When parents of children with Duchenne enroll their children in a clinical trial there is often a sense of participating in something bigger than themselves—of course you hope that your child may benefit from a new therapeutic approach, but you also know that all those days of missed school, waiting rooms, endless syringes, and painful biopsies are contributing to an ever-growing wealth of knowledge that can be brought to bear against this disease. Most importantly, you have volunteered your child and your family for this role knowing that your child might be assigned a placebo or even that the treatment may not work. It feels like an implicit social contract that you will do your part and then the scientists and drug development experts will do theirs.
But what happens when the volumes of data that are generated from a clinical trial, whether successful or not, remain in the hands of a single company or a single academic investigator…sometimes for years while additional analysis is completed and publications are drafted and submitted? Even when immediate release of the data could help avoid duplicative research, improve safety in future trials, or decrease the time it takes to conduct additional trials? Has that implicit social contract been violated? Do the scientists in academia and industry, who have put millions of dollars and thousands of man hours into conducting the research, deserve the time to analyze the fruits of their labor? What is the incentive to do the research if they are forced to give away the results immediately? Consider that the majority of people (according to a recent Research!America poll) who enroll in trials do so for personal gain and not for altruistic reasons—can we expect the academicians and drug industry to behave more altruistically than we do are ourselves?
Although this is the situation now, Society seems to be moving slowly in the direction of making clinical trial data a public, rather than private, resource. In July of this year the Food and Drug Administration (FDA) released a request for public comment on its draft plan for clinical trial data sharing. The FDA plan follows on the heels of a data sharing plan that will be enacted by the European Medicines Agency (EMA) by the end of the year allowing public access to data submitted to the Agency in support of new drug applications. With changes looming and feedback requested by both the US and European regulatory agencies, the Institute of Medicine is developing a report that will describe guiding principles and a framework for the responsible sharing of clinical trial data. I was asked to speak at the IOM Committee’s first public meeting on this topic, as a representative of a patient organization. The meeting took place in Washington DC on October 23rd.
Speakers at the start of the day began by describing the issues that arise around making clinical trial data freely available. They talked about the importance of making available “Individual Participant Level” data, or “IPL” data, rather than the aggregate data that is typically released by trial sponsors in summary format for publication or on clinicaltrials.gov. To be clear, IPL data typically means all data from a single individual who participated in a trial, identified by a random number but not by name—it is “anonymized” individual data. A true re-analysis of trial findings can only be conducted from scratch if this IPL data is available.
So, the important thing to know about IPL data is that, although it may start off as anonymous data, there frequently is so much demographic data included in these files that it would not be terribly difficult to cross-reference with other databases and “re-identify” participants in trials. Experts in this kind of data management admit that it is virtually impossible to truly anonymize IPD data that goes into the public domain in any way that doesn’t render it useless. The problem is particularly true for participants in small trials or rare disease trials….like Duchenne trials. So privacy issues are a concern.
In general, most of the expert speakers were agreed that the benefits of data sharing include:
Overall, the benefits of data sharing are tremendous and not particularly controversial, but the risks of data sharing and how to manage those risks was a major focus of the meeting.
Risks to data sharing were generally defined as:
From a practical standpoint there were questions around what, where, who, when, and how. For example, should all clinical trial data be made freely available as soon as a trial is completed to anyone who asks, or should there be some kind of vetting process to verify that the data was being requested with a reasonable scientific objective in mind? If there is a vetting process, who is in charge of that process—can the trial sponsors themselves determine who has access to the data or should there be an unbiased independent group of decision makers? Should the data be turned over to a third party or be allowed to remain with the trial sponsors? How will the efforts of those who produce the clinical data in the first place be recognized and protected? Will there need to be data standards in place so that trial results can be easily compared and accessed in similar formats? (Answer is yes, but then who decides what those data standards are and how do you enforce them?).
Several groups with data sharing projects up and running described ways this could be done on a smaller scale and a few meeting participants put forth suggested models…but on the whole, this was a meeting that was more about questions than answers. Because the issues are so complicated some meeting participants recommended that we proceed slowly and cautiously with a data sharing plan rolled out over the next ten years.
In the afternoon we heard from “stakeholders” – groups who stood to be affected by requirements to share clinical trial data. Sandwiched between panels of 4-5 academic investigators and industry trial sponsors was a very short session featuring two patient advocates. We were glad to contribute to the discussion, but wished that our panel had more than two people on it.
I had five minutes to provide some prepared remarks. I told the committee about Duchenne muscular dystrophy—emphasizing the clock starts ticking with the diagnosis. I described how our only validated endpoint, the 6MTW, was effective over a narrow window that put boys assigned to placebo at risk of losing ground that could not be regained, even if crossed over to a drug arm. And I pointed out that pooling the placebo arm data generated from multiple industry trials could help us mathematically model the natural history in such a way that we might be able to shorten or eliminate the placebo. I also told them what you told me—it is unethical for investigators not to share the data generated in Duchenne trials for the greater good and that the majority of Duchenne parents were more worried about their rights being “over-protected” against their will than their privacy being violated (within reason).
I left them with two points to consider:
Anyway, it felt like, after the esoteric discussions of the morning, someone needed to ground the discussion. My fellow presenter Deborah Collyar, of the Patient Advocates in Research, did the same and Sharon Terry of the Genetic Alliance made several suggestions for how to manage different preferences by trial participants for privacy.
The interim report of this committee is due in January of 2014. In the meantime, PPMD, in collaboration with the Muscular Dystrophy Association, and the Critical Path Institute will continue to work toward a private data sharing initiative. If it works, we’ll write the report afterwards.