Coyne - What it takes for Queen Mary to declare a request for scientific data “vexatious”

[chipmunk1]

So, the usual BPS inference: environmental=psychological.

I read an article by Per Fink the other day, which promoted the old idea that gastric ulcers are caused by mental stress (or at least, stress is still a major component). His evidence? Greater reported incidence of ulcers after major natural disasters. Here again, we see the environmental=psychological conflation. Never occurs to him that after such an event, there may be a higher incidence of other infections, poor nutrition, hunger, thirst, disease, etc.

PS interesting article @soti!

when they see "environmental" they think "mental". So if you say you have environmental illness you have already lost.

and earthquackes cause peptic ulcers.

 

[Sasha]

Excellent comment from @Jonathan Edwards!

http://blogs.bmj.com/bmj/2015/12/16...-data-from-the-pace-trial/#comment-2413103576

I entirely agree with the proposal. We need more scientific and medical colleagues adding their voices to this. I think it is actually a pity even to mention the expression of emotion by patients. It simply reflects exasperation when faced with what is very bad science that purports to provide a solution to the problem. I think enough people have agreed for it to be pretty much a consensus, with the exception of the authors, that the PACE trial, as an unblinded trial with subjective endpoints, fails the most basic test of reliability. The fact that the MRC and Lancet have stood by it tells us nothing other than reminding us of the motto 'nullis in verba'. I assume that Richard Smith has read the paper and has seen for himself that it is uninterpretable.

This is a Canute case, indeed. We just need the tide of colleagues to start flowing in - and with respect, stop the discussion about patients' emotive response.

Yours sincerely,
Jonathan Edwards
Professor Emeritus
Department of Medicine
University College London

A whole load of comments have been approved now, including both of mine.

Although it needed saying, and everything was held in moderation so no-one could see the points that others were making, en masse I think we're now looking a bit top-heavy on "patients don't despise the mentally ill" argument - which has now been very well made - and we should be doing more to congratulate Richard Smith on his stance in favour of releasing the data.

There are some things in his post that we won't like but, eyes on the prize! Eyes on the prize! It's a HUGE DEAL that the former EDITOR OF THE BMJ has publicly and on the BMJ's OWN BLOG given KCL and QMUL a STIFF KICKING AND IS RECOMMENDING RELEASE OF THE DATA .

He may not be perfectly aligned with us on all matters but this is a huge, huge win and James Coyne has worked hard to get it.

So let's savour the moment and let's THANK THE GUY !

Good for him: good for us.

 

[Sasha]

Although Coyne didn't specifiy it was a PLoS One request in his letter, his tweets and blogs made clear that was the case. Which means that:

That's not correct, Simon - he did specify it was under the PLOS data-sharing arrangement.

Dear Professor McCrone
I have read with interest your 2012 article in PLOS One, “Adaptive Pacing, Cognitive Behaviour Therapy, Graded Exercise, and Specialist Medical Care for Chronic Fatigue Syndrome…”

I am interested in reproducing your empirical results, as well as conducting some additional exploratory sensitivity analyses.

Accordingly, and consistent with PLOS journals’ data sharing policies, I ask you to kindly provide me with a copy of the dataset in order to allow me to verify the substantive claims of your article through reanalysis. I can read files in SPSS, XLS[x], or any reasonable ASCII format.

Thank you in advance. I look forward to your response.

https://jcoynester.wordpress.com/20...lare-a-request-for-scientific-data-vexatious/

 

[snowathlete]

FOI vs PLoS release of data
I think there's some confusion over what James Coyne asked for (brought on by King's treating the request as an FOI). Coyne was asking for release under PLoS One rules, and those applying to that paper in 2012 aren't for full open access, I think, but are restricted to release to academics. Which is all Coyne wanted.






Although Coyne didn't specifiy it was a PLoS One request in his letter, his tweets and blogs made clear that was the case. Which means that:
a) FOI is no defence, since the authors agreed to release the data as a condition of publication and
b) the data isn't for public use (as I understand it, though PLoS One rules now are for public release of data).

As it's not for public use, the anonymisation bit doesnt come into it, as far as I can tell (though presumably there are basic rules about not including names, dates of birth etc.)

Also, PLoS rules provide no scope for policing which academics get the data, as Ben Goldacre is suggesting should happen through YODA. I don't have a big problem with YODA, which sounds like a worthy intiative, but it's not appropriate for a PLoS data request as it's more restrictive.

In short, I think Ben's got it wrong (or is simply being unreasonable), but that the real issue in this particular case is release of the data to any researcher who wants it, not full release to the public. A separate FOI request for a fully-anonymised (and limited) set of PACE data has been approved by the Inforomation commissioner, but is being appealed by QMUL.

Let me know if I've got this wrong

Thanks Simon, good post which clears things up for me: I hadn't realised the old PLOS policy in action when the paper was published only required data for academics on request. I had thought it was for anyone as one of Coyne's tweets suggested anyone could request it from PLOS.

Interesting that PLOS are more progressive than Goldacre (even their old policy. The new one even more so). Guess he's getting left behind a bit.

 

[Chrisb]

Both before I got the diagnosis and after, and well before it became the rage (mid 1980s - early 1990s), I voluntarily gave various psychotherapeutic approaches (both talking and pharmacological therapies) a long hard go, and none of them did a damn thing for me. In fact, in a number of ways I came out of it a lot worse than I went in.

This argument from them that it is just a fear of the stigma against mental illness is a complete crock of shite, and must be vigorously confronted and refuted.

You seem here to be crediting only Sharpe and White. I believe the proper abbreviation for the research is "shitely" to give equal prominence to the other protagonist.

 

[Sean]

Plenty of blame to go around.

 

[Sidereal]

New FOI request:

http://www.virology.ws/2015/12/17/a-request-for-data-from-the-pace-trial/

A request for data from the PACE trial
17 DECEMBER 2015
Mr. Paul Smallcombe
Records & Information Compliance Manager
Queen Mary University of London
Mile End Road
London E1 4NS

Dear Mr Smallcombe:

The PACE study of treatments for ME/CFS has been the source of much controversy since the first results were published in The Lancet in 2011. Patients have repeatedly raised objections to the study’s methodology and results. (Full title: “Comparison of adaptive pacing therapy, cognitive behaviour therapy, graded exercise therapy, and specialist medical care for chronic fatigue syndrome: a randomized trial.”)

Recently, journalist and public health expert David Tuller documented that the trial suffered from many serious flaws that raise concerns about the validity and accuracy of the reported results. We cited some of these flaws in an open letter to The Lancet that urged the journal to conduct a fully independent review of the trial. (Dr. Tuller did not sign the open letter, but he is joining us in requesting the trial data.)

These flaws include, but are not limited to: major mid-trial changes in the primary outcomes that were not accompanied by the necessary sensitivity analyses; thresholds for “recovery” on the primary outcomes that indicated worse health than the study’s own entry criteria; publication of positive testimonials about trial outcomes and promotion of the therapies being investigated in a newsletter for participants; rejection of the study’s objective outcomes as irrelevant after they failed to support the claims of recovery; and the failure to inform participants about investigators’ significant conflicts of interest, and in particular financial ties to the insurance industry, contrary to the trial protocol’s promise to adhere to the Declaration of Helsinki, which mandates such disclosures.

Although the open letter was sent to The Lancet in mid-November, editor Richard Horton has not yet responded to our request for an independent review. We are therefore requesting that Queen Mary University of London to provide some of the raw trial data, fully anonymized, under the provisions of the U.K.’s Freedom of Information law.

In particular, we would like the raw data for all four arms of the trial for the following measures: the two primary outcomes of physical function and fatigue (both bimodal and Likert-style scoring), and the multiple criteria for “recovery” as defined in the protocol published in 2007 in BMC Neurology, not as defined in the 2013 paper published in Psychological Medicine. The anonymized, individual-level data for “recovery” should be linked across the four criteria so it is possible to determine how many people achieved “recovery” according to the protocol definition.

We are aware that previous requests for PACE-related data have been rejected as “vexatious.” This includes a recent request from psychologist James Coyne, a well-regarded researcher, for data related to a subsequent study about economic aspects of the illness published in PLoS One—a decision that represents a violation of the PLoS policies on data-sharing.

Our request clearly serves the public interest, given the methodological issues outlined above, and we do not believe any exemptions apply. We can assure Queen Mary University of London that the request is not “vexatious,” as defined in the Freedom of Information law, nor is it meant to harass. Our motive is easy to explain: We are extremely concerned that the PACE studies have made claims of success and “recovery” that appear to go beyond the evidence produced in the trial. We are seeking the trial data based solely on our desire to get at the truth of the matter.

We appreciate your prompt attention to this request.

Sincerely,

Ronald W. Davis, PhD
Professor of Biochemistry and Genetics
Stanford University

Bruce Levin, PhD
Professor of Biostatistics
Columbia University

Vincent R. Racaniello, PhD
Professor of Microbiology and Immunology
Columbia University

David Tuller, DrPH
Lecturer in Public Health and Journalism
University of California, Berkeley

 

[Cheshire]

Wow, great letter!

We can assure Queen Mary University of London that the request is not “vexatious,” as defined in the Freedom of Information law, nor is it meant to harass. Our motive is easy to explain: We are extremely concerned that the PACE studies have made claims of success and “recovery” that appear to go beyond the evidence produced in the trial. We are seeking the trial data based solely on our desire to get at the truth of the matter.

Will Queen Mary University dare calling this request "vexatious"?

 

[Sidereal]

Will Queen Mary University dare calling this request "vexatious"?

I'm sure they will.

 

[A.B.]

Another small victory. Streisand or snowball effect?

 

[JamBob]

Wow, great letter!



Will Queen Mary University dare calling this request "vexatious"?

They'll probably reject the request as vexatious based on the fact that none of the eminent requesters are trained "Health Economists".

 

[A.B.]

They'll probably reject the request as vexatious based on the fact that none of the eminent requesters are trained "Health Economists".

They are requesting fatigue, physical function and recovery data.

 

[TiredSam]

They are requesting fatigue, physical function and recovery data.

That's easily dismissed, as none of the requesters are trained Fatigue Economists.

 

[Sean]

They'll probably reject the request as vexatious based on the fact that none of the eminent requesters are trained "Health Economists".

Maybe they aren't. But I bet they know and get on well with a few.

 

[Bob]

FOI vs PLoS release of data
I think there's some confusion over what James Coyne asked for (brought on by King's treating the request as an FOI). Coyne was asking for release under PLoS One rules, and those applying to that paper in 2012 aren't for full open access, I think, but are restricted to release to academics. Which is all Coyne wanted.

Good points, Simon. Another point to note is that Richard Smith seems to be advocating for unrestricted access to data...

Finally, the universities may have failed to notice that customs around sharing data in science are changing rapidly. We have recognised that huge value is lost by scientists taking their datasets to the grave with them. More and more funders of research require the release of data they have funded, and journals like F1000Research require authors not just to make their data available on request but actually to submit their data with their study so that anybody can use the data either to confirm or refute the study or do other studies.

I fear that QMUL and King’s are defending the indefensible and like King Canute failing to stop a tide that is coming in fast.

http://blogs.bmj.com/bmj/2015/12/16...-data-from-the-pace-trial/#comment-2413103576

 

[Simon]

Good points, Simon. Another point to note is that Richard Smith seems to be advocating for unrestricted access to data...

Yes, though not all academics are, and the primary goal in this case seems to be have the data analysd by academics, such as James Coyne, to ensure it's done properly.

So things have developed, but from my perspective the key things is that James Coyne gets the data he wants, and equally any other academic who wants access can have it, as per PLoS rules in force at the time of the paper's publication. That takes issues over FOI legislation/exemption, and anonymity, out of the equation. If an anonymised dataset becomes available to all, that's great, but I wanted to point out that's not what James Coyne asked for, and not what really matters in this case - in my view.

 

[Gijs]

Lets not forget that the data of the pace trial could be right as well.

 

[Cheshire]

Lets not forget that the data of the pace trial could be right as well.

Right in what? In proving that people with a SF36 score of 60 are healthy?

Their spin results are yet very poor and prove nothing close to recovery happened. If the hidden measures of the 6 minutes walking tests (even low) could have been used to sell CBT/GET, they would have been.

 

[worldbackwards]

Lets not forget that the data of the pace trial could be right as well.

I'd like to see them analysed with reference to the original protocol. That won't be pretty.

 

[soti]

Data can't be right or wrong. I think what Gijs wanted to get at was, maybe the data do support the authors' conclusions.

Already, even without data released, it's pretty clear that the methodology used makes the results in PACE "uninterpretable" -- i.e., we can't use the results in the paper to support the authors' conclusions, and specifically, we can't exclude the null hypothesis of no effect.

I don't myself know the analyses that people wish to run on the data. It's theoretically possible, as Gijs points out, that a different, more reliable method used to analyze the raw data would support the authors' conclusions. In principle, we have to wait and see what such a method spits out.

In real life, however, if such a method (e.g. the published protocols) supported the authors' conclusions, they probably wouldn't have gone to so much trouble changing their method, systematically introducing bias in the desired direction. Bit of a smoking gun there.