UK CMRC Nov Board meeting - Grand Challenge confirmed, Davey Smith keen

[A.B.]

What were the 'false allegations' they were complaining about? The rest of the letter makes them sound like dicks anyway, but maybe there was a false allegation?

I believe it's this article: http://www.independent.co.uk/news/science/me-bitterest-row-yet-in-a-long-saga-8348389.html

A British psychiatrist should be stripped of an award, fellow scientists said last night, as one of the most heated debates in medical science continued.

Professor Simon Wessely, one of Britain's foremost experts on ME, won the John Maddox Standing up for Science honour earlier this month. The prize was created by the journal Nature and the charitable trust Sense about Science. It was given to Professor Wessely for "courage" in speaking out about his studies into ME in the face a prolonged hate campaign and death threats.

 

[Bob]

If anyone else is struggling with the link to the minutes, this one works:
https://www.actionforme.org.uk/uploads/pdfs/CMRC-draft-minutes-111115.pdf

 

[Esther12]

I believe it's this article: http://www.independent.co.uk/news/science/me-bitterest-row-yet-in-a-long-saga-8348389.html

Would have been nice if they'd been clear about what they thought was a false accusation. Nothing clearly inaccurate there.

 

[user9876]

Seems odd to me, given the Science Media Centre's hopelessly skewed coverage of the PACE trial mediation study (which didn't actually show the claimed causal mediation) and follow-up study (a null result).


Good for Tuller and Coyne to make the points that patients apparently can't make, I really appreciate what they do...


... but think you might be missing important something here: The Grand Challenge.

They are talking about a 10,000-strong cohort. Nothing like that has been done in mecfs before. A £4.5m grant target from the Wellcome Trust, who basically joined the CMRC because they liked the look of the Grand Challenge. And potentially a whole lot more from other funders inc the MRC -frankly £4.5m will probably barely cover the cohort and sample collection, in my view. And 'omics galore; this is one area where mecfs could be well head of most other diseases. And George Davey Smith, one of the very best researchers out there.

All of this was put in train long before Tuller and Coyne published: Stephen Holgate in particular has been working on this for years, and deserves credit for his perserverance and achievement (OK, no funds secured just yet!).

When a much smaller versions of this study (OMF/Ron Davis, and NIH clinical study) were announced, people were rightly delighted. I think this could be much bigger and better, and by a long way the most exciting research study to be considered in the UK.

There are some things that worry me about the grand challenge and 'stratification' to form clusters. One of the issues is just how variable (over time) are peoples symptoms and various readings. Sometimes I have the impression that they can change quite a lot over time and this would have a significant effect on any clustering approach. This would be particularly true if they take a snap shot over 10,000 patients and then try to cluster rather than taking a more longitudinal approach. As a group I don't feel they listen to patients and this perhaps means they won't get the range and variability in symptoms.

In terms of blood results etc I tend to think that a detailed study needs to be done to see which are stable and which may be shocked by things like exercise. If a given measure is taken but affected by previous values and exertion then whatever measure is take is meaningless without the context. So with a large sample there could just be large variability unless carefully planned - but as far as I can tell medicine doesn't tend to think in terms of dynamic systems especially where we don't know to what the responses may relate.

 

[Bob]

IMO, the whole purpose and meaning of the CMRC rests on the success of the large study; and everything else about the CMRC can be considered froth in comparison.

If the big study is a proper solid study then the CRMC will have been massively worthwhile. I like the idea of the large study. And in principle the issue of a broad cohort isn't too problematic because participants can be sub-categorised by various diagnostic criteria (CDC, NICE, IOM, CCC, ICC).

If the study uses the UK's biobank, then I won't have any problems with the patient cohort. If I remember correctly, the biobank patients are diagnosed using CDC and CCC. The biobank patients would be a heterogeneous group, but if the researchers recognise this and are aiming to distinguish subsets (which they are) then this isn't a problem.

But... I agree with what people have said about the major issue of bias on the collaborative board. And if Esther Crawley and colleagues are in charge of recruiting (e.g. recruiting from their clinics) for the study, then we potentially have a major problem, whatever the official recruitment criteria. It has been said that Crawley retracts a CFS diagnosis if her patients don't improve after 'treatment'. In which case, long-term patients are 'disappeared' from the stats. And, although we will never know the exact details, the PACE trial excluded patients with neurological symptoms, and there would be similar potential opportunity to exclude severely affected patients with neurological symptoms from this big study for similar reasons. Such a scenario would obviously introduce serious bias into the study, no matter what the recruitment criteria are. So they need to be removed from the recruitment process to gain the confidence of the patient community. (Crawley and her colleagues don't recognise ME/CFS anyway, because they don't understand the illness: they are only interested in false illness beliefs.)

If the study doesn't have the confidence of the patient community, then we know what happens. The patients' concerns are ultimately validated, and the aftermath is a long-drawn-out-process, divisive, and painful for all sides.

 

[Simon]

Does Esther (SMILE) Crawley have the background to identify the 'critical scientists' for this 'high-level science-led endeavour'?

I suspect she had a big hand in bringing George Davey Smith into the fold (they are both at Bristol).

Prof George Davey-Smith signing that letter along with Professor Sir Mansel Aylward et al doesn't exactly suggest a maverick, anti-establishment free thinker tie or no tie.

Well,we still don't know what the allegations referred to are. But sod the tie, he wrote this:
Problems of reporting genetic associations with complex outcomes (The Lancet, 2003)

Basically it says all gene association studies done to that point - including all those published in the Lancet - were a waste of space: "it didn't exactly make us popular", he said at the CMRC conference. It's been cited over 1,000 times and revolutionised the field: since then, driven by funders such including Wellcome Trust, studies have involved large collaboratives and huge samples. Almost no gene associations published up til 2003 replicated, pretty well everythig published since 2005 has held up. Takes nerve to stand up and tell everyone - including the journal publishing your article - that they've got it all wrong. Then there's this. Davey Smith is the real deal.

 

[Valentijn]

Davey Smith is the real deal.

Then my concern is that he doesn't know much about ME, and is going to be getting the wrong impression from the wrong people. There's still a lot of ways this can all go very wrong.

 

[Simon]

Yes, sounds great on paper. My concern would be, given the involvement of certain usual suspects, whether these 10,000 patients are actually going to have ME or some kind of ill-defined unexplained fatigue and depression. I actually think small but extremely well characterised patient cohorts are a better starting point which is what Ron Davis is doing.

I suspect case definition (the broadest one used) will be a key part of the debate at this project developed. And I'm sure the use of severe bedbound cases will be discussed too (extreme cases have proved useful in studying other diseases.

But say mecfs is at least a dozen different diseases (turns out breast cancer is 13+, asthma is 8 distinct grroups) - a small sample is unlikely to be truly well-defined, more a hodge-podge of different subgroups. That would make it very hard to interpret the findings. A large group gets round this. Plus you need to include plenty of people at the margins, to define where mecfs groups end and more general chronic fatigue starts. That said, I would be deeply concerned if they used something as broad as the oxford definition for all patients, as it risks making most of the sample a group that might be of marginal interest (as opposed, say, to sampling Oxford-criteria patients to make sure they are well-represented, but using somethiing a bit tighter for selecting most of the cohort).

There are some things that worry me about the grand challenge and 'stratification' to form clusters. One of the issues is just how variable (over time) are peoples symptoms and various readings. Sometimes I have the impression that they can change quite a lot over time and this would have a significant effect on any clustering approach. This would be particularly true if they take a snap shot over 10,000 patients and then try to cluster rather than taking a more longitudinal approach

In terms of blood results etc I tend to think that a detailed study needs to be done to see which are stable and which may be shocked by things like exercise.

Agree, that's a good point about longitudinal factors, and I think stress-testing (including cognitive challenges) will be critical.

btw, done for today, back tomorrow

 

[Marco]

I suspect she had a big hand in bringing George Davey Smith into the fold (they are both at Bristol).

Well,we still don't know what the allegations referred to are. But sod the tie, he wrote this:
Problems of reporting genetic associations with complex outcomes (The Lancet, 2003)

Basically it says all gene association studies done to that point - including all those published in the Lancet - were a waste of space: "it didn't exactly make us popular", he said at the CMRC conference. It's been cited over 1,000 times and revolutionised the field: since then, driven by funders such including Wellcome Trust, studies have involved large collaboratives and huge samples. Almost no gene associations published up til 2003 replicated, pretty well everythig published since 2005 has held up. Takes nerve to stand up and tell everyone - including the journal publishing your article - that they've got it all wrong. Then there's this. Davey Smith is the real deal.

So why does a scientist who doesn't trust small scale gene association studies feel it necessary to sign his name to a letter only supported by a few not exactly impartial researchers about alleged abuse?

 

[Sasha]

Then my concern is that he doesn't know much about ME, and is going to be getting the wrong impression from the wrong people. There's still a lot of ways this can all go very wrong.

Then we need to be sure that George Davey Smith is talking to the right people, such as @Jonathan Edwards.

How about inviting him to the IiME conference? Or pre-conference colloquium?

 

[A.B.]

George Davey Smith's twitter account shows him following Coyne and Kindlon... or is that Kindlon and Coyne following him? Still not sure how to use twitter.

Perhaps we should invite him to phoenixrising?

 

[Tom Kindlon]

George Davey Smith's twitter account shows him following Coyne and Kindlon... or is that Kindlon and Coyne following him? Still not sure how to use twitter.

Perhaps we should invite him to phoenixrising?

He doesn't follow me but I do follow him (which is less interesting I know)

 

[Esther12]

He doesn't follow Coyne either. That must have been people who follow him.

 

[snowathlete]

The concern I have, simply put, is that some of the people involved are not very good. In fact, some have a shockingly bad history behind them in ME/CFS. They are experts by default, not because they have the ability but because they are the only ones who have been allowed grants for funding in this country, and that's because they view the disease as one of "false illness beliefs". If you run projects with bad people in place you tend to get bad results, whether its medical research or anything else, and these are some of the worst scientists out there.

It's also a big ask to expect any patient to welcome the involvement of persons who have caused detriment to patients with their prior involvement in the field. They don't seem to care about the fact there are real people suffering from this disease who they cause harm to, it's all about them and their careers. And I'm not repeating what someone has told me, I've seen them in action for myself; some of those mentioned have nothing but disdain for patients. Contrast that to the OMF/Ron Davis, their work is 100% about helping patients who they respect and listen to and never bad mouth or risk harm to.

It sounds like there are some good people involved in this project as well, so I hope it works out. I hope those coming in to work on ME/CFS make great efforts to seek out real patients for themselves and look objectively at some of the prior actions of their new colleagues that have caused such upset so that they understand what they are really dealing with. Otherwise despite their best efforts and intentions they could find themselves being manipulated.

So I'm sorry to say I do feel rather cold about this, as I do for much that comes out of the collaborative because there are bad eggs involved. I will be delighted if things turn out wonderful despite this, and they produce ethical and quality research.

 

[jimells]

It will be interesting to see who is appointed Principal Investigator(s).

 

[SOC]

If the study uses the UK's biobank, then I won't have any problems with the patient cohort. If I remember correctly, the biobank patients are diagnosed using CDC and CCC.

Who is diagnosing them using the CDC and CCC? If it's doctors who can't distinguish between exercise intolerance and PEM, then the biobank could be full of patients with other fatiguing conditions. Also, doctors with biases toward psychogenic theories are going to "see" ME/CFS in patients they expect to see it in and convince themselves they're applying the CDC or CCC criteria. It's not that hard to stretch and bend either of those definitions to include someone with fatigue and some other symptoms. Not that the CDC criteria (which one?) is particularly good at distinguishing ME from other fatiguing illness even when applied by moderately knowledgeable doctors.

I know some people are desperate for something good for ME to come out of the UK, but realistically, does this look like it's not going to be one more cluster f*** for patients when there are SO MANY biased people and so few truly experienced people (about biomedical factors in ME)?

This has FUBAR written all over it. It's a big study with an extremely diverse patient group which may or may not be clearly ME patients. This is the kind of thing that ends up with "No consistent finding, therefore there are no identifiable abnormalities in ME. We studied 10,000 people so this must be considered complete and definitive. No biomedical findings found. ME must be psychogenic. End of discussion." I can see the BPS crew drooling over this one -- a huge slopply-done pretend biomedical study that proves nothing physical is wrong with us. There goes all our funding. The truly good researchers won't be able to get funding because this big study just proved there's no consistent biomedical finding in ME. Look what PACE and other BPS research did to destroy biomedical funding for ME. This could be thousands of times worse because it superficially looks like biomedical research.

Don't think they don't know what they're doing. The BPS folks wouldn't be involved if they thought they were going to loose careers and reputations as a result of this research. They know it's very easy to screw with biomedical studies by choosing a poor cohort, looking in the wrong places, improperly interpreting the results, and so on. SW has done some of that already, iirc.

At this point I don't see that there's anything we can do about this. It's going to happen whether we like it or not. If it sets back biomedical research another 20-40 years... time I don't have... I'm going to be mighty furious at everyone who supported this, ignoring the extreme potential for harm when BPS so-called ME experts join with intelligent, but unknowledgeable researchers. With the BPS crowd as their go-to experts, do you think these researchers are not going to have an extremely biased view of ME and ME patients?

 

[medfeb]

I have a similar concern regarding what disease they are studying.

Holgate's interest in the IOM criteria and capturing "the whole population" is not encouraging. The IOM criteria are not intended for research. Even for clinical use, the lack of exclusion criteria and the lack of standards for how you assess symptoms like PEM means the criteria are likely to be diagnostically unreliable. Jason's study suggested that the IOM criteria encompasses a substantially bigger set of patients than even Fukuda because IOM does not exclude any illnesses. Peterson has also raised concerns with the IOM criteria being overly broad. And this will be a real problem if the investigators think the disease is psychogenic since primary psychological illness is included.

I get the point about research identifying many subtypes of breast cancer. But I assume they all had breast cancer to begin with and then subtyped from there.

So the question for me is whether they will create a 10,000 patient cohort of patients that have ME and use that to identify subtypes of ME? Or are they going to end up with a non-specific collection of fatiguing conditions and then try to use big data approaches to make sense of it? Maybe I am missing something but I don't see how big data approaches will make sense of a 10,000 patient cohort if you don't have a tight handle on what disease is in there.

 

[Dolphin]

(A lot of the points I would have made have already been made)

One thing that stands out to me is that £4,000,000/10,000 patients = £400 per patient. And drawing blood (and probably often transporting it somewhere), collecting questionnaire data, etc. costs money. So one wonders how much will be left for biomedical tests. What I could see happening is that a lot of cheap data will be collected from questionnaires and not much biomedical will be measured. As @user9876 mentioned, symptoms are often not very stable.

 

[Dolphin]

I imagine the most likely way they will recruit 10,000+ patients is through the existing CBT/GET/similar NHS clinics. In some ways, this might bring down some costs so has advantages. But if such clinics are used, it would mean it would be more likely those of the CBT/GET school of thought would have influence.

 

[Sean]

Hmm. All a bit troubling.

Having Crawley in such an influential position is troubling enough on its own.

Hope Holgate knows what he is doing, and doesn't get outmanoeuvred by the psychs.