In the standoff over release of the PACE PLOS One trial data, has the journal just blinked? by James Coyne, 7 April 2016

I just received (April 7, 2016) another communication from the Managing Editor of PLOS One reporting  an emerging position on my gaining access to the PACE trial data, promised to be available as a condition for publishing in PLOS One.

The negotiations are not over. But is there a signal that PLOS One is prepared to render meaningless their distinctive commitment to full data availability and sharing?

Any retreat from full and unrestricted availability of data has implications for those of us who work for free as Academic Editors, those of us who submit our manuscripts to PLOS One rather than somewhere else because of its data sharing policy, and even those institutions who been willing to subsidize hefty publication costs because of the distinctive claim that data from published articles will be available to all.

If PLOS One abandons or even waters down its commitment to data sharing, will we abandon PLOS One?

As I’ve stated before, despite being in one of thousands of PLOS One Academic Editors, I have no special insight or influence over the decision-making in this matter. Indeed, the PACE investigators have more insight into what is going on, not only because they are directly involved in the negotiations, but because maybe PLOS One could be granting them an extraordinary latitude in whether they can refuse or bureaucratize what had been my simple request for the data.

The e-mail:

“Dear Jim,

Thank you for the correspondence regarding the article by McCrone et al.

As you know we are actively following up on the request for data from the article. We sought advice from two editorial board members in order to establish what data we would expect the authors to share in the context of the current analysis, and we have followed up with the authors. The authors have raised concerns about the need to protect patient privacy as well as the specifications outlined in the consent obtained at the time of recruitment for the trial; our follow up with the authors is ongoing.

We plan to contact Queen Mary’s University of London to discuss our policy and position in relation to the sharing of data from this study.

I would like to take this opportunity to outline what we can and cannot engage in as part of our follow up as editors.

The policy that applies to the article indicates the following:

Publication is conditional upon the agreement of the authors to make freely available any materials and information described in their publication that may be reasonably requested by others for the purpose of academic, non-commercial research.

Availability of data and materials. PLOS is committed to ensuring the availability of data and materials that underpin any articles published in PLOS journals. PLOS’s ideal is to make all data relevant to a given article and all readily replaceable materials immediately available without restrictions (while not compromising confidentiality in the context of human-subject research).

The policy therefore outlines that data should be shared for the purposes of academic research and without compromising confidentiality in the context of human-subject research. Our goal is for authors to release as much data as possible, however the policy cannot supersede the requirements for ethical/data access oversight that may apply to the use of datasets involving human subjects.

It is not within our remit as editors to judge what the requirements for consent, patient privacy or data de-identification should be, such considerations are handled by institutional committees (such as ethics committees), which are equipped to assess and rule on such matters. In the context of our policy we will follow up on any and all requests for data underlying publications in the journal, but there may be situations where such follow up establishes that access to the data requires a process of evaluation by an ethics committee. If that is the case, our position is that the readers requesting the data should pursue such approval process in the context of a defined research proposal, in the same manner as the researchers who undertook the work described in the article did.”

Commentary

In making this new statement available, I believe the scientific and larger communities have the right to know what is going on and should have the opportunity to comment. We are all stakeholders and although PLOS One does not apparently have a mechanism for soliciting such comments, I think we can be heard and influential.

Here is some preliminary observations concerning the communication from PLOS One –

It’s reassuring that the journal reaffirms its commitment to making all data relevant to the article available, but it’s worrisome that PLOS One is legitimizing previously expressed objections of the PACE investigators.

Namely, sharing the data would be inconsistent with their purported commitment to protecting privacy and to honoring the conditions laid out in obtaining patient consent to participation in the trial.

In a decision last winter, the UK Information Commissioner’s (IC) rejected arguments from the PACE investigators that fears of breaching confidentiality because of the inability to anonymize data precluded making the data available. As we now know, anonymizing can be accomplished with readily available technologies.

The recruitment and consenting materials that I have reviewed do not mention any commitment to restricting access to anonymized data.

Indeed, the Medical Research Council (MRC) funded a sister trial to PACE for which data has already been readily available. I would be curious to know if there are any relevant differences in recruitment and consent materials between the two trials.

Furthermore, in recent arguments for rejection of a request for data from a PhD physicist, the PACE investigators did not even mention anonymizing or any concern for keeping the promises provided in obtaining patient consent as reasons for denying her the data. A letter from Peter White to the Wall Street Journal acknowledged the PACE trial data can be effectively anonymized.

Finally, the PACE investigators have disclosed that they have already shared the data with researchers outside their group.

It’s not clear whether the requirement for review of a request by an ethics committee for data opens new loopholes for withholding the data. I’m concerned about a non sequitur being introduced here: that the application to the committee has to involve a pre-specified research plan.

Where did that requirement come from? The PACE investigators notoriously changed their endpoints from their protocol. This raises curiosity not only about what results would’ve been obtained with the original endpoints, but also about what may be uncovered by more exploratory, even forensic, analyses probing what might have led to the decision and what the PIs might be avoiding.

Having the data from published papers available allows independently checking the basis for the original investigators’ claims. This becomes particularly important – as in the case of the PACE trial – when investigators have conflicts of interest and when the conflicts of interest were not disclosed to the study participants.

However, another widely accepted rationale for data sharing is that clinical trials are expensive and impose considerable burdens on patients who participate in them. The best use of the public funds that supported trials and of the patients’ contribution is to make data available for exploring research questions that were not anticipated, without the added expense and patient burden of conducting a whole new trial. I anticipate some anomalies being uncovered that would best be examined in analyses explicitly labeled and interpreted as exploratory.

It’s a relief that the PLOS One administration seems to be avoiding any endorsement of the kinds of loyalty or character tests that were so pivotal in refusals to release the data to date. The PACE investigators may have turned my request for the data into a matter falling under the freedom of information act and rejected the request because my alleged vexatiousness and impure motives in wanting the data.

I’m sure that many of you are disappointed that after all this time, an email from the Managing Editor of PLOS One did not simply provide simple instructions how I and others could access the PACE trial data.

I had also assumed that my request would yield circumstances by which everyone could access the data. It now appears that anyone wanting to access the data, needs to anticipate long delays and a frustrating and potentially uncertain bureaucratic process.

Am I missing anything in this communication?

We need to keep in mind that we always have the option of demanding a retraction of the PLOS One article if the investigators set unreasonable conditions for independently scrutinizing their dubious claims.

And then there is the much-anticipated final decision on April 22, 2016 concerning Queen Mary University London appeal of an Information Commission decision requiring release of the PACE trial reported in The Lancet.

We don’t share data.’ Why Peter White’s Wall Street Journal letter can be ignored, James Coyne,  26 Mar 2016:

James Coyne explains why he thinks the arguments in Peter White’s letter to the Wall Street Journal can be ignored. He closes with:

Author, historian, and social critic Bertrand Russell (May 18, 1872–February 2, 1970) could offer him some useful
advice:

• Do not think it worth while to proceed by concealing evidence, for the evidence is sure to come to light.
• Do not use power to suppress opinions you think pernicious, for if you do the opinions will suppress you.
• Be scrupulously truthful, even if the truth is inconvenient, for it is more inconvenient when you try to conceal it.