Share Trial Data? JNJ Says Yes, Others Hesitate

Should clinical trial data be shared with any researcher who wants to study a medical question? Or does trial data belong to the original investigators?

That's the question that is being debated in the pages of The New England Journal of Medicine (NEJM). The issue was raised by a proposal in early 2016 from the International Committee of Medical Journal Editors (ICMJE) that set out proposed requirements for sharing data. If implemented, the proposal would make trial authors include a plan for data sharing as part of their trial registration and would require authors share deidentified individual-patient data within six months of publication. The ICMJE wrote that they "believe that there is an ethical obligation to responsibly share data generated by interventional clinical trials because participants have put themselves at risk."

Proponents of sharing clinical trial data say the practice will increase transparency and help speed medical advances in drugs, devices, and diagnostics. But some researchers say data sharing could lead to more inaccurate analyses and pull much-needed time and funding away from new trials. 

One major supporter of sharing data is medical giant Johnson & Johnson. As reported by MD+DI last year, Johnson & Johnson collaborates with the Yale University Open Data Access (YODA) Project to share clinical trial data on the drugs, diagnostics, and medical devices. Researchers interested in using that trial data submit a research proposal and if granted access after a blinded review by the YODA Project, they must sign a data use agreement and can access the data remotely through a secure platform. 

At the time, Joanne Waldstreicher, MD, chief medical officer at Johnson & Johnson, wrote in an e-mail to MD+DI, "We hope that in ensuring systematic, objective access to our clinical trial data we will support the overall principles of greater clinical trial data transparency throughout the industry and make possible research that ultimately enhances public health and advances science and medicine."

Waldstreicher and Harlan Krumholz, MD, of Yale, authored one of four opinion columns on the subject in the August 4 issue of NEJM, sharing their collaborative experience as an example of a data sharing platform that works. In a blog post, Waldstreicher wrote, "We believe that granting an independent academic group ultimate authority regarding the release of clinical trial data provides a fair and unbiased approach."

They wrote that in its first year, Johnson & Johnson enabled data from 123 trials to potentially be shared. The YODA Project received 29 requests for data, all of which were approved. In her blog post, Waldstreicher noted that so far, there are two papers in development from this shared data.

To Krumholz and Waldstreicher, the question is not whether trial data should be shared, but how best to do so. They wrote, "The next steps in data sharing will involve leveraging the progress and experience to date to develop consensus standards for implementation, with attention to affordable ways of sharing data sets responsibly and giving credit to the researchers who generated them." 

Data sharing has another well-known advocate, Senator Elizabeth Warren (D-MA). Warren wrote in her NEJM opinion column that the pluses of the practice include new angles for analysis, large-scale analyses across numerous trials, identification of conflicts of interest, and the potential that the proposed requirements could encourage more timely trial reporting. Recent research has shown that although trial results are supposed to be reported to the ClinicalTrials.gov website within one year of completion, a minority of trial results actually are reported in that timeframe. 

On the other end of the spectrum, the researchers of The International Consortium of Investigators for Fairness in Trial Data Sharing took issue with the ICMJE proposal. They wrote in their opinion column that although "at least for large trials, there may be a case for sharing data in an appropriate and timely manner, [but] we do not support the ICMJE proposal as it currently stands." The group wants the proposed six-month post-publication requirement for data sharing to be lengthened considerably, to a minimum two years, with another six months added on for every year the trial took, up to a maximum five years. In addition, they suggested that if a benefit of data sharing is the potential for independent confirmatory analyses, instead, "an independent statistician should have the opportunity to conduct confirmatory analyses before publication of an article." 

One of the consortium's chief concerns is that the "too short" six month requirement would de-incentivize trial coinvestigators and site investigators from taking part in a trial since they usually take months or years to fully publish multiple analyses from one large trial. "Once the investigators who have conducted the trial no longer have exclusive access to the data, they will effectively be competing with people who have not contributed to the substantial efforts and often years of work required to conduct the trial," the consortium authors wrote.

The consortium authors added that they think the ICMJE proposal might actually lead to longer delays in publishing trial results. They also proposed that groups or researchers requesting access to data be required to give financial compensation to the original researchers for time, efforts, and to defray the costs associated with sharing the data.

Warren wrote that the costs of data sharing "can and should be built into the grants, cooperative agreements, and contracts that researchers negotiate with trial sponsors . . . and over the long run, data sharing may help reduce costs by allowing researchers to avoid duplicating trials or to answer questions without undertaking a separate data-collection effort."

Another proposal comes from the opinion written by The Academic Research Organization Consortium for Continuing Evaluation of Scientific Studies--Cardiovascular (ACCESS CV). The group, which shares data from cardiovascular trials, backed the practice of data sharing, but suggested a framework with a different timeline and infrastructure. The ACCESS CV approach would allow researchers interested in a new analysis to submit a request to the tiral's publications committee after the results are published. After a two-year period, these analysis requests would be systematically reviewed by a group of ACCESS CV members not affiliated with the original trial, the original principal investigator, a trial statistician, and one member of the data and safety monitoring group. This 2-year lag would also give the original researchers time to conduct their own analyses, they wrote. Requests and decisions would be posted on an ACCESS CV Web portal within 60 days. In addition, there would be the ability to validate a trial's primary results no later than one year after publication.

"Such a process would avert the problem of multiple groups proposing the same analysis and would require investigators to have prespecified analysis plans," the ACCESS CV authors wrote.

Which approach and proposal--including some collaborations and recommendations not mentioned here--will win out? It seems this debate is still in its early days.

[Image courtesy of PAKORN/FREEDIGITALPHOTOS.NET]