Unexpectedly, it seemed I was preaching mostly to the converted when I spoke at the recent US Institute of Medicine Workshop on Sharing Clinical Trial Data. I was given the job of the opening “inspirational” talk which was to emphasise the benefits of sharing data from clinical trials. The meeting planners anticipated that later speakers might focus heavily on the potential problems and pitfalls of data sharing. The hope was that starting off with an upbeat elaboration of the benefits would provide a useful counterpoint to any doubt and pessimism that might follow. As it turned out, though, most of the speakers and attendees at the meeting did not seem to need much convincing. Everyone professed a commitment to the idea that sharing clinical trial data is a Good Thing. Most of the talks and much of the discussion focussed on the specifics: what to share, who to share it with, and how to do it.

The devil, of course, is in the particulars, but it was heartening to learn that for nearly every potential problem a solution has been invented or at least imagined. One excuse that’s off the table: the claim that there is nowhere to store data. Online services like Synapse and Dryad make that argument obsolete. A speaker from Medtronic described the time and expense the company invested in de-identifying a dataset for independent researchers as part of the exciting Yoda project. Subsequent speakers, however, challenged the idea that de-identification will always be expensive and difficult. Andrew Vickers, a research methodologist at Memorial Sloan Kettering Cancer Center, said de-identification of a large dataset from a trial he supervised was a matter of a few hours spent writing appropriate software code. One novel solution: instead of sending massive amounts of data to researchers, datasets could be standardised and configured so that researchers could “send the question to the data.”

Many of the talks made it obvious that if researchers and companies are not prepared to share clinical trial and other research data, then patients are poised to take matters into their own hands and “the door will be kicked open.” Privacy advocate Deven McGraw discussed “patient facing models of data collection, delivery, and access.” She sees great promise in that approach since large organizations and companies often are bound by regulatory commitments—think HIPAA and the common rule—that limit their ability to re-use and disclose information. She also noted that “if we are struggling with notions of who owns data we are starting in the wrong place” since “ownership is not conducive to a sharing world.”  She said a better concept is shared rights and responsibilities, and noted that “even if data are owned other people may still have rights to it.”

Patients and potential research participants who do not like the idea of participating in trials if data are owned and controlled by a for-profit company or academic institution will increasingly be able to bypass these entities and give their data directly to researchers or aggregating sites that in turn make data available to scientists. But will they actually do so? In fact, such projects are already underway. John Wilbanks of Sage Bionetworks described how he obtained his genotype through an online company and then syndicated the results through openSNP, a company that makes genetic information uploaded by individuals available to interested scientists.

Within days he received emails from researchers around the world. From one he learned that he was unlikely to go bald. Another told him there was “no suggestion of consanguinity” in his pedigree. “I’m from Tennessee,” he said, “so that was good news.” He acknowledged the substantial risks in this approach since most of it occurs outside of any regulated system. “I would rather see the health system doing this for me but I’m not getting the service from the health system to help me as an individual…Things will exist to serve me whether they come from the system or not.” Many other speakers discussed similar projects, all of which aimed to facilitate the move from a situation where research participants feel like they are “sending information into a black hole” to a world in which people control their own medical data and can reasonably expect to get something back from researchers. This, one speaker predicted, will “begin to accelerate the system exponentially.”

In my introductory talk I told the audience that in seeking to liberate clinical trial data we were engaged in one of the important struggles of human knowledge. I was confident, I said, that the benefits of sharing clinical trial data would greatly outweigh the harms. At the end of this historic meeting, I still felt the same way.

Elizabeth Loder is US research editor, BMJ