Challenges and solutions to sharing health research data
In this blog post, Robert F. Terry, Manager of Research Policy at TDR and Phaikyeong Cheah, Co-ordinator of the Mahidol Oxford Tropical Medicine Research Unit Data Access Committee and Amanda Blatch-Jones, Senior Research Fellow at the NIHR, explore the importance of data sharing, the scepticism surrounding this practice and what needs to happen in order for it to become second nature.
The sharing of all research outputs, including de-identified data, should be a fundamental part of the research cycle. For public health, with its focus on maximizing health benefit for as many as possible, access to a wide range of high-quality data in a timely manner is essential to understand the spread and control of diseases in our increasingly inter-connected world. Of course, this is in a future where we hope to be able to bring together data from genetics, patient records and the wider environment to understand better how, for example, antibiotic resistance develops and spreads.
The situation now is that for many health research areas, optimal sharing of data remains the exception rather than the norm. For example, recent work commissioned by Wellcome and TDR showed that even with respect to emergency public health situations, such as disease outbreaks, only 2 trials out of 58 surveyed provided any link in their registry entry to the background data. Here, data sharing becomes even more critical in order to be able to address the research uncertainty. With the advent of post-publication peer review – such as that enabled by the F1000 publishing model for instance– allows for data to be shared now while publications can follow later –undoubtedly a win-win situation.
Reticence to sharing
Researchers have presented a range of reasons to explain their reluctance to share their data in our various discussion with them. A common explanation is there is a limited academic benefit to them in sharing, plus curating data takes time and money, so beyond a public good argument, they feel there is little incentive to share. Furthermore, for the researchers who work in low-income countries, they have told us that they feel that by sharing their data before they have had a chance to fully analyse them, makes them open to exploitation by researchers in high income settings. To put it bluntly, they feel like data exporters.
The pressures of research assessments are also another limiting factor. For example, in the UK, the Research Excellence Framework places much burden on researchers to seek high impact journals, with limited focus on the implementation and benefits to the public. If assessments were more about the implementation, making better use and re-use of data and collaborations across disciplines, then we may start to see a greater demand for sharing research data in its entirety.
Actions need to come from all directions; the researcher, public, funding organisations, charities, academic institutions and industry. If we really want to reduce and eliminate scepticism then we need to look at changing the culture to support and embrace openness, transparency and inclusivity. If the value and kudos of data sharing are equal to the demands of institutions and/or research assessments then we might begin to see a change outside of academia as well.
Everybody has a responsibility to ensure that data is shared in a robust, safe and efficient way. If we can ensure and demonstrate that data are protected in a secure way, then the boundaries on withholding are weakened. This can be achieved with good data management processes, procedures and well-trained data managers. This includes investments and commitments by research groups, institutions and funders in good data management infrastructure, skills training, data management hardware and software.
There also needs to be proactive activities put in place to engage researchers. For instance, when the IDDO worked with researchers in malaria, tuberculosis and Schistosomiasis, we found that if the researchers from a particular area are brought together to have an input into the design of a new sharing platform, they engage more in the subsequent sharing. One key element is understanding what the research agenda with data will be – can they see the purpose? IDDO has a number of examples of this where research groups have been established to investigate the pool of data and this provides a real example of the scientific benefit sharing the data will bring.
To this end, there needs to be consistency across disciplines and institutions, funders and industry to agree upon and establish a core data sharing policy for all to follow, which can be built on and adapted to fit the relevant type of data. If there are different policies, then there is risk that on one knows which one to follow and there could be a case for not sharing. Indeed, funders have a key role to optimise the value of data sharing and institutions can enable academics to enhance the movement for data sharing to be or part of normal practice.
Would you like to contribute to our Real Impact blog? Find out how.
Recent News & Blogs
- Discover impact news from across the globe
As the new year dawned, the open access landscape was already in full discussion as cOAlition S requested feedback on their proposed framework for transformative journals. Here I explain our position on this framework and our aspirations for the future. At Emerald we are committed to providing a leading service for open research, in a …Read Article
Stay In School campaign: Operationalising the extra medical measures to address the issue of fistula in Northern NigeriaDr. Bankole Allibay
Fistula and Extra-Medical Solution The orthodox response to gender-based sicknesses is scientific medical treatment. Although not incorrect, field experiences show that these treatments are often times reactive, and do not prevent a repeat incidence. Social performance practice recommends proactive social risk diagnosis and proactive risk management procedures, rather than reactive impact management. Such is the …Read Article