Another Type Of Risk With Clinical Software That Does Not Often Get Discussed.

This article appeared a little while ago.

Birmingham to offer PICS licence-free

11 November 2013   Lis Evenstad

University Hospitals Birmingham NHS Foundation Trust will offer its Prescribing Information and Communication System to the NHS on a licence-free basis instead of “open-sourcing” it.

EHI reported last week that NHS England is working with four different organisations, one of which is Birmingham, to put their products on a new open source electronic patient record systems framework.

The trust’s medical director, Dr Dave Rosser, told EHI that Birmingham would not be offering its PICS e-prescribing system open source because it is too risky.

“We think open sourcing PICS would be dangerous. It’s too complicated a programme with very complicated code,” said Dr Rosser.

“It would be risky to say the least. It has 600,000 lines of code and it is all interactive. It’s very easy to make a change in one part that changes something in another part that not even the programmers can predict.”

Speaking at EHI Live last week, Beverley Bryant, director of strategic systems and technology at NHS England, said that up to £20m of the £250m first round of the Safer Hospitals, Safer Wards Technology Fund will go towards developing the open source option.

Twenty trusts have expressed an interest in using open source systems. IMS Maxims, HP and the US Veterans Association system, VistA, are also expected to be on the framework that trusts will be able to call-off against.

Lots more here:

http://ehi.co.uk/news/EHI/9028/birmingham-to-offer-pics-licence-free

Continuing on from other software safety risks we have discussed on the blog, a really interesting issue raised here with the complexity of an application being such that there are real concerns with having others modify it, as would be possible with open-source, because there is a serious risk of breaking the software.

This is interesting given that going open source has often been associated with improvement in software quality and reliability. Clearly the authors think the learning curve with their software is such that only those who are very involved can cope with.

Do you think the authors are right, or are they underestimating the smarts out there in ‘geek land’?

David.