Public health and technology leaders convened in Ottawa at CityAge’s 10th annual Data Effect event last week to discuss data stewardship and sharing as the key to foster innovation and create an integrated public health data ecosystem in Canada.
Chief scientific officer at Ottawa-based Bruyere Research Institute, Kumanan Wilson kicked off the discussion by highlighting the potential of data to advance research and implementation of new health technology.
The development of the mRNA vaccine “in an unbelievably short period of time”, with “basically lines of code” is one such advancement that trailblazed during the pandemic, Wilson explained. The latest Omicron specific versions of the vaccine, for instance, were released prior to phase three trial, and hence needed a new data system to not only test its safety and effectiveness, but also to conduct post-market surveillance with ongoing real world data. Canada, he added, “has the best data” and “can do this better than any other country in the world”.
For Christopher Allison, director general and chief data officer at Public Health Agency of Canada, “there’s a need to take advantage of a terrible crisis to move things forward”. He stressed the importance of favoring an ecosystem and open science approach that links across different sectors to maximize the value of data available on the market.
All involved in health care, including health care providers, provinces, territories, municipalities or indigenous communities need data to play their part. The goal, Allison said, is to “go from the thousands of siloed, fractured health systems, to that vision, where the patient, the human being, is at the center of health care, and we’re providing effective services that can actually work across all of this.”
Data stewardship is different from data governance or management, Allison stated. “Stewardship means you do need to protect it, you do need to make sure it is ethical, you need to make sure it’s not doing any harm as a base layer. But then you also need to make sure that the right people have it, so that they can do their work.”
Therefore, an ecosystem approach also means bringing experts from different areas, including technologists, scientists, epidemiologists, and policy people together to explore the subject and solve issues around the standardization of data, for instance, Allison affirmed. People at senior levels of organizations “devote their lives to one thing” and often “do not have a deep knowledge of what these things actually mean”. Data stewardship encompasses multiple sectors and they all need to take part in that conversation to fully exploit the potential of data.
Privacy policies are another hindrance to data sharing. Senior vice president and general manager, Canada Research Chair in Medical AI, Replica Analytics and University of Ottawa, Khaled El Emam pointed out that access to data should be broader than just to academia and government, and encompass startups and other companies where innovation is also taking place.
“So many reports have been written about this problem. And they all pretty much say the same thing. They describe the same problem and prescribe the same set of solutions. I think we really need to start acting on these recommendations, we shouldn’t be writing more reports.” lamented El Emam.
One of El Emam’s recommendations is looking to other jurisdictions. He gave the example of the CPRD (Clinical Practice Research Datalink) model in the UK, which is essentially the commercialization of the NHS dataset, allowing researchers to pay for anonymized data. He claimed that Canadian analysts have to resort to these very datasets from the UK or the U.S. where it is readily available, instead of using local data.
The General Data Protection Regulation (GDPR), El Emam added, as a standard is a good regulation to follow, but has made data access and sharing very challenging in Europe. “So we have to be cautious also about who we are emulating and who we are learning from, and take the good bits, but not necessarily the bits that didn’t quite work as well.”
The panelists provided a few more recommendations to enable broader data access and sharing, including leveraging synthetic data to set up pipelines and showcase potential, mandating data sharing targets, or creating incentives for large data custodians and educating them on what is allowed and what’s not.