We’re just getting used to it, but will healthcare data become federated in the future?
An overwhelming trend in 2018 will be increasing restrictions around releasing healthcare data for analysis, just as we begin to connect with it for deeper insights.
This has obvious implications for industry. Maintaining timely access to robust healthcare data is vital in demonstrating value and outcomes, as well as forging partnerships with patients. Real world data – observational data, reflective of routine healthcare delivery – is also the key to establishing the holy grail of outcomes-based pricing.
So, why the concern about releasing healthcare data? The issues are four-fold. Firstly, pooling data from different settings makes it richer, but also more attractive to cyber criminals. Secondly, citizens aren’t generally in favour of agencies combining their data beyond what is necessary, especially since the care.data fiasco.
Thirdly, there is the ongoing apprehension about industry access to patient data. Finally, fear about increased responsibility and fines from breaches of the forthcoming European General Data Protection Regulation isn’t helping. In response, the walls are going up and this has led to researchers conducting analysis in strictly controlled environments, only seeing fake data or not seeing data at all.
Ever since Sir Nick Partridge’s 2014 review of data, held by the Health and Social Care Information Centre (HSCIC), the writing has been on the wall. He recommended ‘That the HSCIC actively pursues a technical solution to allow access to data, without the need to release data out of the HSCIC to external organisations’. In short, researchers will be expected to enter a secure ‘sandbox’ environment to analyse data, instead of it being released.
Consider also the development of the Simulacrum – a ‘synthetic’ oncology dataset modelled to correspond in distribution to the real thing at Public Health England (PHE).
It means researchers query fake data, and only if PHE deems these queries appropriate will it conduct the analysis itself.
This has also shaped thinking behind the Digital Healthcare Data Federation. The phrase ‘data federation’ may sound confusing, but the issue is that European data controllers don’t want to breach social contracts with their citizens by transferring data outside their nations. To enable data analysis across Europe, a federation where each constituent nation retains some autonomy, will respond to requests for data queries, but researchers never get to see the raw data.
It may appear that we retain the same levels of access through these approaches, but there is a stark difference between having appropriately de-identified data transferred into our hands and this new trend. No one wants sensitive patient data to be mishandled, but a lack of visibility may throttle innovation. Industry’s relationship with healthcare is about to become even more complicated, if that could be possible.
The NHS had collected vast amounts of patient data since 1989, before launching its care.data information-sharing facility in 2013. The ‘announcement’, however, was botched as the NHS’s leaflet to every household became lost among other junk mail. care.data was closed in 2016 amid concern from the public about confidential medical records getting in the wrong hands.