So I built an electronic patient record system that’s been running since 2007. It was, as far as I am aware, the first implementation of SNOMED CT in a live clinical system. Most users have never known that SNOMED is its lingua franca, and most think it is a bespoke ‘database’ for their use, because I built multiple subject and specialty-specific “prisms” through which to see, what is essentially, a single record.

There is no better iterative fast feedback loop than being a user as well as a developer. It is difficult to know what works and what doesn’t work without an exploratory phase of design and implementation.

That software now supports the care of thousands of patients across Wales, over 126000 lines of code and on today’s estimates would cost $6 million to develop.

But as with any large software application, it’s starting to become more difficult to make changes. Early design decisions accelerated progress but are unduly limiting now. Early assumptions about how and where the application might be used are now proven wrong.

Such issues are common in all software projects.

In NHS Wales, we have made amazing progress on some core national services solving difficult problems; we now have services for shared clinical documents, investigation results such as laboratory and radiology as well as a national staff directory and an enterprise master patient index providing a single cross-organisation view of patient identity.

But the clinical application that staff use to access data in those services - Welsh Clinical Portal - will never be an electronic patient record. Early design decisions have bound it tightly to the local patient administrative system in each organisation and tight-coupling means that changes are difficult to make, and some health boards have requests for change that have waited years to be implemented.

It is time we modernised the way we think about and develop health and care technology.

One of the most effective ways of doing that is to recognise the benefits of modularisation, in which we break apart the problems we face to make them easier to solve and recompose those modules in interesting ways in order to solve problems for professionals and patients.

Understanding how and when to modularise requires significant technical and domain knowledge taking into account existing legacy services and the commercial market. We can use open technical standards to decouple those components. This is one way that we can pay down some of the technical debt built up over years.

The NHS Wales Informatics Service is being transformed into a new statutory body on 1st April 2021.

Now is the ideal opportunity for it to become the digital platform on which we build a seamless, effective and safe health and care service for Wales.

It can and should be a provider of national shared services - much like Amazon has “Amazon Web Services” - with small, long-lived teams of experts providing a range of data and software services as well as support for provider organisations to make use of those services to improve their services.

Its mission cannot simply be “driving change” in the health service - there’s plenty of evidence that health boards are desperate to improve services using digital technology - it must be focusing on the work that can and should be shared across health and care in Wales - and working to enable change by the provision of a suite of open data and computing services.