Ever since the first G8 summit on dementia in 2013, the OECD and international community have placed dementia high on the agenda. Elettra Ronchi, Senior Policy Analyst at the OECD, opened the meeting by highlighting that clinical trials are failing and there is little transfer of biomedical innovation to the point of care. The risks and rewards for innovators are simply not aligned, and there are disincentives and barriers to data sharing in dementia research.
Setting the scene in socioeconomic terms, Anders Wimo, Professor in Geriatric General Medicine at the Karolinska Institutet, emphasized the growing worldwide cost of dementia – estimated to be a total of 818 billion US dollars for 2015. To put this into context, if global dementia care was a country, it would be the 18th largest economy in the world.
In this digital age, big data holds much promise for the acceleration of dementia research. Eric Meyer of the Oxford Internet Institute described four existing data sharing initiatives related to dementia – Alzheimer’s Disease Neuroimaging Initiative (ADNI), AddNeuroMed, the UK Biobank and Swedish Brain Power – demonstrating that there is no single way to build and share big data but that “openness by design” helps.
Big data for dementia may come from many different sources, from mobile phones and loyalty cards to electronic medical records. Making better use of data (whether big or small), and sharing it, is likely to advance dementia research, as it opens up more opportunities for insights from both well-established and new or unlikely sources and types of data.
One theme resonating throughout the workshop was the importance of trust, that is: trust from the public in consenting to their data being used responsibly; trust within the research community to re-use data appropriately and give credit where credit is due; and trust in the quality of the data itself. In dementia research especially, there are privacy concerns and legal barriers to sharing patient data, with a growing need for new models of consent to be established.
In dementia research especially, there are privacy concerns and legal barriers to sharing patient data, with a growing need for new models of consent to be established.
Although the benefits of sharing data might seem obvious, a cultural change is needed – researchers who share their data are not generally rewarded for their efforts. We also face a technical challenge, with a need for more training and data scientists in the lab.
Another obstacle discussed was the lack of infrastructure – without it, the sharing and re-use of data in dementia research cannot be adequately governed. Conversely, there seem to be no barriers to not sharing data, which is seen as part of the problem.
Although not yet the norm in dementia research, incentives are starting to emerge. From a funder’s perspective, Laurie Ryan discussed the NIH’s data sharing policy and recently launched Accelerating Medicines Partnership-Alzheimer’s Disease (AMP-AD) initiative. Sustainability, however, appears to be an issue as questions were raised regarding how funders can support data sharing beyond the end of the project and whether there is support for repositories.
We then heard from Julia Wilson of the Global Alliance for Genomics and Health on their interesting efforts to catalyze data sharing with initiatives like the Beacon project, testing the willingness of international sites to share genomic data. This was followed by Mark Thorley providing an overview of CODATA activities, ranging from data citation principles to policies and best practice recommendations.
Publishers can also play a key role in incentivizing data sharing, as outlined by Andrew Hufton, Nature Publishing Group, noting the importance of clear and transparent policies, and the benefits of publishing ‘data papers’ in journals such as Scientific Data, GigaScience and F1000 Research, “rewarding those that share above and beyond existing standards.”
Last but not least, citizen science may have a part to play. Ernst Hafen discussed the plans for MIDATA.COOP, an initiative encouraging citizens to contribute to data integration and personalized health, knowing that their data is being put to good use.
When it comes to data sharing in the neuroscience community, neuroimaging has proven a great place to start.
When it comes to data sharing in the neuroscience community, neuroimaging has proven a great place to start. It has given to rise to some promising initiatives, as David Kennedy (International Neuroinformatics Coordinating Facility) and Arthur Toga (ADNI) discussed. ADNI, in particular, has led to the launch of GAAIN, a portal for Alzheimer’s research data. We also heard about Dementia Platforms UK from Claire Mackay, which, like GAAIN, aims to act as an enabler, making it easier for a governing structure to be implemented rather than forcing people to share data. In a similar fashion, the NACC, led by Walter Kukull, links data from different Alzheimer’s research centers across the US – resulting in more collaborations and publications.
And of course we need only look to the physical sciences to see a successful model of data sharing – Salvatore Mele, Head of Open Access at CERN and Strategic Director for INSPIRE, gave an inspiring talk about the evolution of open access and open data practices in the field.
The next steps
So we know the obstacles, incentives and enablers, but how do we make data sharing in dementia research a reality?
The workshop concluded on three key action points:
- To develop better metrics: what is the impact of data sharing at the level of the individual, institution or nation? What are the costs? How much data sharing occurs? Answering these questions will facilitate a shift in the culture and reward system for data sharing.
- To establish a set of core principles: ensuring data sharing practices are prioritized in IRB and ethical review boards, recognizing and rewarding data sharing, and adopting ORCID.
- To coordinate funding: forming a dementia funders forum, promoting institutional collaboration, incentivizing with challenges and competitions, and supporting a sustained infrastructure.
In order for the above to work, openness, collaboration and trust is crucial. And if successful, I’m optimistic that these efforts will accelerate dementia research and beyond, perhaps to other areas of the biomedical sciences.