Forthcoming EU guidance could significantly change the way that innovative medical software is approved for use – or not – in the NHS. Key to regulating the industry proportionately is to understand that different risk profiles exist, and any attempt to create a ‘one size fits all’ approach, for the sake of simplicity, should be resisted.
The need for regulation
Health-related software is an increasingly dominant part of our lives – be this a wearable or an app tracking the quality of your sleep. And whether this is a trend you’ve come to loathe or love, it’s hard to dispute that this is also becoming a vital part of the health service too. Increasingly we are seeing the use of software to help inform clinical decisions and even predict deterioration in patients, some of which utilises AI techniques.
All health-related software naturally generates some concern and controversy. What happens if the system, the information, or the decision itself is wrong? Who is to blame? ‘Organic’ human errors are tolerated more than the errors of automated systems (as self-driving cars are demonstrating) and the presence of regulation helps inspire trust in an area in which humans are naturally cautious.
In the UK and the EU, the EU Medical Device Regulation (MDR) lays the foundations for how software for use by clinicians is (or should be) regulated. On top of this, guidance is drafted to support the MDR’s implementation – and it is new draft guidance from the European Commission, which is currently out for comment, that demonstrates regulation is far from easy. So hazy are the definitions underpinning this complex topic that together the MDR and the guidance could change entirely how the industry is regulated – particularly for low risk software – and risk closing the door on healthcare innovations when the NHS needs them most.
Everything with a screen
While the importance of regulation is not under dispute, the what and the how is causing concerns amongst many software companies, including those we work with. Taken together, the new guidance and the MDR may result in a profound shift in burden onto developers, ramping up the risk classifications assigned to what – in some cases – has hitherto been regarded as low risk software.
Under the new guidance seemingly everything with a screen is being thrown into the ring, purely because it has some passing link to software. On a wide reading of the draft guidance, the mere provision of medically related text, if provided to clinicians via a screen, could be subject to stringent regulation; despite it providing no more information than a medical textbook. In fact, potentially far less risk than textbooks, because information provided electronically can be updated far more quickly than information provided through a textbook. The new guidance falls into the trap of assuming that information that is provided through software – whatever the content – is intrinsically of a higher risk than information that is not. This is what risks the future of simple, impactful tech solutions that have real merit.
The need for flexibility
In a climate where, according to the BMJ, 55% of our clinicians are close to burnout and emotional exhaustion, assistive technologies can allow clinicians to focus on the jobs that only they can do. This is particularly key if the ambitious productivity targets that have been set for the NHS have any chance of being met. Regulation and guidance needs to facilitate rather than frustrate the adoption of these technologies.
The MDR may have been passed, but opportunities remain to ensure software developers retain a degree of flexibility:
- As stated above, the European Commission’s guidance on the classification of medical software for the purposes of the MDR is currently being consulted upon – albeit in a closed fashion
- Further revisions to the European Commission’s guidelines on the ‘qualification and classification of standalone software issued in healthcare’ are planned
- The UK’s MHRA is considering publishing its own guidance on the implementation of the MDR if needed
Getting all this guidance right will be key to minimising the regulatory burdens the MDR places on software developers, whilst ensuring the effectiveness and safety of medical device software is retained.
The B word
Beyond this, however – and with the prospect of a ‘no-dealer’ as PM – the MHRA would be forgiven for thinking again about the ways in which divergence from the MDR might accelerate the development and adoption of medical software in the UK. In thinking about this, the MHRA would do well to consider the example of the US’s 21stCentury Cures Act of 2016 which, among other things, lifted low-risk medical software out of the regulatory scope of the Food and Drug Administration (FDA).
The UK political situation may be tumultuous, but in considering a variety of Brexit outcomes we do have the opportunity to radically rethink the way that we support software innovation (even if we ultimately decide to align with the EU system). Thinking radically does not mean abolishing regulation, or even embracing regulation-light, but instead querying where the risk actually arises from. The UK can position itself at the forefront of medical innovation, harnessing the opportunities that arise from being the largest integrated healthcare system in the world. All of this must, of course, be carefully balanced with safety standards and effectiveness, but these things are not mutually exclusive.
It is possible for us to maintain these high standards while freeing low-risk, outcome improving devices from unnecessary regulation; if only we see that not all technology is made equal.