‘Appy in the NHS?

A question was asked from the floor at DH’s National Information Board in September:

“The Government has banned apps that replicate websites. What does that mean for the NHS?”

I don’t have an answer, but spending some time in this space, here’s some thoughts on apps used by the NHS itself for direct care (ie, not considering those for patients or secondary uses).

For a start, there are various security critical tools which need OS-level verification. The HSCIC demo’ed an app accessing the NHS Spine using an iOS version with the equivalent security of a smartcard.

Things locked to an NHS smartcard are probably candidates only for rich device-locked apps (due to the security work, the fingerprint is a nice touch) and probably shouldn’t be web apps. Accessing PDS – the NHS’s ‘reverse phone book’, used to look your details up by your NHS number – via a web app from any device on the planet seems unwise.

Another example of apps given was the acclaimed “Mersey Burns” app, which automates the process of working out the details of fluids to give to burns victims (which varies by location, scale, severity of burns and the person’s physical characteristics). It’s basically a bit of maths with some sliders for degrees, and pictures of a body for locations.

Does that have to be an app?

With the advent of a good, standard platform in almost ubiquitous use, and new communications technologies, it may seem wise to use this as the app platform for the foreseeable future.

But there, it’s the year 2000 and I was actually talking about was XP and IE6 – the effects of which are still taking a significant amount of resource to clean up, 15+ years on.

Following GDS’s articulation of Loosemore’s Logic, which has held true even 2+ years on – despite a lot of attempts and failures to prove it wrong – why is the burns app an app? It’s not doing anything that you can’t do in a browser, it’s not processing directly identifying information and there’s no need to transmit it anywhere, nor save it beyond a page reload.

As a “web app”, it would do pretty much the same thing interface-wise, but equally, there would be the option of running some variant of RCTs (effectively impossible in a deployed app) at large scale, and iteratively evolving the tool over time, and learning easily.

It is difficult to tell clinicians to always update their apps every time, if they don’t know what those updates will do, nor how to reverse them. If you have a deployed install base, everyone needs to initially install the software. (Need to borrow someone else’s tablet? Hope they have all the same apps as yours in the same place; which might work for those who work in only one place, but the locums will have problems.) If there’s an urgent update, how do people know?

Many such apps – which, to restate, shouldn’t be handling patient records – could be web tools, linked from NHS.uk for clinicians; there doesn’t have to be one right answer, and there can be several “burns apps” because there’s no one right way to do certain things.

It also solves the problem of rich patient apps interacting with each other, probably badly or not at all, to do particular tasks (such as a patient record app recording burns information) since those are all done through a webview, rather than you liking one company’s app in one area, but not liking their inferior other tool in another.

Small pieces, loosely joined, is what makes the no-apps vision work: being of the web, not just ‘on the web’.

Unfortunately, the business model for small pieces loosely joined is probably one that is in the interests of the clinicians and experts, though it is clearly not the favoured business model of those who like building apps for devices and charging ongoing fees for them. Or some NHS Institutions.

We’ll see what NHS bodies do, and more importantly, are asked to do.


(Note: while I sit on the National Information Board as a lay member, my involvement in the NHS apps work is only what is in the public papers. As a result, all information here is in the public domain, and I have no internal knowledge of what the NHS is doing next on apps of any kind.)

(Since it always keeps coming up; for now, almost all the tablets will likely be iOS/iPads; but that is simply due to the quality of offerings; with full web apps, a shift to an alternate platform is simple, and can be driven by (clinician) demand.)

posted: 24 Sep 2015

Blockchains on a Catapult

Several sectors, including Government, are putting significant effort into investigating “authoritative registers you can trust”, or what are sometimes called “distributed ledgers” – the main implementation technology for which are blockchains. Even the BBC has made a radio programme about them.

While widespread awareness of blockchains is relatively new, they are a mature design, building on a well-established theoretical underpinning. The main reason for the variety of terms currently in use to describe blockchains is that officials and high-ups are somewhat nervous about referring to the most famous example to date: the digital currency bitcoin.

The general adoption of blockchains promises to fundamentally transform how data is stored and used. The UK model for encouraging and promoting such transformations is the “Catapult” network. It is puzzling therefore that, despite their clear applicability in a number of areas of national interest and a growing acknowledgement of the need to regain the missing trust in data, blockchains do not appear to have been deemed a priority. Read more…

posted: 14 Sep 2015