Wednesday 21 December 2011

The extra costs of the big IT systems

My book The Human Element came out last month and it has been fascinating how people have responded – I got a Facebook message last week from a doctor in New Hampshire explaining how the fake efficiency being peddled by managers at her local hospital was adding to costs.

Then sometimes I run slap bang against new evidence myself. Because I hadn’t until last week used the new NHS Choose and Book appointment system.

The first I knew that I was using it was a letter asking me why I hadn’t called them up to book my appointment with a consultant. I hadn’t in fact received the original letter from my doctors, but didn’t know that at the time.

I called their appointments line. They couldn’t talk to me because I didn’t have a password.

Now I’ve recently been refusing to accept passwords. I have vast numbers of them already for every public and private agency I deal with. It means I have to keep them in a notebook.

Far from making my identity more secure, it is actually making it less so – like the phenomenon of people putting their daily password on a post-it note on their office computers.

So I said I wasn’t accepting any more passwords. Long conversation with the manager who eventually put the phone down on me. Stand off. They said I had to get back to my GP. My GP’s receptionist said I had to go back to the appointments line.
I felt excitingly embattled, but my resolve crumbled when I got my GP on the phone. I don’t want to make her life any more difficult, after all, so I meekly accepted my password.

It is (surely you’re not going to publish your password? I am, really, it’s no use to me) ‘estate tomato’.

The magic words finally got me through the appointment system phone line. They gave me another number to call, which got me through to King’s College Hospital. There a very nice lady said they would send me an appointment date in the old way.

I asked why that required a password like estate tomato and four phone calls. She said that, in practice, the new Choose and Book appointment system – imposed at vast expense by one of the ubiquitous IT consultancies – was so useless that they had opted out.

Why was it so useless? Because as students of the systems thinker John Seddon will know, these centralised IT solutions can’t deal with diversity. They are inflexible and therefore lock in costs.

Was there really a problem of people not being able to prove their identity under the old system? People stealing each other’s hospital appointments? I don’t believe so, yet the IT consultants have clearly made out that there was.

Maybe you need extra safeguards if you have a huge national system. It is another way that big scale solutions lock in costs.
So here we are, another way in which the legacy of New Labour still infects Whitehall. A more expensive system which works less well than it did before because it rules out informal solutions – very urgent appointments, anything which requires face to face judgement.
Inflexible solutions means higher costs. No wonder the deficit is rising.

What can we do? Well, we can refuse to accept any more passwords, but – I have to admit – I failed on this score at the first hurdle. I’ll do better next time.


Michael Gradwell said...

A very interesting post. My Trust trialed this wonderful knew computer system and I was recently told by one receptionist (now safely retired) that the system was so slow in giving out appointments that they started timing them and it took around ten minutes! Some people walked away. Talk about false efficiency.

Edis said...

Every Liberal should know about the concept of the 'Law of Requisite Variety', one of the historic fundamentals of 'systems thinking'.

As Vladimir Nabukov (reputedly) said 'life spills over the edge of every cup'

Richard Gadsden said...

The great strength of the internet is that it is "loosely coupled". It's lots of small systems with standard connections between them, and the connection standardises the minimum amount.

This leaves each system with the maximum amount of flexibility to suit local circumstances.

One reason why the Internet's TCP/IP stack prevailed over the OSI stack was that it was so much simpler, because it didn't try to deal with as much stuff, so each manufacturer could add on a TCP/IP stack to their computers without having to change fundamental aspects of the way their computers worked so they worked the way OSI expected them to.

What the NHS should have done was work out standard interchanges for data between systems; then each hospital and each GP practice could choose their own preferred systems and capabilities, but the systems would all be able to transfer data to one another.

Take "choose and book". Why not just define two standard interfaces:

1: Query a hospital for availability of a particular type of appointment.
2: Create an appointment for a patient at a particular time and date.

Then each GP system can query all available hospitals for availability, make a decision according to the programmed criteria (which might be some clever automated combination of cheapest, soonest and best, or might be as simple as "ask the GP" or "ask the patient") and then book the appointment.

Create the minimum central standardisation necessary and then let each separate organisation decide what to do.