The Department for Health and Social Care recently announced the creation of a new dataset. This dataset is - quietly - very important for anyone who cares about adoption of innovation, and new treatment guidance:
“NHS England commits to the development of a local formulary national minimum dataset […] to increase visibility of local variation in the implementation of NICE guidance…” (from the DHSC review of medicines pricing, and the 2024 VPAG scheme)."
We love the idea, and this blog explains why. Back in 2019, we wrote an entire research grant application proposing to do exactly this, in a project called OpenFormulary. We probably don’t want to do that work today, six years later. But we’re happy to help, so we’re sharing our thoughts to inform the discussions. (We’ve also been thinking through - we’ll share more another time - what teams like ours can helpfully do with their many failed funding applications!).
First, for those new to the space, here’s a brief background
A formulary is a list of medicines recommended for prescribing in an organisation or local area.
What do we mean by 'local area'?
“Locally” usually means at Integrated Care Board (ICB) level or hospital — but in practice, formularies can exist in all sorts of NHS organisations. We even know of ICBs that have two different formularies.That sounds simple, but there’s a lot to unpack. Firstly (out of scope for this post) there are the good and bad reasons to have different formularies in different areas in the first place. And after that, there’s a world of complexity. For example, there’s enormous variation in how formularies are produced, how their content is described, and how they are published online. So you can look at them one by one, but if you want data on them all, in one place, you have a problem.
Because of that, there’s very little known about formularies. For example: nobody knows how much they vary between organisations; nobody knows how many there are; and, crucially, nobody knows how successful they are at influencing local treatment choices.
We think formularies probably are a very useful tool in managing the NHS; and that they probably work well in some areas, but not so well in other areas. We think there is a huge opportunity to learn more about them, spread best practice, and improve how the NHS drives change in treatment choices when there are changes in the evidence, the prices, or the clinical guidance.
Our original “Scrapheap Challenge” proposal
Back in 2019, our OpenFormulary proposal set out a pragmatic approach to tackle this:
-
Build an empty national database for local formularies, with a sensible pragmatic data schema, ready to hold the information.
-
Create simple user-friendly tools to help regions update their data, with an offer of them getting back reports describing adherence to the formulary in their local area, as the “carrot” to encourage them to do this.
-
Develop an automated importer, in collaboration with software companies who already provide formulary management tools, to make the work of extracting information easier, and help keep things up-to-date when online formularies change.
-
Build public-facing dashboard tools on OpenPrescribing, showing adherence to (and deviation from) local formularies, on multiple different prescribing measures, so that clinicians, patients and NHS managers can understand local clinical activity.
-
Generate high-quality, open research (by us, and other teams using our tools and data) to describe things like: the extent to which local formularies vary between each other, and on what types of content; the extent to which local prescribing is influenced by the content of local formularies; and factors influencing the speed of change in local clinicians’ treatment choices when the formularies in their area change.
It was a bit of “scrapheap challenge” approach: pragmatic, scrappy, open, and driven by our desire to make useful tools for the NHS via services like OpenPrescribing.
Sharing what we know
We’re delighted to see this topic being picked up again at national level. We suspect the best solution might look a little different now, and there are probably several ways to deliver on this ambition.
Do we want to build OpenFormulary ourselves right now? Probably not. We’re quite busy with things like OpenSAFELY, and OpenPrescribing Hospitals. But we are very happy to share our ideas, the lessons we learned, and our advice on what might work (and what probably won’t). Our teams know a thing or two about NHS data infrastructure, medicines data, and running big open platforms.
So if you’re working on this, or interested in doing so, then get in touch!