevant – it’s 5 in 2000 not 2 in 74 – do some maths on it and let me know the likelihood of dying in previous trials versus this unmedicated trial

  • Dear Fiachra
    I wouldnt doubt your experience for a minute – I am pleased that you found it helpful. Many would say similarly of medication.

    I am only asking that people are allowed to assess the evidence for every intervention whether its psychological or drug based. Especially non-scientists ie the vast majority of people needing these drug and talk therapies – we have a duty as scientists to make clear to the public…what are the pros and cons

    The problem is that drugs are painted as bad and psych interventions are painted as good – it is of course not that simple. If you are radically changing someone’s thinking with drugs or talk – we need to be aware of the chances of benefit and of course, the chances of harm.

    We are all aware of the negative consequences of antipstchotics (and maybe the evidence on the potential benefits) – by contrast, we know next to nothing about the harm of CBT (and where evidence exists, it suggests it may be harmful in psychosis – I gave one referenced example above)

    Re the benefits of CBT for psychosis, these are sadly -at best – minimal (according to the evidence from large groups of individuals in trials – at least in terms of reducing symptoms
    themselves)

    Finally, we might also ask why CBT for psychosis researchers do not routinely examine potential harm in their RCTs – we wouldnt allow drug companies such leeway…would we?

  • Factchecker
    I have read the new study published today http://ajp.psychiatryonline.org/article.aspx?articleID=1831621 and it does not contradict anything we say in our latest meta analysis.

    Just saying this does not make it so – if you feel it does, then by all means please be specific in this forum

    The paper you mention has one ‘accidental death’ – it alters nothing – even if its 1 death (accidental) in 50 trials of medicated individuals (2500-3000 individuals) vs 2 deaths in one trial of (74 individuals reduced to <40 at trial end) unmedicated indviduals

    Nobody denies deaths in antipsychotic trials – its an irelevant argument to what I am saying

    Perhaps even tell us who you are if you are going to accuse me of smears

  • Beyondlabelling
    I cant recall a 140 character discussion on Twitter after 12 months, nor do I see if you replied to me when I said I didnt understand your point?

    The reason I ask is because if you read that whole paper to which you refer, you will see that Leucht et al directly examined abruptness of withdrawal in the results and say
    “Abrupt or gradual withdrawal of prestudy
    antipsychotic drugs did not change relapse risk” – so perhaps you didnt read it fully

  • John
    in reply to “I note Keith Laws does not look at the evidence of increased relapse rates when people come off drugs.”

    I gather from your comment that you have not yet read the paper being discussed here. It is freely available on link in my blog http://keithsneuroblog.blogspot.co.uk/

    The individuals tested in this study were either never medicated in their life or drug free for a minimum of 6 months

    So the adverse events are not a reaction to drug withdrawal

  • John
    you say “There is then the problem that CBT therapists may not have much experience of dealing with people who are very distressed”

    According to many influential CBT for psychosis advocates, CBT does not need to be delivered by people with highly specialist training e.g. as here advocated by authors who are members of the NICE committee or have been members http://www.rima.org/web/medline_pdf/BrJPsychiatry_40.pdf
    In this study they were trained for 10 days to deliver CBT for psychosis

  • I know Jonathan, the culture is different. Bear in mind in the UK, CBT is to be “offered to all people with schizophrenia” according to national guidelines (regardless of issues of heterogeneity)

    Re adverse events, again I should note that 10% in that study is undoubtedly an underestimate because the researchers lost contact with so many participants for whom the experience was presumably not tolerable! And one-third chose to become medicated duting the trial. My point is that these adverse events have not happened in trials of 1000s where paticipants were medicated.

    Also note that TAU does not mean TAU in this trial – many of the controls were simply ‘discharged’, with no care – just followed up by the researchers

    Nobody (including me) would argue that medication has no problems, but it has to be weighed against the risks of no medication – and this trial in some resepcts highlights that dilemma

  • Dear Rossa
    people may choose to follow paths that are ‘unevidenced’ – that is their choice – but in the UK NHS, we cannot afford to direct taxpayers money towards interventions that are shown to be of little or no use (at the expense of pursuing possibly more helpful alternatives).

    By the way, please note that I said “at best we have 6% & 3% who ‘may’ benefit” – and put ‘may in quotes – it is also possible that they ‘change’ after CBT not because they benefit but because their condition is worsened by CBT. And that is important to know – CBT for psychosis therapists have avoided documenting possible ‘harm’ of CBT therapy. Very few studies examine this but where they do, harm has been documented e.g. http://www.ncbi.nlm.nih.gov/pubmed/22759932
    We would not stand for this in drug studies – why should we in psychological interventions?

    Please also note that Dr Steingard nor anyone else here (on my quick look) mentions the fact that serious adverse events (SAE) were documented in 10% of participants in this specific trial with unmedicated individuals – including e.g. 2 deaths, one attempted suicide and one serious case of threatening another – this was in 74 people at start of study and only 51 remaining at the end (how ‘tolerable’ is this)
    – compare with 50 trials of CBT in medicated individuals covering nearly 3000 people – not one death ever reported

  • Dr Steingard,
    if we take it that 50 RCTs exmaining CBT for ‘schizophrenia’ have been conducted over 20+ years, we might assume that they have managed to capture a reasonable degree of heteroegenity (acute, chronic etc) – see our meta analysis published Jan 1 2014 http://bjp.rcpsych.org/content/204/1/20

    If you accept that premise….then we have ‘no evidence’ from RCTs that benefits comes from CBT – in fact the trials show that 94% and 97% of the CBT and control groups overlap on positive and negative symptoms respectively at the end of treatment – in other words, we with regard to positive and negative symptoms – at best we have 6% & 3% who ‘may’ benefit – this does not seem – to me – to be sufficiently large to warrant investing public money and hope in such an intervention.
    If someone can show it works for a predefined subgroup, that is great, but it has never been achieved – and these RCT participants are the ones most likley to benefit (having been screened into demanding studies that last months)

    Re Cognitive remediation, I dont know the package you mention and havnt seen their papers, but am familiar with the general literature – and it indicates a benefit in the order of just under half a standard deviation improvement in cognitive performance.

    So, cognitive remediation currently offers more hope than CBT – but addresses cognition rather than symptoms (which may, indeed be the problem for CBT – it is being sold to sufferers as a quasi-neuroleptic)

  • Dr Steingard
    Do you want to discuss heterogeneity in the paper that you blogged about here or are you shifting to discussing my own research (something in particular)? Or is it the general issue of heterogeneity in any kind of research with human beings?

    Re ‘Cognitive Enhancement’, are you referring to what is more commonly called ‘Cognitive remediation’?

  • Steve – Treatment as Usual (TAU) in this study has no relationship to what we normally understand by TAU – take a look at the paper – many patients in TAU were simply ‘discharged’ – that is not TAU

  • Saying “I know of people who have figured some of this out on their own. I gave an example above” is not part of the process of science of course…I am sure some homeopaths could give anecdotes of success

  • 1) The effect size for the PANSS change shown by CBT at 18 months is not significantly different from the PANSS effect size change shown by TAU at 9 months (i.e after no intervention whatsoever)

    If you read my post, I argue that all we are observing is random symptom fluctuations in a purportedly unmedicated patients

    2) Re “>50% symptom changes”, many of those showing improvement were medicated – see Table 4 in the paper

  • The stats I present are not complex – Effect sizes are essentially the mean of CBT minus the mean of TAU (divided by the standard deviation)
    Compare this to the inordinately complex stats performed by the authors of the paper!

    Anyway, looking at my post, you will see that if you compare the CBT and TAU groups at the end of the intervention (9 months) – they do not differ on total PANSS symptoms, positive PANSS symptoms or negative PANNS symptoms.

    In other words CBT has no impact on symptoms of psychosis – which is completely consistent with our recent meta analysis of 50 RCTs of CBT for psychosis in medicated patients (British Journal of Psychiatry) http://bjp.rcpsych.org/content/204/1/20