Data sharing has the ability to strengthen medical research, writes Senator Elizabeth Warren (D-Mass.) in a commentary appearing in the New England Journal of Medicine this week. She further applauds the International Committee of Medical Journal Editors' proposal that would tie publication to the sharing of de-identified data.
"By requiring data sharing as a condition of publication, journals can help synchronize and expand existing data-sharing practices," Warren writes.
The ICMJE proposal, which was published online in January in the Annals of Internal Medicine and other journals, also says that researchers should include a data-sharing plan when they register their clinical trial.
"The ICMJE proposal has the potential to spur the transformation of the research culture so that data sharing becomes the norm. Indeed, if properly implemented, the proposal could promote a culture that maximizes the contributions of patients who volunteer to participate in clinical trials," Yale School of Medicine's Harlan Krumholz and Johnson & Johnson's Joanne Waldstreicher write in their own commentary, which also describes the Yale Open Data Access Project.
In a separate perspective piece, the International Consortium of Investigators for Fairness in Trial Data Sharing argues that before the ICMJE's proposal is enacted there needs to be a discussion of the benefits, risks, and opportunity costs of data sharing. The group, which does not support the ICMJE proposal as it currently is, says the time period the ICMJE guidelines give in which data must be shared — six months after publication — is too short and will undermine researchers' ability to publish secondary articles based on additional analyses of their dataset. Instead, the group suggests a minimum of a two-year, and up to five-year depending on the trial's length, exclusive use time period for the primary researchers.
In addition, the International Consortium of Investigators for Fairness in Trial Data Sharing says independent statisticians should be able to review analyses prior to publication and that any researcher who was not involved in an initial trial but who wants to access to the resulting data should pay for that access.