¡Ay, Lorenzo!

By Danny Bluestone on November 20th, 2012

Nations, industries and companies invest billions in digitising their services as a way of saving money. That was the plan at the NHS, but the organisation has ended up spending £7 billion on an incomplete, ineffective IT solution.

The NHS National Programme for IT was initially commissioned in 2002. The aim of the project was to give healthcare professionals, such as doctors and pharmacists, the ability to access patients’ NHS medical files from anywhere in the UK. So, if a patient who lived in London came down with the flu on their weekend away in Oxford, any doctor, hospital or GP would be able to access the patient’s entire medical history, to see, for instance, if they have an allergy to a certain antibiotic.

A great idea in theory, but in practice the project, nicknamed “Lorenzo”, has proven to be remarkably unsuccessful. Over the last 10 years, the scheme has been hit with constant delays and uncertainty, making it a difficult project to manage. This has resulted in just 10 of the 166 trusts from the north, Midlands and east of England receiving a basic system.

Lorenzo has since been sold on to an American IT services provider CSC, who are continuing the project.

As far back as 2007, it was easy to see the flaws in the project. South Norfolk MP Richard Bacon cited a number of reasons for previous IT flops in public sector organisations like the NHS. These included contracts being signed too quickly, without the government having full knowledge of what they wanted to achieve; IT companies not fully understanding what they were commissioned to produce and stakeholders (in this instance, clinicians, doctors and nurses) not being consulted about what they actually want from the system.

These shortcomings point to the fact that a “user-centred design” (UCD) approach to IT projects in the public sector would lead to far better outcomes.  As this approach involves a representative amount of all relevant stakeholder and peer involvement from the planning phases of the project, each project would be much more likely to achieve its goals.

According to a healthcare report in Information Week, one of the biggest complaints to come out of the disastrous NHS project was the Department of Health’s reluctance to engage clinicians in the decision-making. Yet clinicians are a key target audience of the system. For that system to be useful, it should be designed with clinicians’ requirements in mind.

If medical professionals were engaged from the beginning of the project, then the software could have been built with a clear understanding of the end user, their tasks and the environment they work in, rather than IT professionals trying to second-guess what they believe to be the best technical approach, interface or even clinical process.

Newcastle upon Tyne’s NHS Foundation Trust Hospitals were part of an independent scheme using a different solution called Cerner Millennium, which was successfully launched in 2009. The system is now being used by all medical staff in the local area, with all the project directors agreeing that “the input and co-operation of hospital staff have been essential to the successful introduction of its Cerner information system”.

Neil Watson, Director of Pharmacy and Medicines Management at the trust, says: “Clinicians from all backgrounds have recognised that this is not an IT project and that to make this work it requires clinical buy-in. No matter how much you push from an IT point of view it has to be about clinical change.”

This “UCD approach” entails producing a bird’s eye view, called a schema, and plan of all the target personas including how they engage with each other and through what interactions (ethnographical studies can be useful at this point).

After the user requirements are clarified, application and flow of control diagrams are agreed with all stakeholders. The main part of the strategic phase is the production of interactive wireframes which are demonstrated to all the target audiences (in this case all departments in the NHS, including all specialties) to get detailed feedback on interfaces, processes and workflows.

Another downfall of the NHS project is that it hasn’t been managed with what we in the industry call an “agile” mindset. By adopting the scrum development method, which is an iterative (or agile) approach, the project could have been built and released in manageable sprints that could have been tested one at a time.

With hindsight, the NHS could have benefited from an Agile UCD approach, where IT professionals would consult the clinicians and various other stakeholders on the design and development of the software.

The IT companies involved could then continually change and fine tune the system as it released bite-sized chunks (each undergoing user testing), rather than designing the software using a Waterfall based model and testing everything at the very end of the project and then realising that things are not working.

It is apparent that medical professionals were not even informed how this system would help them do their jobs. By using a project management tool with various privilege levels, all stakeholders could have been informed of the goals, progress and milestones.

National level stakeholder interviews would have given the IT companies an understanding of the type of tasks clinicians would need to perform.  We would also have recommended a research phase upfront to see how other countries tackle these sorts of projects.

The NHS continue to face the challenges of integrating all the UK’s medical files into one system. If and when they achieve it, it will be the largest collation of civil data in the world, and would save money for the National Health Service.

So it is frustrating to see these mistakes being made both here, and in other public sector IT projects, when there is a viable solution to managing a project of this size and scope. The answer, of course, is a solid approach applying user-centred design and agile development processes.

Danny Bluestone is managing director of digital agency Cyber-Duck