Knowledge for Growth is Belgium’s largest life sciences conference, and with a mix of thought leaders from across the ecosystem, always has a finger on the pulse of current topics. This year, the event focused heavily on data science, with a range of keynote talks and sessions dedicated to the important topic. One of the afternoon panels in particular, Understanding Data, held a lot of insights from Liesbet Peeters (Prof. Biomedical Data Sciences at Hasselt University) and Bart Vannieuwenhuyse (Data Sciences Lead at Janssen Benelux Campus).
This article by Amy LeBlanc was part of a series on data science in the lead up to Science for health 2021, where Liesbet Peeters and Bart Vannieuwenhuyse took part in the panel.
Data sharing is possible
“I believe that our healthcare system of today is just not good enough.” Liesbet Peeters said, in opening the panel session. “It urgently needs to be disrupted with insights gathered from big data. More specifically, from real-world data, which is any data gathered within the context of clinical care.”
Peeters went on to share a touching personal story of a PhD student named Glen who developed Multiple sclerosis (MS). While working on a data science project, Peeters grew frustrated when it took five years for the research consortium simply to find, access and reuse pre-existing data to gain insights on the disorder.
“That’s not good enough,” Peeters reiterated. “During that time, Glen deteriorated from a perfectly healthy badminton player to someone who had trouble finding a job.”
“If the sociological and technical barriers drop, it is possible to reuse data. We just have to be willing to do so.” – Liesbet Peeters, UHasselt
When the pandemic hit, many people with MS had urgent questions that couldn’t be addressed due to a lack of data. In response, Peeters led a global data sharing initiative to find answers as soon as possible. The resulting collaboration between the MS Data Alliance (a global nonprofit collaboration that Peeters is chairing) and the MS International Federation ended up involving about 80 countries from around the globe. What’s more: the initiative was set up in record time – instead of the usual five years, this global data-sharing network was up and running in just one month!
“It can be done,” Peeters says. “If the sociological and technical barriers drop, it is possible to reuse data. We just have to be willing to do so.”
Peeters emphasized that data shouldn’t be open – it should be FAIR: Findable, Accessible, Interoperable and Reusable. She stressed that we need to invest much more time and energy into helping people who understand the value of data and are willing to work with it but lack the expertise to do so.
Data quality needs improving
About 70% of healthcare information is currently stored in non-structured information holders, where it needs to be extracted before it can be used. If we want to start using routine care data for research purposes, the first step will have to be an improvement in data quality, says Bart Vannieuwenhuyse:
“Of course, even better would be to have structured data entry at source, but that’s a story that has been going on for more than 30 years and it’s still not being done. So, first focus: improving data quality. The second step is to create networks of larger datasets. Because the data of a single hospital are insufficient: we don’t have enough volume to look at specific subsections, to have controls, etc.”
Read this article to learn how an Israeli healthcare provider is unlocking health data for research insights.
To obtain data volumes of a high enough quality, granularity, and richness to be useful for research, Vannieuwenhuyse says we need to think about how we can create a network of datasets. This then leads to a range of technological questions such as: how are we going to do the standardization, and how will the governance work?
According to Peeters, there are many solutions for these challenges, but we need to overcome the sociological and political issues first:
“Technically, it’s possible to share data,” she says. “But how do you convince a multi-stakeholder ecosystem to be a community of trust and transparency, that facilitates and accelerates bona fide research at the same time as it protects citizens from abuse? There’s currently no platform for the debate. We manage to organize the individual stakeholders – academics, pharma, regulatory, healthcare professionals – but we haven’t brought them together. We really have to invest time in this debate and empathize better with each other’s issues, concerns, and return-on-investment models.”
Decisions need to be made
“Data saves lives, and we should value data as we value everything else related to healthcare.” – Liesbet Peeters, UHasselt
One of the key stakeholders on the topic of health data is, of course, the government. When asked about its role in Belgium, Peeters was clear:
“First of all, the government has to take its responsibility to facilitate, coordinate and moderate the multi-stakeholder debate. There are urgent and critical questions that are not being addressed (for example: what Common Data Model we will be using?). We really have to make sure that we make these decisions as a multi-stakeholder ecosystem. Then, Belgium should prepare to start participating in the European Health Data Space. We’re all cheering for it, and it will happen soon, but to the best of my knowledge there’s no Belgian roadmap to prepare ourselves.”
Vannieuwenhuyse agreed, and underlined the need for financial resources to drive these initiatives:
“The Netherlands, the UK, Germany, Finland, France; all of the surrounding countries around Belgium are way ahead of us,” he says. “We have a lot of catching up to do which will not happen by itself.”
Read this article to find out how Belgium compares with other EU countries for data science.
Peeters’ take-home message was emphatic:
“Data saves lives, and we should value data as we value everything else related to healthcare. To gain important and urgent insights, it is crucial that we improve the basics of data science in Belgium.”
