Health sector has a 'first mover disadvantage' in data use, says Ben Goldacre
The old ways aren't the best ways when it comes to data infrastructure
The health sector in Europe and North America lags at least a decade behind retail and finance when it comes to digital maturity, said Jennifer Pougnet, Global Data Policy Strategy Lead at pharmaceutical giant Roche, speaking during a panel session at a recent Open Data Institute (ODI) event in London.
It also ranks poorly when it comes to innovative business models, as the initial response to Covid demonstrated.
Because of the pressures on health systems, data infrastructure, which will be increasingly crucial to managing health services in the future, has become a "nice to have, rather than something that's a key enabler in achieving the goals" said Pougnet.
"There's a lack of translation between the technical decisions being made and the health system and business decisions that are being made."
Roche, along with the ODI, Arup, the BBC, Microsoft, Refinitiv and Sopra Steria, is working on a data governance playbook to try to bring the two sides together, and prevent data governance and infrastructure from being sidelined as a purely technical endeavour, as it frequently is today.
First mover disadvantage
The NHS is sitting on 73-years worth of incredibly valuable health data, said Ben Goldacre, doctor, science writer and Professor of Evidence-Based Medicine and director of the Bennett Institute for Applied Data Science at the University of Oxford. However, this data tends to be stuck in silos, held in non-compatible formats, used for individual projects then discarded, without much thought on how to add value by sharing it widely in a secure, anonymised, standard and reproducible fashion.
The health sector has a "first mover disadvantage", according to Goldacre. Back in the 1980s, there was a broad realisation that computers could revolutionise the health sector's sprawling paper-based record keeping systems and governments invested accordingly. The trouble is, the initial systems worked well enough that they have become ossified.
"What happens when you're the first mover is you end up with legacy ways of working," Goldacre said. "You end up with legacy teams and institutions that can unintentionally obstruct innovation, but also by having lots of people with a great deal of expertise for the cutting edge of 20 or 30 years ago, you have a lot of expertise at just about getting by."
The result, he went on, is those 'experts' bulldozing legacy systems through, because they understand them and because they work, after a fashion, but not nearly as well as a modern, automated data management and governance system would. But the people in strategic decision-making roles don't know the difference.
"They can see that something has been delivered, but they often don't have the technical skills to tell the difference between something that could have been done 30 years ago versus something that was an efficient reproducible analytical pipeline where everything's been properly assured, where everything's got great technical documentation, where you could produce 20 slightly different forks of the same thing really quickly and efficiently."
The problem is structural and financial. What's needed, according to Goldacre, is more technical people "with the knowledge necessary to understand the guts of the machine" in executive decision-making roles, plus proper investment in the data infrastructure. "You've got to put a little bit of money into get over the entropy hump of automating your previous bulldozer processes."
Everyone thinks it's a good idea until you explain how they have to change their ways of working
Which is not to underestimate the concerted effort required, said Tomas Sanches Lopez, Director of Data Curation and Open Data Service Directorate at NHS Digital, repeating the need for senior management to have a proper and full understanding of the importance of health data for research, treatment and decision support.
"The thing about data management, data standards, data quality, metadata, and so on is everyone thinks it's a good idea until you explain how they have to change their ways of working and the investment that is required. And then, all of a sudden, it's more important just to get something out of the door."