Merits of open source

From Ganfyd

Jump to: navigation, search


The Merit of Open Source for medical record and management software.

In order to make distributed medical records work, and to support a movable patient record, systems of terminology are required. In order to make this work well we all need to use the same system of terminology, or we need to be able to use all systems that are available. The history of the NHS attempt to achieve this with the Read Code demonstrated that persuading everyone to buy the same system doesn't work.

Galen - the medical terminology system - recently went open. Anyone can get it, anyone can use it. This is a good thing. (given the slogan of Galen - "Making the impossible very difficult" - people may be forgiven for thinking that it would be easier to start from scratch and reinvent a collection of wheels themselves…usually it isn't, in the end. A great merit of open source is that you can reuse parts of other projects)

Why is IT so little used?

Look for a moment at a hospital or general practice. Its origins are probably before the NHS, and it is expected to continue for several years, perhaps until after the NHS. Its culture, and ways of working have changed greatly but always been within its control and have evolved.

Look at the supplier of its IT. Its origins are quite likely within 5 years, and few people would guarantee that it will persist beyond tomorrow. An exception to the rule is IBM, which is some 50 years old now.

One of the reasons that IT has not made such an impact on general practice as it should have is that sensible businesses do not make themselves dependent on something that can be removed suddenly if their supplier merges or folds, as rather too many GP system suppliers have done.

The Open source model of development offers a way around this, it has other merits, but this is a sufficient one to make its adoption highly desirable.

Can it work?

Until recently it was suspected that selfishness and other human behaviour would cause any such project to fail. However the rise of Linux, an open source operating system, and of Apache and several other large programs produced by the same mechanisms has demonstrated that it can be done. The current challenge is to produce the elements of complete business systems, and they are well-advanced.

Eric Raymond writes well on the Open Source Model of development[1], that large software projects tightly controlled and planned in a cathedral-like atmosphere - NHS Net and X.400 leap to mind - are outraced, out-evolved and out-performed by software from a horde of booths in the bazaar outside, as the Internet allows the bazaar to function as though it has an intelligence greater than the sum of its parts.

Serious writers of serious software, and system integrators, expect to earn a living at it, the Open Source model does not hamper this. People who acquire your software free will either not use it, in which case you have lost nothing and may have avoided a tedious effort to sell it to them, or they will use it. If they require assistance or development then they are likely to return to you. If there is a fault that bothers them, and they have the facility to do so, they may fix it, and since they would want this fix integrated with the next version, they will give it back to you. Establishing a community of users is the difficult part of setting up functioning large-scale systems, giving the software away helps this.

How Open is Open.

Software has no value, it is not a made good like a car, it can be copied with no incremental cost. The paradox in open source is that by giving away the secrets, companies can make a better living than by trying to protect them. However like any computer industry standard not everyone means the same thing when they talk about Open Software. Companies find it hard to relinquish control, and so instead of using the GNU Public Licence (GPL), which is as open as a very open thing, they make up their own, like Sun's Community Licence. It is better to go for the real thing, let it go as GPL.

Galen is offered on terms that are broadly similar to the Gnu General Public Licence. It might be simpler to just adopt the GPL but this is a laudable step in the right direction.

What is available to put the terminology into?

There are several free software developments aimed at an open source medical record and medical establishment management system. Littlefish in Australia is one, as are Freemed and FreePM which are based in the States. Each has a world-wide community of interest, who at least provide advice and will test and develop within their abilities. Littlefish has recently been given the Good European Health Record kernel, which seems an excellent step by GEHR, and interestingly arose from an original consultancy study by IBM. If the NHS is serious about making Read a widespread standard, then it is past time it was released under similar conditions. The Canadian firm Minoru Developments is actively demonstrating the application of open source elements to specific solutions.

So from a long-duration computer company and from the ethos and community of the Internet, we may soon see a medical record system capable of providing general practices and other health organisations with a record system capable of matching their anticipated longevity. I look forward to it.


[1] The Cathedral and the Bazaar Eric S Raymond [2] GNU Public Licence [3] [4] Open Source Healthcare Alliance

Personal tools