Everything is interoperability

In a system as complicated as the NHS, pretty much all work is collaborative. Getting seemingly simple things done usually means that at least three or four teams are required. In a way, this is really nice: to move forward, we all need to rely on each other. A shared sense of purpose develops. It’s also really frustrating: to move forward, we need to agree on a unified approach, but the system is vast and we don’t all have the same incentives. Much of what we are working on now needs to deal with past decisions that were made under different regimes and different pressures. The wider system wasn’t built in one go – it has accreted over time like so many layers of of sediment. Supposedly, this is the right way to do things. Inevitably, we find places where the intentions and objectives of our predecessors point in a direction that is counter to our own.

We are currently working on several interconnected problems that all hinge on data that describes who someone’s doctor is. A year ago, I did not imagine that this would be a complicated piece of information. I was wrong. How that information is recorded, structured, stored, and accessed has implications for at least four current projects. From here, we’re trying to work out how to thread a needle of user expectations and clinical safety and the reality of our datastores in a way that settles multiple challenges.

This isn’t an uncommon problem or something no one is thinking about. Quite the opposite: I can’t count the number of articles I’ve read about “fixing the plumbing” in the last few years. The question right now is “how will we approach the problem?” It is easy to slip back into spending the bulk of your time thinking about technology instead of remaining focussed on users and what they need to do. What is strange is that everyone seems to know that we tend toward this reversion, and yet it still happens. Perhaps this is because the dimensions of the problem are easier to understand and work on when framed this way, but there are methods for grappling with complex problems. We’re back to service design and systems thinking. Nice to know I won’t be obsolete any time soon lol.

I’m reminded of something Megha said to me a few weeks ago: “Most service design is mostly solving interoperability problems”. This framing feels particularly relevant to working in the NHS. Compared with other contexts I’ve worked in, connecting the miscellaneous parts and layers of the system here is an order of magnitude more difficult. “Make your service interoperable” is one of the points of the NHS Service Standard for good reason! In this particular instance, the concept of interoperability goes beyond data sharing. I’d use it to also describe how teams work together and how concepts are defined. We need to also ensure the humans systems have interoperable (i.e. shared, matching) models in their heads before we can solve the data and service design challenges. It turns out that part is much harder than you’d expect.

Some links:

• Everything is tuberculosis. I listen to audiobooks of pop-science books from time to time (it works like a 10 hour podcast), and this is a great one. It is both a scientific and cultural history of tuberculosis that I would call “a real page turner” if I had read it on paper. The thing that is most interesting about this to me is that it is a great example of the biopsychosocial model that my colleague Lia Ali talks about – yes, tuberculosis is a serious medical condition, but it is also a set of social relations and a mental construct. The best parts of the book are about how the disease divides communities (because of stigma) and shapes people’s perception of self (because of same). Obviously I stole the title of this post from here.
• Letter to Arc members, 2025. My favourite web browser has now been officially declared undead in favour of an AI pivot. I really love using Arc but Dia (the new project) just seems like a Chrome plugin that Google is bound to Sherlock within the next six months. I have no idea how a web browser was ever supposed to be a profitable project and ultimately Arc is yet another story about venture capital and why we can’t have nice things.
• Beyond user needs: Why we need an expanded design philosophy for the digital public sector. I still might write a longer post about seams and seamful design but maybe you should just watch this video by Richard Pope instead. We’re currently testing out a few design changes in the NHS App that embody these ideas.
• Patient data could power the NHS. Much of it is still stuck on paper. The FT with a very clear breakdown of why improving the use of data in the NHS is hard.
• Designed with care: Creating trauma-informed content. Miriam Vaswani (one of the contributors) worked with us on the NHS App for a while and was central to me incorporating trauma-informed principles into our practice. I’ve dipped into a few chapters and so far, so excellent.
• Vibe coding, fireworks and the mortar of government. Benjamin Welby on needing to raise the level of digital maturity in government, in general, possibly via AI, instead of it being the role of specific digital professionals. A great turn of phrase: “digital isn’t a brick in the policy wall of government; it’s the mortar that binds the whole thing together.”
• James’ monthnote for May. I’m also playing about with Copilot (because it is there, because it is the one we have at work) and can confirm that once it is turned on it is everywhere in MS apps. I can also confirm that this is really annoying.
• Problems worth solving, with Rachel Dunscombe. A podcast from Sam Menter (who now runs Healthia, f.k.a. Mace & Menter). This episode on electronic health records is an excellent description of some of the challenges the proposed Single Patient Record will need to address.
• A single patient record for England. Gary McAllister with a proposed approach to developing a Single Patient Record.
• Why is quality so rare? Karri Saarinen (CEO of Linear) on quality and craft, something I have a hard time describing, though lately I have been trying via a set of design principles we’re working on to extend the NHS set.
• Dave McKenna insists they have ham trucks In France. This is the level of group-chat madness I aspire to.