![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
₤14 billion over budget
May. 31st, 2006 07:16 amThe British NHS' project to create a national electronic patient record system is ₤14 billion over budget and over two years behind schedule. My guess is that is a highly optimistic assessment. The project is the antithesis, writ large, of everything I think I know about systems development in general and health care systems in particular. Let's consider a few obvious problems:
1. It's monolithic, one size fits all. The assumption is that the same system will meet the needs of GPs, emergency rooms and the full spectrum of specialists. This is just nonsense. The ways of working, data needs and presentation layer needs vary quite widely. Even within a single clinical area like oncology the needs of different agents are significantly different.
2. It's unnecessarily comprehensive. There is very little need for a GP in, say, Truro to have access to information about a patient who lives in Leeds. There's occasional value in an emergency situation but the price paid in complexity is way too high for the marginal benefit.
3. It's top down and doesn't engage users. No doubt there are 'representative' panels of various stakeholders but it's just insane to think that any significant proportion of NHS staff feel engaged by the process. The risk here is that care providers, especially physicians, will work around the system. Clerical workers or, more likely and worse, nurses will be diverted from their duties to do data entry and extraction for doctors who can't be bothered.
4. Because of the monolithic nature of the system most of the benefits accrue when and only when the thing is fully implemented. This means that while costs (and hassle to future users) are incurred from day 1 of development, real benefits are years away. This makes it really hard to maintain enthusiasm even among people who start out as supporters.
5. Sheer size means that the system is going to be a bugger to maintain and virtually impossible to change as new technologies become available. It doesn't help that this system is going to take so long to implement that when complete it's core technologies will already be obsolete.
The sad thing is there is an alternative. The government could have funded local initiatives focussed on specific needs while setting data standards that would facilitate moving data between systems. (This isn't too hard as generally agreed standards exist for most health care data sets). Communication/interoperability between such systems could be encouraged and made to happen where it was needed. The big advantage of such an approach is it feeds off demand by the people who will use the systems which makes it much more likely that the systems' full potential; will be realised and real benefits generated. It also produces some benefits relatively quickly which helps build and sustain momentum.
Such an "organic" solution is messy, hard to plan and budget for, not susceptible to central control and inevitably will include some failures so it won't ever find favour with bureaucrats and politicians. Genuinely good ideas rarely do.
1. It's monolithic, one size fits all. The assumption is that the same system will meet the needs of GPs, emergency rooms and the full spectrum of specialists. This is just nonsense. The ways of working, data needs and presentation layer needs vary quite widely. Even within a single clinical area like oncology the needs of different agents are significantly different.
2. It's unnecessarily comprehensive. There is very little need for a GP in, say, Truro to have access to information about a patient who lives in Leeds. There's occasional value in an emergency situation but the price paid in complexity is way too high for the marginal benefit.
3. It's top down and doesn't engage users. No doubt there are 'representative' panels of various stakeholders but it's just insane to think that any significant proportion of NHS staff feel engaged by the process. The risk here is that care providers, especially physicians, will work around the system. Clerical workers or, more likely and worse, nurses will be diverted from their duties to do data entry and extraction for doctors who can't be bothered.
4. Because of the monolithic nature of the system most of the benefits accrue when and only when the thing is fully implemented. This means that while costs (and hassle to future users) are incurred from day 1 of development, real benefits are years away. This makes it really hard to maintain enthusiasm even among people who start out as supporters.
5. Sheer size means that the system is going to be a bugger to maintain and virtually impossible to change as new technologies become available. It doesn't help that this system is going to take so long to implement that when complete it's core technologies will already be obsolete.
The sad thing is there is an alternative. The government could have funded local initiatives focussed on specific needs while setting data standards that would facilitate moving data between systems. (This isn't too hard as generally agreed standards exist for most health care data sets). Communication/interoperability between such systems could be encouraged and made to happen where it was needed. The big advantage of such an approach is it feeds off demand by the people who will use the systems which makes it much more likely that the systems' full potential; will be realised and real benefits generated. It also produces some benefits relatively quickly which helps build and sustain momentum.
Such an "organic" solution is messy, hard to plan and budget for, not susceptible to central control and inevitably will include some failures so it won't ever find favour with bureaucrats and politicians. Genuinely good ideas rarely do.
