The data sharing conundrum in the NHS

In the week where the NHS has been declared "the best, safest and most affordable healthcare system" by the think tank Commonwealth Fund*, one thing springs to mind is how far we've come - and how much work there is still to be done.

Let's take patient data as the topic du jour.

The amount of NHS patient data stored and then silo-ed away by commercial clinical applications is surreal to a certain extent. Few years back, the government mandated that the NHS must digitise all health records and go paperless, enables data sharing across the primary/secondary care sector and more recently a call for interoperability and the ability for patients to access their data online and on the go by the end of the decade. The policies and the funding were made available to help with this transition, the technology to enable it has been around for decades (apart from the new comer FHIR), so why are NHS hospitals still procuring applications that locks their data into a single vendor, essentially rendering it completely silo-ed and impossible/expensive to share with other clinical applications?

How big of an issue is this?

Locking patient data into a silo is clearly a problem, the lack of data sharing however could be disastrous to patient care in the long run, with the advent of AI/ML, data is becoming the real enabler in advancing patient care, it is a real asset that hospitals can leverage to their advantage (see Royal Free and Deep Mind collaboration*).

What's the reason?

Information governance around patient data is absolutely crucial so delays in tech adoption is understandably slow but sharing said data could easily be achievable by adopting a strategic and long term approach during the procurement process and empowering the local IT organisation by putting them at the forefront of that process in consultation with clinicians. Incumbents in the commercial sector have a lot of work to do to make data in their silos available and open, adhering to various standard protocols and integration workflows. Understandably, there is some engagement cost involved to integrate with third parties but this ought to be priced up in advance, reasonable and not passed on to the hospital. Most crucially however, technology should be the enabler here to reduce those integration costs. The feedback I receive when I talk to decision makers in various NHS/Private hospitals however was that the technology being advocated by the majority of big commercial vendors falls short of the data sharing spirit. Take VNA as an example; despite the name, VNA by itself is just another silo-ed, Radiology centric system which doesn't play well with others. To mitigate this, some companies have implemented full IHE profiles as an add-on, but all of this still sits on a legacy system that requires propriety data storage and retrieval, thus making for an exorbitant TCO. Diagnostic imaging should be part of the holistic patient record and accessed via EPR (Electronic Patient Record) like any other critical patient data such as GP referral letters and test results.

What's the solution?

• Making interoperability a mandatory requirement for future procurement in ALL departmental areas, from a simple glucose monitor to a fully-fledged PACS system. Some are even calling for this to be a legal requirement.
• AICA (Application Independent Clinical Archive) is gaining a lot of traction lately; some companies have done excellent work in this field.
• A self-regulated body of commercial vendors which ensures all members adhere to the data sharing spirit and put data sharing at the forefront of their product offering.

From my personal experience however, we need buoyant competition in this field, smaller, nimbler companies that disrupts the industry and adopts a modern approach to software design. This will reduce cost and widen the choices available to NHS Hospitals.