We need some substantial changes in the way we think about healthcare information technology in Wales.
I believe that we need to, figuratively speaking, pull the Andon cord and call a stop to the status quo, to stop building significant technical debt and move away from long procurement cycles and success judged by the deployment of different products from different vendors in different organisations.
The Welsh Audit Office (WAO) report strongly criticises the governance arrangements and the imbalanced reporting of success and progress. Our architecture is becoming ever more tightly coupled with, mostly bespoke, point-to-point integration. With each expansion or integration, the whole enterprise from end-to-end becomes ever more brittle and less predictable. We have increasing problems with reliability in critical national systems.
Paradoxically of course, this is itself extremely predictable.
A simple single answer?
Unfortunately, there is no simple single answer to our problems.
The fundamental steps now are to recognise that there are problems and that we need to improve and to be open and transparent about the issues we face.
Unfortunately, the responses at the recent public accounts committee meetings suggest that
- perhaps not all recognise that there are problems
- that there is a lack of ambition to enable transformational change using digital technology
- there is a lack of priority on the delivery of benefit and understanding of how to do that effectively.
I think the solution will, given the NHS is a complex and adaptive environment with lots of moving parts, lots of variation and significant complexity, to look at improving things both from the top-down and the bottom-up.
Top-down : cohesive principles and credible plans
From the top, we need cohesive, credible plans based on broad design and technology principles and standards.
While we have the “Informed Health and Care” strategy, it is only a vision of what is needed and we must ask, where is the the specific executable plan to deliver that vision? There are those of us who are now trying to draft that, but where has it been up until now?
This needs to be a thoughtful and informed piece of work, noting the successes and failures in a non-judgmental way and highlighting the areas of priority, as well as reflections and robust options-appraisals of different options. If it can’t do that, then the top-down delivery plan must create a tangible plan to understand those options to inform the balancing judgement.
It necessarily follows that this work must adopt the broad principles of openness and transparency in which our work as a service is open to inspection and evaluation. As a result, adopting the GDS design principles (https://www.gov.uk/guidance/government-design-principles) has been a good first step, acknowledging that we do not have all of the answers is the second, understanding that we can already define some good broad principles that will enable us to deliver the third, and being open to outside help the fourth! For the latter, we need to bring in expertise on digital transformation and adopt agile methodologies and user interface and needs expertise.
What must follow from this is the creation of an overarching strategic group who is tasked with setting broad directions as well as the principles and standards to be used by any team working to deliver functionality for users across Wales, including professionals and patients. Those overarching principles and technical approach must include a credible plan for a cohesive architecture, making the best use of existing infrastructure and yet permitting delivery of solutions to both our immediate problems, and not limiting our ability to solve tasks we must face in the future.
Bottom-up: empowerment, cooperation and trust
For improvements from the bottom-up, we need to change the narrative and, most importantly the culture, from command-and-control to collaboration and cooperation. We need to focus our resources on that work only the public sector can do and make information technology the enabler for change by the service itself. This means that we provide a toolbox of data and computing services and applications from which the service can stitch together solutions to their problems at a local level and share those solutions between organisations when they are successful. This means that we aim towards an open platform approach in which data is the key and application programming interfaces provide access to important shared software services.
In essence, our platform becomes the servant of the health service so it can work more effectively.
As such, we need a shared platform of discrete modular loosely-coupled components available to use for both internal and external teams. Those components might include citizen and professional identity services, laboratory, radiology and other investigation services, document repositories and a messaging fabric, but critically, they must be open, documented, standards-based and secured with appropriate authentication, authorisation and verifiable logging at the service (API) level. Loose-coupling means that interactions between services and the teams that work on those services are limited, documented and form essentially a service contract. Importantly, this approach means the different components of the architecture can be improved and developed independently; critical in speeding up delivery.
What therefore must follow from this is the need for a national platform delivery group, made up of cross-functional teams bringing together expertise from different specialties such as management, development, testing, quality assurance, safety and operations. These individual teams must be given responsibility to deliver discrete components of a modular, loosely-coupled architecture. For example, these services underpin important outcomes in many different workstreams - citizen identity clear underpins both “Information for you” and “Supporting professionals”, the workstreams supporting patients and professionals respectively.
Our open platform needs robust API lifecycle management to build APIs with a focus on ease of use for application developers (whether internal or external), consistent security policies, automated testing and deployment, documentation and analytics to gain better insights into API usage and performance. Those APIs need to be used by both internal and external developers alike.
Together with these technical changes, we need to build trust and collaboration between organisations. This means shared source code, teams made up of individuals from across Wales, irrespective of their organisations, focused on delivering their specific goals.
In essence, information technology needs to be owned by the service working collaboratively and in partnership, and bringing in commercial expertise and solutions when necessary.
Most businesses recognise that IT does not simply support their business, but it is the way that they do business and we need to do that for health. Such recognition has meant that digital leaders are at board level and that information technology should not be considered a separate department. In Wales, much IT is outsourced to a separate poorly governed organisation.
Striking a balance
There is, of course, a balance and none of what we need to do is particularly easy, but I struggle with the paradox of Wales attempting to run hugely ambitious, top-down, multi-million pound IT projects given both local and international experience of the likelihood of failure vs a genuine lack of ambition to use digital tools to transform how we care for our population and enable an ecosystem of innovation. I personally find decision-making to be opaque and rarely find published appraisals of different options explaining the advantages and disadvantages of any particular option.
It is quite easy to make excuses and blame lack of delivery on how difficult it is to effect change, but I would argue that our culture, our governance and our technical architecture conspire to make our work even more difficult. We should be aiming to stack the odds in our favour!
There are people around who want to make a difference and recognise the need for change. There are pockets of good work at all levels, including government, NWIS and the health boards. I think for many of us it feels as if we are swimming against the current however. It is also important to recognise that there has been historic under-funding compared to peers in England, but it is difficult to justify more expenditure when there has been a consistent historic failure to prioritise and to deliver using the resources we have had.
I would suggest that we need to:
- Perform a national architecture review to help define a modern, standards-based loosely-coupled evolutionary adaptive architecture.
- Change our metrics so that we focus on user need and delivery of benefit, rather than deployment of a specific product. Our metrics should reflect our focus: user need.
- Change governance and ownership. Give the health boards ownership and control, ask them to form a partnership with each other and government tasked to set top-down principles and technical and information standards and provide expertise and oversight to define a national open platform.
- Change delivery - by empowering developer teams whether national, local or commercial to use that open platform to solve problems, and ensure that those solutions can be built collaboratively and shared and scaled when shown to be safe and successful.
- Change development practices - with real agile working, cross-functional empowered teams made up of different specialists working together to deliver.
- Change our behaviour and culture - become open by default, make Wales a beacon of safe, secure healthcare informatics underpinning service transformation and improvement - and teach everyone else how we did it, the difficulties we faced and the strategies we used to overcome those difficulties.
It is possible to have a single organisation acting as both overseer, strategist and software supplier, but that needs a change in behaviour towards transparency so that decision-making is open, subject to external scrutiny and there are clear appraisals of different options. In my view, the NHS Wales Informatics Service fails to behave in this way. My preference is to change the behaviour, but if that is not possible, then we must surely consider re-organisation instead.
I have written previously on many aspects of this. These are the most important bits:
On open platforms in healthcare here:
On governance and top-down vs bottom up here.
On platforms here…