It’s time to democratise NHS data
- A global Britain has nothing to fear from No Deal - September 25, 2020
- A brewing storm: why the West and its allies must be braced for further Erdoğan adventurism - September 23, 2020
- London’s bridges are falling down. The Hammersmith Bridge disaster is the embodiment of Labour rule - September 17, 2020
BY DR ROB SUTTON
An ageing population with an increasing disease burden, combined with ever more expensive treatments, continues to stretch funding for the NHS, where costs which have far outpaced income growth. For decades, governments have taken the same approach: throw more money at the problem and hope whoever wins the next general election manages to fix it.
The trend of increasing costs is set to continue, so the question of sustainability is vital. There is little political incentive to attempt a top-down reorganisation, and given the complexity of the organisation there is no reason to believe an attempt to impose a one-size-fits-all solution would be successful.
It seems a necessity that we work to improve efficiency within the NHS. Given the enormity of the organisation, there is no shortage of areas where improvements could be found, and the application of data science has been heralded as a potential path the success.
The possible applications are myriad: improving patient care pathways, developing risk stratification and prognostic tools, maximising efficient resource allocation, and managing demand could all benefit from the modern tools of analytics. Machine learning, artificial intelligence and data science will each play a part.
As a junior doctor, a large amount of one’s time is spent transcribing notes from one source to another, organising patient discharges, negotiating when they need to move to a different ward, and chasing test results through a labyrinthine organisation. When you hear doctors complaining about the amount of paperwork they are required to do, this is a large part of it.
With the exponentially growing body of data we need to analyse and manage, this problem will only grow. Many tasks which occupy the time of healthcare professionals are dull, repetitive, unskilled and error-prone; others are of a complexity that is beyond human capacity – in short, exactly the sort of tasks which would be ideally suited to a computer. Software applications capable of assisting in these areas would allow us to vastly increase efficiency and allow healthcare professionals to focus on providing patient care. The NHS is in desperate need of a competitive market from which to buy or build such tools.
Yet organisational resistance, red-tape, arguments about data protection and closed shop practices in who can access that data are preventing tech disruption from making our NHS fit for the 21st century. Many practices taken for granted in our hospitals would be viewed in horror by many in the private sector for their wastefulness. Clunky software interfaces are an embarrassment.
This is not to say there has been no progress. The formation of NHS Digital shows that the power of data science to revolutionise healthcare practise is, at last, being recognised, albeit late in the day. Many hospitals now employ teams of data scientists or contract external analytics providers. More broadly in the current government, the formation of a data science unit in No 10 indicates there is a new seriousness about applying analytic techniques to the public sector.
Yet this isn’t enough. Having this data only accessible to a handful of public sector employees who do not have to face competition, and large analytics firms who are contracted on an ad-hoc basis makes it extremely difficult for small, innovative tech disruptors to enter the market and provide software which could radically improve how we deliver healthcare.
The data available for those fortunate enough to have access is also inadequate. Patient data is often siloed within NHS trusts. This is an absurd approach if we are serious about using it to improve patient outcomes, and is recognised as inadequate by those working in the field. A hospital shouldn’t have to wait for an outbreak before it can analyse patient data on best practice. Data should be shared widely, in an accessible, anonymised and consistent format.
There are dozens of simple applications which, if implemented, would vastly improve the ability of healthcare workers to carry out their jobs more easily and make excellence sustainable. Yet the asymmetry of information between a select few with access and potential software developers means the NHS will perpetually be behind the cutting edge in analytics. As long as data is kept in-house we cannot radically innovate. Time, energy and money will be wasted as we settle for organisational mediocrity.
These oligopolistic practices obstruct market entry and peer-review. They allow unscrutinised analysis to drive national policy, as when Public Health England’s Covid-19 death figures were found to be a likely overestimate. When we prevent innovation, we are denying patients the best possible care and wasting taxpayer money. If we make NHS data more widely accessibly, in a secure and anonymised fashion, outsiders can expose weaknesses and develop improvements.
The obvious concern raised would be the protection of patient confidentiality. According to NHS Digital, “many smaller organisations would like to use more data, but do not have … systems in place to keep patient-level data secure.” This is of course a serious issue, but by no means insurmountable, and it is mostly used as an excuse to stifle competition.
Patient data could easily be anonymised while maintaining sufficient granularity to draw meaningful conclusions. Information governance is waved around as an excuse to not bother with new ideas, despite the technical challenges of ensuring security and anonymity being modest. It prevents us from learning lessons and studying best practices.
The fact that so many tasks which desperately require innovation are allowed to exist within such a key part of our national infrastructure as the NHS is a result of an asymmetry of information distorting the barrier to market for tech disruptors. Big data can generate unrealised value in healthcare, but only if it is made available to those who can compete to show us how to best use it.
By providing a secure interface through which any member of the public, healthcare professional, research group or company can study data and build applications in open competition with each other, our NHS will be stronger, safer and more sustainable.
Rob Sutton is a Junior Doctor in Wales and a former Parliamentary Staffer. He is a recent graduate of the University of Oxford Medical School. Follow him on twitter: @DrRobSutton