How do we realise the health benefits of data sharing?
A new article published in The BMJ outlines experiences and investigates the advantages and costs of different data sharing models.
Over the past decade the concept of data sharing has transitioned from being a relatively abstract idea to a recommended practice within the field of medical sciences. However, the sharing of individual-level patient data from clinical trials or epidemiological surveillance still remains relatively closed when compared to other areas of medical science such as genomics.
The benefits of data sharing are well understood, yet there is still much concern around: ‘data grabbing’ by researchers from well-resourced organisations, issues of patient consent and breaches of confidentiality, inadequate infrastructure to safely store and reuse the data and inadequate guidance on how to share data. These issues may make researchers, especially those with limited resources, more hesitant to make their data widely available to the research community.
An analysis article published in The BMJ outlines the authors’ experiences of data sharing and investigates the benefits and costs of different models. The analysis, ‘Beyond open data: realising the health benefits of sharing data’, summarises discussion at a workshop held at the 2016 Geneva Health Forum and organised by the Infectious Diseases Data Observatory (IDDO). The workshop brought together representatives from academia, industry, global health organisations, non-governmental organisations, funders and medical publishers, to examine the premise that data sharing is good for health. By unravelling the evidence behind the concerns around data sharing, the workshop participants sought to determine what is needed to make data sharing work more effectively.
“We need to think more strategically about how we make data more useful for public health, rather than just championing data sharing without due thought for where it might be most beneficial,” said Prof Osman Sankoh, Director of INDEPTH and co-author of the article.
Prof Souleymane Mboup, project coordinator for the West African Network against AIDS, TB and Malaria, and co-author of the article, added: “We need to work out which platforms and technological structures can be shared across diseases and which diseases would benefit most from data sharing. Likely candidates are emerging infections about which little is known, diseases such as malaria where progress is hampered by factors affecting drug efficacy such as resistance to existing medicines, co-morbidities, drug-drug interactions, and pharmacology determinants, and neglected tropical diseases with sparse data and small sample sizes globally.”
The overall consensus at the meeting was that different models of data sharing hold value in different ways, increasing transparency of research and enhancing public health gains. A variety of different platforms and models for data sharing in global health have emerged in recent years, but it is not enough to share data by depositing it in a repository; protocols and processes for these governance structures need to be shared along with scientific questions and ideas to make data sharing more effective. There is a strong need for funding to support long-term infrastructure and data curation to avoid creating ‘data cemeteries’.
The group discussed the concept of ‘research parasites’ - researchers from well-resourced institutions who use data from data generators to publish in high impact journals with little or no acknowledgement for the primary data contributors. The establishment of more equitable and meaningful collaborations with researchers, in particular those from low- and middle-income countries, increases the potential for public health benefit and reduces the perception and risk of ‘data grabbing’. Strong governance structures help to reduce the risk of breaching patient confidentiality and help to ensure that the investment of the researchers who conducted the studies is appropriately recognised.
“We need to work collaboratively at every stage of the process to maximise the benefits of data sharing, and ultimately improve patient outcomes,” said Prof Philippe Guérin, Director of IDDO, and co-author of the article. “The goal is not only to assemble retrospective data but to provide a wide range of tools and resources that support researchers in enhancing prospective data collection and analysis during clinical trials. We have demonstrated the feasibility of such a model for malaria through the WorldWide Antimalarial Resistance Network, and believe that this model can be applied effectively to other diseases.”
Elizabeth Pisani, Peter Aaby, Gabrielle J Breugelmans, David Carr, Trish Groves, Michelle Helinski, Dorcas Kamuya, Steven Kern, Katherine Littler, Vicki Marsh, Souleymane Mboup, Laura Merson, Osman Sankoh, Micaela Serafini, Martin Schneider, Vreni Schoenenberger, Philippe J Guerin. Beyond open data: realising the health benefits of sharing data BMJ 2016; 355:i5295 doi: http://dx.doi.org/10.1136/bmj.i5295 (Published 10 October 2016)