Try to be serious

If a *clinical* system - that is, one which carries active patient records, "goes down" overnight - or any other time - people will die.

I don't know where or what the problem was: this *is* an old US system adapted - somewhat - for the NHS - which is very different from the US with different requirements which have been changed at a bewildering speed since NPfIT was established and a previously free market nationalised - and sold off to the lowest bidder without - I agree - adequate specification.

(does anyone think the OBS was anything other than a vague wishlist?).

IDX was evicted.

I saw Cerner at a Usability session before the contracts were awarded: as it did not (at that time: I haven't seen it since and don't think they *have* a GP module) support Coding in their GP module I was not impressed...

99.9% uptime is not acceptable for life support: what happened to the requirement for backup sites taking over seamlessly?

Hm. Will Patient Choice give us information on the IT used by different hospitals and the performance ratings of those systems?

Why the heck hasn't anyone started with defining standards and a base?

I managed to shove a network under the whole of UK Government a good 12 years ago. It worked so well that AFAIK it has been working seamlessly since, right up to now. How did that get established?

Well:

1 - get everyone with a brain around the table and let them have their say. The clever thing to do is to skim just under the level where it gets political and idiots start dictating things they hear from their friends. This is where you will get your real needs from rather than the bells & whistles politicians want. Some of that is well meant, but you must explain to them that there is a time and place for that.

2 - LEAD, but from knowledge. Consensus is cute but never 100% achievable, and you're better off building a base that supports a wide but controlled set of standards than a "one size fits all" which will never work. If it's your project, you take the decisions. If you can't, tell whoever overrides you to stick the project where the sun doesn't shine and walk. I'm serious here - if it's not your decision you don't want to take the hit when it does what most IT projects have done over the last few years.

3 - OK, for the 1000th time: KEEP IT SIMPLE, STUPID. Stick with variables you can control and establish a foundation and a framework before you get fancy. Here's a start: do all practices and entities have an internet connection? Can they be part of a more enclosed network? Are your standards truly open so you don't end up sponsoring a supplier because your integration gets otherwise screwed?

4 - Stop trying to build mega databases. No, I'm not kidding. Think about the Net. Some bits we can all get to, some parts we can only visit after invitation or by being a member. Maybe there is a better way to get and store information. That would also avoid the "all eggs in one basket issue" all these projects are plagued with.

5 - organise a weekly chat with your gov sponsors. They need to know what is going on, but should not try to influence what works by itself, just support it from the sidelines. Keep them up to date and you'll help each other.

Now *bloody* get on with it.