Skip to main content

When it comes to data and resources, to share is divine (and now required)

Dr. Robin Barr
Robin BARR,
Director, DEA,
Division of Extramural Activities (DEA)
.

Sharing. It is the basic element of socialization. We share feelings. We write to communicate. We talk to each other. Sharing is intrinsically human, and we see it in lots of other species, too. Why, then, is data sharing from our research sometimes so problematic?

As context, I offer a personal anecdote from my early days as a researcher. My colleague and I struggled mightily to publish a particular set of experiments. Some reviewers were on our side and wanted a little more evidence for the phenomena we had observed. Others were hostile, since the findings threatened the assumptions driving their research. These reviewers even questioned our data, asking, “How could you obtain results like that?”

When the paper was accepted, we decided to publish our raw data with the paper. It led to the longest paper ever published in the journal (there was no Internet then!). We shared it because people did not believe our interpretation. Our aim was to allow anyone to analyze the data to try to find an interpretation consistent with their assumptions. It was a good reason to share.

New guidelines for sharing data and resources

I bring this up because NIA has newly published guidelines for sharing data and resources. The central idea is that, going forward, we will require a sharing plan for both data and resources from most applications that we receive.

The new guidelines allow for considerable flexibility in most aspects of the sharing plan. For example, you can write the expected costs of sharing into the competing budget. The guidelines discuss privacy concerns and consent issues and acknowledge that truly open sharing is not always the best option—or may be beyond the ability of the investigator team to allow. We hope you find reasons to share other than “NIA made me do it,” and that our new policy will serve as a prompt to stimulate that human tendency to share.

NIH as a whole is also working on updating its 2003 data-sharing policy. An RFI for other views on what NIH should do to update the current policy received many responses and revealed the spread of norms across different fields. In some areas, such as genetics and survey research, sharing standards are well established. In fact, sharing clearly advances these fields. In others, such as clinical trials, legislation had to be introduced to make the results available because too many results were buried in hard drives somewhere and not even publications showed the outcomes. In Europe, at the moment, sharing is complicated by the European Union’s new General Data Protection Regulation (GDPR) law, which tightens privacy protections on data.

It is simpler then for NIA to shape our own guidelines for now. More than ever, the science of aging is a growing field. Data- and resource-sharing are important stimulants for that growth. We should capitalize on our investments in primary research to achieve fresh bounty from them.

What do you think? Please share your thoughts with a comment below and help us spread the word about our important new data-sharing policy!

Comments

Submitted by Peg Graham on August 21, 2019

I come to the NIA via the SBIR pathway and hope that the NIA will continue to provide companies with a means of shielding proprietary information in applications from public disclosure. Having said that, I fully support sharing, especially across disciplines. During my 30+ years in healthcare, I have observed the "cost" of inward-looking practices and policies, casting doubt and shunning the views of those whose credentials are different. Particularly stunning when there is so much noise about "patient-centered" - yet, no one really is listening. Time for change.