Visions for IT in the NHS, with various levels of ambition, have come and gone.
Now more than ever, it is in our interests to make digital transformation successful – it is vital for the service, for patients and tax payers.
I applaud the honesty in Matt Hancock’s Tech Vision where the document says, “We don’t have all the answers – this should be the beginning of an open conversation about how we can iterate to best achieve what is needed and work with the many brilliant, forward-thinking people in the system to get it right.” After all, we’d agree that the problems with digital technology over the last 15 years are far more about execution than vision.
Digital transformation is certainly a challenge and not only from a technical aspect. In my previous blog I mentioned hearing Graham Evans, CTO at North Tees and Hartlepool NHS Foundation Trust discussing his experience in the NHS and with digital transformation. Looking to other industries he highlighted that the challenges to successful transformation are really 10% technical/informatics and 90% everything else (people, process, culture and organisation).
Failing fast, learning quickly
Make no mistake, there are huge challenges in “the 10%” technical work that must now be overcome quickly, especially when it comes to interoperability. This is integration of the next generation. Our standards and systems must negotiate security matters, and access and return information in a structured way so it can be reasoned on for intelligent display and clinical decision support.
With far more demanded, far more can go wrong – and if we are going to fail, we need to fail fast. That means doing something agile, based on solving a particular clinical need or problem; that way we can learn quickly. We look to do something along these lines:
Defining a standard information model
We must first define a standard information model for the problem. Whilst this should be done for a locality, it must be fit to work nationally. That is why, in England, doing this with the Professional Records Standards Body (PRSB) is so important. When talking about medications, allergies or encounters, we must be able to ensure we are all talking about the same thing.
A computable format
The next challenge is to implement the model using associated standards including HL7 FHIR, Dictionary of Medicines and Devices (dm+d) and SNOMED, into a computable format to allow data to be exchanged using standards APIs and documents that have to be defined.
Right now, in 2018, we still have information in patient records on drugs that is free text, so the standards need to have a minimum structured data set recognising the reality out in the service. Over time, this must change as these standards will allow NHS and social care systems to classify and code diagnoses, procedures, events and medications. All systems must be able to accept these standards so that we can make sure information flows at scale across the population without different APIs being developed for different systems.
Test harnesses and hackathons
Setting up of test harnesses and running hackathons for the implementation of the information model (using FHIR profiles and APIs) is key to starting to identify what works – and importantly our first chance to fail – fail fast and fail early.
Clinical validation and assurance of implementation made by suppliers
You might have defined a specification where you expect technology to behave in a certain way, but once you build it, you need to understand if it works in reality, and whether clinicians can work with it. Everything we do must be clinically driven and technically informed. This is not just about the technology, standards and data models delivering against a specification – it’s about clinical acceptance, validation and assurance.
Technical implementation of interoperability in a specific care landscape with specific systems
Having done hackathons – where the model has been proven in a test environment, we need to take that next step to try it out in the user environment. We must identify if it works in the complexity of the real world and that means planning early to identify first of types and actually use it for real.
Achieving transformation with limited resources
Outside of the technical challenge, transformation is especially difficult in the NHS which is under financial pressure and where the workforce is already overloaded. But the NHS no longer has a choice – so how can we be successful?
Work with the willing
Organisation leaders and clinicians have to be embedded in digital decision making and implementation. However, start with those people who are prepared to work together on the transformation journey and they will win over the hearts and minds of others to embrace and lead change. Typically, success breeds success which often means the slow adopters will come along in due course.
Understand process change
Implementing healthcare technology has a profound impact on processes of care. By listening to staff, mapping out their workflows, and identifying their pain points and gain creators, we can provide them with tools that make a real difference.
Develop a realistic roadmap
Your integration roadmap must be clinically driven and prioritised, tempered with what is achievable. This means starting small and focussing on high value data that can be shared with systems that have some kinds of API that supports the exchange of data.
The fact is that most staff doing the day to day work know what the issues are. By allowing them to contribute solutions and have them added to roadmap or even implemented immediately if they can be, will increase the value and relevance of solutions and the buy-in of staff.
So, whilst there is a lot of technical work that has to be done to define standards, there needs to be considerable effort to effect the necessary process and cultural change. This is not something that can be imposed “top-down”, but rather collaboratively with co-produced deliverables embracing all stakeholders, especially those at “the sharp-end”.