Jan 28

Jan 28 Should Clinical Trial Data Sharing Be a Precondition for Refereed Journal Article Acceptance?

Clinical Trials, Cloud Computing, Collaboration, Data Access, Journals, Open Data, Project Management, Research Data, Research

By Dennis D. McDonald, Ph.D.

To appreciate some of the implications of the recent proposal by the International Committee of Medical Journal Editors (ICMJE) to make consideration of journal article acceptance contingent upon the author’s agreement to share de-identified clinical trial data with other researchers, as reported by NPR, some context is appropriate. Relevant considerations include:

The role of journals and the overall research communication process.
Increasing use of the cloud for data collection and distribution.
Evolving efforts to share clinical trial data.

Journals

Journals do a lot more than just referee and publish research articles. They also serve as proxy gatekeepers for professional career development, and they provide a snapshot in time of progress in the myriad of research areas that exist.

One can argue that they are less important now as communication media than they were once upon a time, given the many formal, informal, and social media channels for collaborating and sharing scientific and medical research.

Regardless of any waning power, however, refereed journals still have an important role to play even as article publication and distribution has moved to the web. The significance of the ICMJE proposal is real as, according to some estimates, only 50% of clinical trial research ever makes its way to publication in refereed journals.

The Cloud

Clinical trial data, much of which has been traditionally shrouded in pharmaceutical company secrecy, is becoming more open as institutional support (for example, see Yale’s YODA project) and commercial infrastructure supporting electronic clinical trial data management (for example, see Pharma Tech Outlook’s top 10 clinical data management solution providers) continue to grow.

ICJME’s proposal occurs at a time when movement of clinical trial data to the cloud for management, access and analysis is the norm, even as questions are raised about the security of such systems (for example, see Big Pharma’s Bet on Big Data Creates Opportunities and Risks).

While it’s impossible to predict how rapidly standards-based clinical data will be universally cloud accessible, steadily increasing digital access seems inevitable, bringing with it improvement in opportunities for sharing and reuse. ICMJE’s support for sharing this appears to be not pioneering as much as part of an overall trend.

Sharing

It’s important to keep ICMJE’s reasons for requiring data sharing in mind:

Sharing data will increase confidence and trust in the conclusions drawn from clinical trials. It will enable the independent confirmation of results, an essential tenet of the scientific process. It will foster the development and testing of new hypotheses. Done well, sharing clinical trial data should also make progress more efficient by making the most of what may be learned from each trial and by avoiding unwarranted repetition. It will help to fulfill our moral obligation to study participants, and we believe it will benefit patients, investigators, sponsors, and society.

That’s a powerful set of goals. They go to the very heart of why research findings need to be shared not only in clinical medical research but in all areas of science. Validation, reproducibility, avoidance of unnecessary duplication are all – in both theory and practice — highly valued professionally. Limiting one’s review of research to only the words, data, equations, and figures displayed in the traditional refereed journal article will certainly hobble the processes of reproducibility and validation, especially if the original research is unavailable or restricted.

Still, when it comes to making research data consistently available, the devil will be in the details. Control of how this sharing will take place and how accomplishment of the above goals is measured will be in the hands of many different parties that do not always share the same goals and objectives. These include professional societies, national and international standards bodies (e.g., CDISC), academic institutions, commercial publishers, cloud vendors, software developers, research funders, pharmaceutical companies, hospitals, and others.

With so many players and with no single governing body or process, it’s likely that clinical data sharing may take some time to become the norm, even as high-profile organizations such as Johnson & Johnson move in that direction. The mechanisms by which data are made discoverable and analyzable will need to be designed and supported by entities that are willing and able to support the costs.

Conclusions

Clinical trials data are generated by an intervention of some sort (for example, mechanical, pharmaceutical, or behavioral) with human beings. Patient safety and the safety of those who might also be treated based on products emerging from clinical research must be paramount. As was shown with the clinical research associated with paroxetine (brand name Paxil) minor variations in how data are captured and coded can have significant impacts on subsequent findings and recommendations.

Assuming privacy and security questions can be addressed, the cost of (a) making raw data available to independent researchers at a level of detail that supports re-analysis of source data, along with (b) an understanding of the processes used to generate the data, eventually need to be understood by everyone involved in the process. We appear to be moving in that direction.