Medical Software's Free Future

BMJ 2000;321:976 ( 21 October )

Editorials

Medical software's free future

Open collaboration over the internet is changing development methods

The government in the United Kingdom spent £7.1bn ($9.9bn) on information systems in 1998-9, of which £1bn was in health care. Yet information systems are difficult to commission, purchase, and evaluate, and the results not always good.1

As computer hardware becomes an ever cheaper commodity with ever increasing power, it is clear that software is the rate limiting step in system development. Software is slippery stuff: its possibilities seem almost limitless, but implementing a system competently is a difficult activity that commands premium rates of pay. A lot of its cost lies in planning, implementing, and monitoring and enforcing exchanges between the parties involved, who might be, for example, a hospital wanting to buy an information system and a system supplier. Such exchanges have high transaction costs.2 The relationship between an information systems supplier and its clients has, according to transaction cost economists, the quality of "information impactedness": a state in which one of the parties to an exchange is much better informed than the other, and the other cannot achieve information parity, except at great cost.

Even when a system is successfully commissioned, the costs can remain high. Once a customer is "locked into" proprietary software, its makers can demand premium prices, safe in the knowledge that the client would find it even more expensive to change.3

It is such forces that have led to the rise of free software---most notably the GNU/Linux operating system, which is freely available for download from the internet.4 (An operating system, such as Microsoft Windows, is the essential software that runs a computer's basic functions.) Free software differs from proprietary software in several important respects. Most importantly, its licence (the General Public License (GPL)) encourages free copying, distribution, and modification of the software.5 There is only one catch: users must make any modifications that they make to the software available to others on the same basis that they received it. This virtuous cycle of development has, over the past decade, created a commonwealth of high quality software.

Free software facilitates the provision of common software components. As well as the saving on licence fees, it allows software engineers to concentrate on the important part of system development: customising components for the organisation that they serve.

There are other advantages. It is reliable and secure: source code can be inspected for bugs and security flaws before it is compiled for use. It can be maintained even if the developers who originally produced the software are no longer available. Many high quality components exist ready made, which allows new projects to build on the existing base of code; developers can spend their time creatively exploring new and unsolved problems rather than duplicating effort.6

Free software concepts make particular sense in medicine: although peer review has its problems, medical knowledge is becoming more open, not less,7 and the idea of locking it up in proprietary systems is untenable. And professional staff should not invest time learning the user interface of proprietary systems that may change, be withdrawn, or be arbitrarily "upgraded" for commercial reasons. Much better instead to invest time on a system licensed under the General Public License that will always be free.

The European Union has already embraced open source: its fifth framework programme (which will fund 3.6bn Euros of research and development over the next 5-10 years) places a strong emphasis on projects which will yield open source software as one of the outputs.8 Next week the NHS Information Authority hosts a seminar to consider the implications of the free software movement for its future strategy. If it chooses (as it should) to use and encourage open source development methods throughout the organisation, it will find a host of high quality programmes already under way across the world.9 Leveraging this effort should reap rewards for managers, professionals, and patients alike.

Douglas Carnall, associate editor.

BMJ

1. House of Commons Public Accounts Committee. Improving the delivery of government IT projects: report and proceedings London: Stationery Office, 1999 (HC65).

2. Mick SS. Explaining vertical integration in health care: an analysis and synthesis of transaction-cost economics and strategic-management theory. In: Mick SS, ed. Innovations in health care delivery. San Francisco, CA: Jossey-Bass, 1990.

3. Shapiro C, Varian HR. Information rules: a strategic guide to the networked economy. Boston, MA: Harvard Business School Press, 1999.

4. Debian 2.2 GNU/Linux operating system. www.debian.org

5. Free Software Foundation. GNU General Public License. In: DiBona C, Ockman S, Stone M, eds. Open sources: voices from the open source revolution. Sebastapol, CA: O'Reilly and Associates, 1999. www.gnu.org

6. Raymond ES. The cathedral and the bazaar. Sebastapol, CA: O'Reilly and Associates, 1999.www.tuxedo.org/~esr/writings/cathedral-bazaar/

7. Smith R. The future of peer review. In: Godlee F, Jefferson T, eds. Peer review in health sciences. London: BMJ Books, 1999.

8. European Commission. Information society technologies: a programme of research technology development and demonstration under the 5th framework programme. 2000 work programme. Brussels: EC, 2000 ftp://ftp4.cordis.lu/pub/ist/docs/b_wp_en_200001.pdf (downloaded 11 October 2000)

9. Valdes I. LinuxMedNews Open Source Medical Project List. www.linuxmednews.com/linuxmednews/955216338/index_html (downloaded 11 October 2000)

http://bmj.bmjjournals.com/cgi/content/full/321/7267/976