Users? or Constituencies?
When I was at Caradigm (a Microsoft, GE joint venture), we implemented among other products a Single Sign-On (SSO) solution, one login for many apps. This was not technically a 'clinical' application, it was used in other industries, but it saved doctors hours every day. Its implementation involved working mainly with IT. Maybe there would be some clinical 'champions', but you could work behind IT's existing service desk. You didn't need to interact with the end users. It was much simpler. Easier to implement.
When I was at Interneuron, the most challenging implementation we did was our Electronic Prescribing and Medicines Administration (EPMA) product. This was much harder. It involved working with IT, doctors, nurses, physios, pharmacists and hospital executives, all with their own and different objectives.
As a junior doctor, I witnessed the implementation of the much more ambitious National Programme for IT. It aimed to revolutionise how the NHS worked, and it would touch every part of it - everyone would be involved (whether they liked it or not). I can still recall my genuine excitement about the possibilities.
I was the 'junior doctor representative', went to meetings, workshops, briefings. The programme promised many things, but in the end all that was essentially delivered was digital radiology, replacing the silver halide films I no longer had to carry around. The rest of the programme ultimately failed to deliver the comprehensive care records system it had set out to build.
At first glance, the difference between these implementations could be attributed to the number of users involved. That's how vendors often charge for their software. You can pay per seat, per user. IT will naturally think this way, most of the industry does. The more users, the harder the implementation. But let's challenge that. The other major variable in these examples is not just the number of users, it's the number of different user groups.
You can define a hospital system as a collection of different constituencies that is designed to work in silos. You have 'the clinicians'. This group includes doctors, nurses, physiotherapists, allied healthcare professionals, etc. They care about their patients, the problems that resulted in them being in hospital and how to diagnose and treat them.
Then there are the managers. They may reflect the departmental specialties of the hospital, Orthopaedics, Cardiology, etc., or the clinical services the hospital provides, Care of the Elderly, Paediatrics, Emergency admissions. These are operational people, not clinical. They worry about waiting lists, referral to treatment times, average length of stay, patient flow and hospital budgets.
Then there is IT. Traditionally interested in infrastructure, desktops, hardware and networks (tin & string). This is a difficult constituency to recruit for. You aren't just competing with other hospitals, you are competing with other industries. The IT director at your hospital could just as easily be working at Morrisons as they could Moorfield's.
All these different groups report to the executives. The managers, to operations, IT to finance, the clinicians to the medical director. Hospitals are organised this way for good reasons. The constituencies have different goals. Clinicians focus on patients, managers focus on flow, IT focuses on infrastructure and security. But they share the overall purpose of running a hospital, and so report through necessarily different management structures to the board. The system works well, by design, until someone tries to change something across constituencies.
I was working in a hospital as a surgical trainee, during the planning for a new treatment centre. There was tension between the managers and the clinicians. The managers had decided that two surgical specialties were basically the same, the juniors should be able to manage both groups of patients. It would save money and time, the juniors were not specialists, it would simply be more efficient.
The clinicians were horrified. It took years to train juniors to recognise the serious post-operative issues that require urgent action. Standards would fall, patients would be harmed.
I managed to resolve the conflict by walking with the managers around the new building. There were two wards, side by side, one for each specialty. But the builders had put in doors that connected the wards.
My proposal was that if we opened the doors between the wards near the entrance, we could keep all the (high turnover) day-case patients towards the front, the juniors could cross cover each other for these simple cases. The complex, longer-staying patients for each specialty would be moved to the back, and the doors between them locked. There would be no cross cover of these patients.
Everyone was happy with the compromise. Both objectives were achieved.
Implementations are easier when they target a single user group. This was the main reason why the national programme was able to only deliver the Radiology solution - it was primarily aimed at a single constituency. But when they cross constituencies, things get much harder.
So when an implementation is not going well, ask yourself, have we really understood our constituencies?