The complexity of cracking open the data door
Data-sharing advocates have been calling on manufacturers to release the raw data from clinical trials so that researchers can look at all the parameters related to specific drugs, not just the primary or secondary endpoints.Last week the trade group PhRMA and its European counterpart said their members are ready to do so. The data door may be opening, but the transition will be a complex one.
The commitment by PhRMA, and the European Federation of Pharmaceutical Industries and Associations, came as the European Medicines Agency and the European Parliament consider more liberal data-transparency rules. The proposal to amend the EU's clinical trials directive is scheduled for debate in Parliament in November. And court cases are pending—AbbVie sought an injunction to stop the EMA plan to release detailed information that it holds on its drug Humira for RA starting in January.
As such, industry's proactive effort, set to start Jan. 1, may have headed off something of the less-nuanced sort with a principled approach that would allow access to clinical trial data in a more open forum.
"Our motivation here is to expand the science to bring in more and more of the qualified, legitimate and responsible researchers, so that ultimately we advance our knowledge and our ability to serve patients,” said John Castellani, PhRMA president and CEO, during a briefing for reporters.
Unlike the European-style plans to release anonymized trial data as soon as a drug is approved, industry's model comes with other caveats. Data-sharing is subject to terms necessary to protect not only patient privacy but also confidential commercial information, which is necessary to maintain incentives for continued investment in biopharmaceutical research.
It's that last point where the principles differ from the EMA proposal. “PhRMA and EFPIA back AbbVie's claim that clinical study reports are commercial confidential information, and I see the principles document as an effort to define an alternative vision,” said Peter Doshi, a postdoctoral fellow in comparative effectiveness research at Johns Hopkins University School of Medicine.
For years, industry has voluntarily shared data with third party investigators, says Doshi, who has been conducting systematic reviews of the medical literature since 2009, mostly with regard to neuraminidase inhibitors.
“I see the recent principles document as a formalization of this historical standpoint which states clearly that the intention is to share data with only ‘qualified scientific and medical researchers' doing ‘legitimate research,'” he told MM&M by email.
Data requestors will be required to submit a research proposal to document this, something the EMA's proposal doesn't call for. To vet the requests, each company must establish a scientific review board that will include one or more outside scientists and/or HCPs.
“We expect (the boards) will be populated by experts from within and from outside the company,” Castellani added. “They will disclose decisions for all to see.”
As an academic researcher, Doshi said he'll be looking at the speed at which applications for data are reviewed, and whether the review process will transparent “such that outsiders can judge for themselves whether the decisions are swift and fair.”
As it stands now, under the FDA Amendments Act of 2007, firms must disclose trial results on ClinicalTrials.gov upon product approval, but not the detailed patient-level data. So far, GlaxoSmithKline and Roche have taken steps to go further, appointing independent review boards and promising data openness. For the rest, evolving procedures to allow more transparency won't be easy.
“In the US, it's going to take some time,” not only to establish the board but also for each firm to draft and approve its own set of internal standard operating procedures, or modify existing ones, said Bob Paarlberg, a regulatory consultant who has lectured about clinical trial data transparency. “It's not something that's going to happen (by) next week.”
But it has to happen soon. And it's not unreasonable to think there may be lingering resistance. Both trade groups spoke out against the Europeans' data-sharing plans in February. And the (UK) Guardian reported on a leaked email that detailed plans by the two groups to mobilize patient groups against the changes.
“When there's a discovery like [the memo leaked to the Guardian], that suggests the internal battles haven't quite ended,” said Stephen Friend, president, Sage Bionetworks, a non-profit foundation in Seattle.
“Probably in all companies there are basically subgroups of pundits that are battling inside—‘We cannot let that data out.' ‘Yes, we should.'—so it's impossible to ask for a coherent stance in any one company or across all pharma,” he said, comparing the debate to pharma's own version of the eternal battle between Jungian dark and light forces.
Nevertheless, all members of the trade groups have signed on, said Castellani, a sign that more people are recognizing the power of sharing data. Together, they represent 80% of manufacturers. (That doesn't include the small-to-midsize companies.)
Each company uses a different data structure to record its clinical trials, and this poses another hurdle for compliance. “Not everyone would necessarily be using the same formats for the data, and so in a sense you could be comparing apples to oranges,” said Paarlberg.
This could become an issue when releasing information to the general public. Lay people could reach invalid conclusions.
“[Transparency] needs to be done in a responsible manner,” said Paarlberg. Misinformation “could cause harm for patients in, say, not taking their medications—‘I read this so I'm not going to continue.'”
To prevent that scenario, the principles call for companies to provide a factual summary of clinical trial results and make the summaries available to the public. “[The raw data] will not be generally understandable to the public that looks at it,” said Castellani. “So what we've adopted in our principles is a bifurcated approach”—one for researchers, the other for participants and others.
Access to more data may not necessarily lead to a groundswell of newly published research. In the past, some journal editors have been hesitant to publish manuscripts of negative trial results. Although some of the strongest proponents of data-sharing have been journal editors. The BMJ this year began requiring the relevant anonymized patient-level data from trials to be made available for third-party study as a condition for publishing academic papers on drugs or devices.
But having a commitment to share results on all new drugs and all new uses for existing drugs, whether approved here or in the EU, positive or negative, could help researchers assess the risks and benefits of products. It could also help researchers learn from failed trials. Data will include investigational medicines whose development programs have been discontinued.
Friend doesn't see industry's attempt being derailed by hypocrisy. “I am optimistic that companies, in general, have been moving toward more access to data and more sharing, and that's generally where it's going,” he said.
The ball is now in individual companies' courts to get up to speed. Added Paarlberg, the consultant, “I haven't had my phone ring yet based on the principles, but it's an evolving topic for all of us that have worked on disclosure.”