The International Committee of Medical Journal Editors (ICMJE) recently put forth a proposed set of new requirements for sharing data that was generated by interventional clinical trials. The ICJME believes there is an ethical obligation to responsibly share such data because the participants in the trials put themselves at risk.
Essentially the ICMJE is proposing that as a condition of consideration for publication of a clinical trial report in their member journals, the authors must share with others the deidentified individual patient data (IPD) that is underlying the results presented in the article, including any tables, figures, appendices, and other supplementary material, no later than six months after publication. This proposed requirement will include all data underlying the results of the article’s findings, as well as any necessary metadata.
As you can imagine, there are strong opinions on both sides of this proposal. Those who are arguing for it, claim, “many funders around the world – foundations, government agencies, and industry – now mandate data sharing.” You know, that whole, “everyone else is doing it, we should too,” mentality that our parents warned us about when we were younger.
Those who are against it have a multitude of opinions and reasons for being against it. Some go so far as to refer to those who support data sharing as “data parasites,” since they latch onto research that has already been painstakingly performed and utilize it for their own purposes.
Analysis
This new proposed rule may sound like a nice idea, having the ability to reexamine high-quality information for the possibility of new information being found, potentially resulting in higher patient satisfaction and longevity. However, as just about anyone who has ever managed clinical studies, performed data collection and analysis, or curated data sets knows, there are a litany of concerns over such a proposal. Dan L. Longo, M.D., and Jeffrey M. Drazen, M.D., penned an editorial laying out some of their concerns from that perspective.
One such concern is that someone who is not involved in the generation and the collection of the data will not understand the choices the researchers made in defining the parameters. Some specific questions raised by Longo and Drazen included, “How heterogeneous were the study populations? Were the eligibility criteria the same? Can it be assumed that the differences in study populations, data collection and analysis, and treatments, both protocol-specified and unspecified, can be ignored?”
A second, very valid, concern, is that an entirely new class of research person will emerge – someone who had nothing to do with the design and the execution of the study, but use another group’s hard-earned data for their own ends. These “stealers of data” can then use the data to steal research productivity planned by the data gatherers, or even use the stolen data to disprove the original researchers analysis.
Data sharing may not be all bad, depending on how it is performed and what requirements are in place. Longo and Drazen posit, for example, that if data sharing were to work symbiotically, with collaborators whose collected data might be useful in assessing your hypothesis, it might be beneficial for all parties involved, including patients. Throughout the symbiotic relationship, the two (or more) teams of researchers work together to test a hypothesis and report new findings with coauthorship, acknowledging the group that proposed the new idea and the investigative group that pursued the data and allowed it to be tested.
It is interesting to note that four days after jointly submitting an editorial piece with Dan Longo, M.D., Jeffrey Drazen, M.D., walked back part of it. He clarified that the New England Journal of Medicine, the forum for the initial editorial, is “committed to data sharing in the setting of clinical trials.” He went on to comment that he believes “there is a moral obligation to the people who volunteer to participate in these trials to ensure that their data are widely and responsibly used” and that “researchers who analyze data collected by others can substantially improve human health.”
In the walk back, he concludes by once again bringing up data sharing through collaboration, signaling that such form of data sharing may be palatable to more than just the “data socialists” who want to take your hard-researched data a mere six months after the publication of your findings.
Saurabh Jha, a radiologist in Philadelphia, summed it up nicely,
It takes a lot of effort to generate data in biomedical sciences. To expect researchers to surrender the data for the greater good is fuzzy, and lamentably boring, adolescent naivety. If we do not recognize the self-interest of researchers, data socialism, like other forms of socialism, is condemned to failure.
If you would like to provide feedback on the ICMJE proposal, you may submit your comments and concerns to the International Committee of Medical Journal Editors by April 18, 2016.
For More Background David Shaywitz, MD has a nice series of articles in Forbes on this topic:
Data Scientists = Research Parasites?
Biden Cancer Project: An Opportunity To Implement Data Sharing Incentives
Do We Really Want To Separate Clinical Data Gathering And Data Analysis?