Why the IDRC won’t rush its open data policy

More and more research funding agencies are mandating open data policies. This is a trend not just in industrialised countries, but in middle-income countries such as Brazil and Colombia too.

From bilateral aid agencies to philanthropic foundations and national research councils, institutions are recognising the potential of open access to advance science, collaboration and innovation.

The International Development Research Centre (IDRC) is no exception. The centre also believes open research data fulfils a fundamental scientific principle of enabling others to confirm or challenge research.

Its Open Data Statement of Principles, launched in November last year, underscores the importance of open research data to advancing the centre’s mandate. It signals to researchers that data management is to be conducted in a manner that maximises the potential for data sharing — but recognises that doing so should not place an undue burden on researchers. It also focuses on responsible sharing of data, outlining legitimate reasons why some research data should not be shared.

The Statement embraces an incremental and learning-based approach to opening research data. Why incremental? The decision to ‘go slow’ comes from an assessment of capacity and culture within and outside the organisation – and others might learn from this experience.

An evolution

But first, let’s look at the bigger picture.

Although research funding agencies are moving towards mandating free access to research data, their policies vary. A key variation is the amount and nature of support ‒ financial or in other forms ‒ that is provided to their grantees. The IDRC’s discussions with some of these agencies have revealed that implementing and monitoring their open data policies can be challenging.

At the same time, publishers too are increasingly requiring authors to make their data open.  A recent study found that 10 per cent of the top 100 social science journals had an open data policy in 2003, and this almost quadrupled in just over 10 years – rising to 39 per cent by 2015. This trend towards more rigorous open data policies is comparable in the life science journals as well.

It’s a move that can bring significant benefits. One powerful example of this comes from the Human Genome Project’s pioneering decision to make genomic sequencing data open. There may be disagreement on how to quantify its economic impact ‒ one study estimated a 141-fold return on the original US$3.8 billion investment over 20 years – but there is broad consensus that this mega-science programme was a significant catalyst for genomics research and the healthcare economy.

Another example comes from India’s Open Source Drug Development (OSDD), a consortium hosted by the Council of Scientific & Industrial Research (CSIR). A recent internal assessment of its work reports how OSDD benefited by pooling research infrastructure and making it open: it reduced the cost of neglected disease research, and focused a previously disparate research effort into a networked community that could then work collaboratively.

One initiative, to re-annotate the genome of the microbe responsible for causing tuberculosis, was completed in four months at zero additional cost as a result. The impact may be more incremental than a new drug discovery, yet it matters to advancing research.

By encouraging and supporting the research community to share their research data, the IDRC expects to see similar dynamics emerge in the research it funds: new collaborations, new directions and utilization of data for economic, social and environmental benefit. This includes addressing gaps that limit research in low- and middle-income settings. Data sharing should also enable researchers in the South to increasingly build on the work of their peers.

Pacing change

The agency’s Open Data Statement follows from the launch of an Open Access policy in 2015. This recognised that publications from publicly funded research should be freely available and openly online, and created flexible mechanisms for grantees to contribute their research to the global pool of knowledge.

The Open Access policy focused on sharing outputs, such as journal articles, books and grey literature from IDRC-funded research activities. While it  encouraged data sharing, this was not mandated.

Since then, and looking to inform its position on open research data, the Centre assessed its internal capacity to support open data and responsible data management. It also scanned other research funders’ policies and practices, consulted grantees, and worked with a cohort of researchers to develop guidance for managing and archiving research data.

The results of that analysis concluded that the time is not yet right for IDRC to pursue a mandatory open data policy that would apply to every research project.

Making research data open is more involved than supporting open access to research publications. To be successful, the management and sharing of open data needs to be integrated into the overall research life cycle. This requires additional resources and attention from research teams. As might be expected, the technical expertise of researchers and the capacity of their home institutions to curate research data varies.

IDRC recognises that some research institutions and grantees will seek additional technical support and will look to IDRC staff for guidance. This expectation informs the incremental approach as both staff and grantees acquire new skills and deepen their understanding of the issues involved. There will need to be informed dialogue, for example, on how to protect human subjects and intellectual property. Another issue is compatibility with privacy laws when sharing research data, or when the nature of the data and the context merit restricting its access.

Without appropriate technical and financial support, an across-the-board policy requiring the development of open data management and sharing plans could easily become a compliance exercise, and fail to change research culture for the better.

Learning from this preliminary work, the IDRC’s new principles make clear that it plans to build on its existing open access policy gradually, and with a learning-based approach. In this way it will continue to contribute to public dialogue and learning on opening research data in developing country contexts.

Over the next three years, and before revisiting its approach and commitments, the IDRC will strengthen the expertise of staff and researchers to promote open access to research data. For now, the immediate focus is on making open the data that is most useful and most likely to be reused, and working with those researchers who support that vision.

Naser Faruqui is Director of IDRC’s Technology and Innovation Program, and chair of the centre’s Open Data Working Group. He can be reached at [email protected] Co-authors Katie Clancy, Daniel Gregoire and David O’Brien are members of the working group.