Freedom of information: can researchers still promise control of participants’ data?BMJ 2016; 354 doi: https://doi.org/10.1136/bmj.i5053 (Published 21 September 2016) Cite this as: BMJ 2016;354:i5053
- Nigel Hawkes, freelance journalist, London, UK
The consent agreements researchers make with patients participating in trials will have to be rethought, experts have warned, after researchers were forced by a court ruling to release data from a controversial trial despite not having obtained the participants’ authorisation to do so.
Queen Mary University of London said this month that it has released the data from the PACE trial (a randomised evaluation of pacing, graded activity, cognitive behaviour therapy, and specialised medical care for chronic fatigue syndrome1) to a patient who had requested it under the Freedom of Information Act. The university had long battled against the request, citing concern that the participants could be identified if the data were released.
Frances Rawle, head of corporate governance and policy at the Medical Research Council, says that it will now be necessary to rethink how consent forms for future trials are drafted if it is no longer possible to assure participants that data will not be shared.
“People who run trials have to say what will happen to the data they collect, so we’ll have to look very carefully at the wording of consent forms,” she says. “As things stand they are no longer consistent with the Freedom of Information Act.”
If people have to be told that data may be shared, even if only after anonymisation, it could affect recruitment for trials, she said.
Questions over the data
The PACE trial, which was published in the Lancet in 2011,1 included 641 patients and remains the most important research to date into treatment for chronic fatigue syndrome (also called myalgic encephalomyelitis). But patients and experts have challenged the trials’ methods and its influential finding that cognitive behavioural therapy and graded exercise therapy are more effective than adaptive pacing therapy or specialist medical care. Activists and charities that represent patients favour adaptive pacing, which involves a gradual increase in activity aiming to improve people’s capacities over time, and reject psychological treatments.
Alem Matthees, who made the freedom of information request, had, like many patients, questioned the research methods and the original findings, saying the researchers had not analysed the data as set out in the original trial protocol.
Experts have also questioned the published results. Former editor of The BMJ, Richard Smith, called for the release of the data in December 2015, accusing the researchers, along with Queen Mary, of “defending the indefensible.”2 And earlier this year, 42 experts wrote an open letter calling for a third party review of the data.
Vincent Racaniello, professor of microbiology and immunology at Columbia University in New York, who organised the open letter, said the decision of the tribunal was “a victory on multiple fronts: for patients seeking the proper therapies, but also for transparency in science and the need to have data available for review by others.”
What patient information was released?
Matthees has been sent what he asked for, Queen Mary says, but it has not been posted on the university’s website. This is not obligatory, though most government departments do post freedom of information releases on their websites. The data are now in the public domain, and Matthees can make them available if he wishes. It was not possible to contact Matthees before this article went to press. Anybody else who wants the same data can now apply to Queen Mary for it.
What Matthees sought was a spreadsheet listing the baseline and 52 week follow-up data for all trial participants, including their physical function scores, their fatigue scores (in two forms), whether they were still classified as having chronic fatigue syndrome by the Oxford caseness classification at 52 weeks, participant and doctor related impressions of overall health, and walking distances.
This is by no means all the trial data. “There are about 3000 variables in this trial and we’ve released 12,” says White. “These are the primary outcomes and two secondary outcomes, plus some other things. It’s a limited release because that’s what he asked for.”
Matthees told the commissioner he wanted to analyse the data using the original 2007 protocol. One of the principal complaints of critics is that the protocol was changed after data started to come in and that if this had not been done, different conclusions would have been drawn.
White says these changes were made before any data were examined, were ratified by the trial steering committee, and reported in the 2011 Lancet paper.
To anticipate any analysis that Matthees plans, the PACE team published its own this month.3 4 What is the result of going back to the 2007 protocol? “It makes not a ha’porth of difference,” White says. “Cognitive behavioural therapy and graded exercise therapy are still statistically significantly better than pacing or specialist medical care.”
The patient advocacy organisation #MEAction and others have claimed, however, that these new analyses show only a third as many patients improving as the Lancet paper found. “They’re not comparing like with like,” White says. “They are comparing one measure with a completely different one—it’s apples and pears.”
Landmark tribunal ruling
In August a tribunal rejected Queen Mary’s argument that to release the information would risk identifying patients who had participated in the PACE trial.5
The PACE trialists, led by Peter White, professor of psychological medicine at the university, had answered critics who have made legitimate scientific points and shared the data with other research groups. But they had been unwilling to share it more widely. A total of 37 freedom of information requests were made, many successfully rejected by Queen Mary under section 14 of the act, which covers vexatious applications.
But when Matthees, a patient with chronic fatigue syndrome from Perth in Western Australia, made a request in March 2014, Queen Mary did not play the vexatious card. Instead it rejected the request under section 40, which relates to the protection of personal data, and section 41, information provided under a duty of confidence. It also attempted to invoke section 22A, which was written into the act in 2014 to protect medical trial data from improper use. However, this did not come into force until October 2014, some months after Matthees had made his request, and the commissioner ruled it could not apply retrospectively.
That left the personal data issue. The commissioner found that anonymisation of the data would be an adequate safeguard. In doing so he disregarded the editor of the Lancet, Richard Horton, who argued that there was “a fairly small, very vocal and very damaging group of individuals who have, I would say, actually hijacked this agenda and distorted the debate so that it actually harms the overwhelming majority of patients. This community actively seeks to identify and attack those who are associated with the PACE trial.”
The Lancet has since had nothing to say about the ruling.
Queen Mary took the case to a tribunal, which ruled that if the university was prepared to share data with other researchers it was tacitly acknowledging that anonymisation is effective and dismissed as “implausible” claims made by data security expert Ross Anderson, a professor at Cambridge University, that anonymisation could be broken by linking different databases. By a majority of two to one—the lay member dissenting—the tribunal ruled that Matthees should get the data he asked for. After mulling it over, Queen Mary agreed to comply.
See related blogs by Peter White, lead researcher, and a patient on thebmj.com.
Competing interests: I have read and understood BMJ policy on declaration of interests and have no relevant interests to declare.
Provenance and peer review: Commissioned; not externally peer reviewed.