As I’m sure every NHS manager reading this right now knows – demand from patients in the NHS comes in enormous peaks and troughs, which really hurts performance and budgets. If you plan for peaks you have too many underutilised resources during quiet times and, if you plan for the troughs, you can’t meet the demand targets during busy times.
At the moment, what we see in the NHS, is that very little data is used for forecasting and, the mathematical models deployed are, in my opinion, entirely inappropriate for providing accurate predictions of patient demand.
There is a wealth of data both internal to the NHS and externally – either public and free or available through subscription. Consequently, patients could be better served and costs managed more efficiently by using more information, more sophisticated technologies and better statistical models to more accurately forecast demand, right down to different shift levels.
Right now, there are large ‘systems integration’ projects integrating the systems of primary and secondary care. This, of course, is necessary, but it’s a really traditional approach to IT. In parallel to these integration activities, the NHS should be aggressively deploying data and analytics solutions for patient care and running efficient organisations.
But, it’s not happening. This data isn’t being used. Currently, the deployment of analytics could be best described as tentative or exploratory.
Skills, data sharing laws and funding failure
Data isn’t being used and I’d say this is predominantly due to a lack of internal know-how. Feel free to comment below, if you think otherwise. But, at a national level, there are some good skills in data analytics at, for example, ‘NHS Digital’. [Although, the recent announcement that NHS Digital’s executive director of information and analytics, Professor David Hughes, would be leaving the organisation in August might well set this back].
But, on the whole, analytics skills do not exist within CCGs or secondary care trusts that are staffed for traditional IT. This needs to change.
In the meantime, I think local trusts will need to work with externals – trustworthy and reliable third party suppliers – until this skills gap is filled. I appreciate that will sound bias on my part, but I don’t see any other way of filling the skills gap in the short term.
The secondary issue for lack of uptake of analytics is complex laws around data sharing – and fear of getting it wrong.
Research by the Wellcome Trust last year found that only a third of people felt they had a good understanding of how the NHS uses patient information. People said they were generally comfortable with anonymised data from medical records being used for improving health, provided there is a public benefit, but many were uncomfortable with the idea of companies accessing their health data.
Funding is also an issue. The EPSRC has recently funded five research centres around the UK that will apply mathematics and statistics to prediction models in order to help clinicians tackle health challenges such as cancer, heart disease and antimicrobial resistance. But, this isn’t really going to help NHS managers right now – it’s a long term solution.
Again, in the short term at least, I think trusts will need to work with those providers who can promise business insight and ROI. They need to find a provider which says, ‘if you spend money, we will save you money’.
Do you see other challenges to the successful deployment and usage of analytics by NHS managers? How do see analytics helping? What can suppliers do to support you?