The Discoverability Challenge – How Can We Make Research Data Easier to Find and Use?

Enhancing the discoverability of public health and epidemiology research data is a key to ensuring that it gets more widely used. This was the topic of a recent workshop hosted by the London School of Hygiene and Tropical Medicine, where researchers and data experts explored the findings of a recent Wellcome Trust report on data discoverability. Dave Carr, Policy Adviser at the Wellcome Trust, highlights some of the key themes that emerged from a lively and productive debate.

The Wellcome Trust is committed to ensuring that the data outputs generated by the research we fund can be accessed and used in a way that maximises the health and societal benefit. We are a member of the Public Health Research Data Forum, which brings together a consortium of like-minded funding global funders with a shared vision of increasing the availability of health research data, in ways that are equitable, ethical and efficient.

While some research disciplines have well-established community-level resources that store and curate datasets and make these available to potential users, this type of infrastructure has been much slower to develop in other fields.

For public health and epidemiology research, the vast and rich datasets collected from human populations in the course of research are often held locally by the groups that have gathered the data. In many cases, there is no easy way for potential users to find out whether a particular dataset exists, let alone to gain access to the data in a useable form. This severely limits the potential value that may be derived.

The need to make data more readily discoverable to users is widely recognised as one of the fundamental barriers to more effective data sharing. On behalf of the Public Health Research Data Forum, the Wellcome Trust commissioned an expert team to explore how the public health and epidemiology field could best take on this challenge. Their report was published in July 2014.

The workshop gave participants the opportunity to discuss the team’s findings with a panel of invited experts – including Arofan Gregory of the Open Data Foundation, Steve Kern of the Bill and Melinda Gates Foundation, David Leon of the London School of Hygiene and Tropical Medicine, Brian Hole of Ubiquity Press, and Matthew Wollard of the UK Data Service. The panel discussion was chaired by Jimmy Whitworth, Head of Population Sciences at the Wellcome Trust.

One of the important messages was that making data discoverable is certainly not an impossible challenge. The required technology already exists and successful approaches have been applied in related fields (such as the social sciences). These could be built upon and adapted to provide workable solutions for public health and epidemiology research data.

Panellists and delegates highlighted several pioneering initiatives that are already widening access to research datasets. But the picture that currently exists is a fragmented one, with a lack of overarching community standards and agreed best practice across the piece. A key challenge for any initiative to enhance data discoverability would therefore be linking up existing initiatives and effectively building on what is already in place.

Two other messages emerged very strongly from the discussion. Firstly, that data usability is at least as big a challenge as discoverability, and as it is inextricably linked, cannot be ignored.

The adoption of robust approaches for collecting adequate metadata is critical for both discoverability and usability. Metadata must be gathered as the research is conducted (and not added in as an afterthought).

Secondly, in order to encourage researchers to put in the effort required to make their data discoverable and useable, appropriate incentive structures must exist. This point resonates strongly with the findings of a report of Expert Advisory Group on Data Access on incentives and culture change for data sharing that was published earlier this year.

In building support in the research community, it was emphasised that the principal argument must always be that enhancing the discoverability and usability of data enables better science.

Over the weeks ahead the Trust and our partners in the Forum will crystallise plans for taking forward the report’s recommendations. We are committed to ensuring we progress in a way that best meets the community’s needs, and want the issues to be discussed and debated as widely as possible. This workshop provided an excellent first step in this progress, and we would welcome your further feedback and comments.

The workshop was filmed and is available to view on the London School of Hygiene and Tropical Medicine’s Vimeo Channel. The report, “Enhancing Discoverability of Public Health and Epidemiology Research Data” is available from the Public Health Research Data Forum site.

Image credit: Lederle Laboratory - G. Terry Sharrer, Ph.d. National Museum Of American History/NIH, Carrot on a string – by nist6dh on Flickr CC-BY-SA