Information in practice

Back to basics on NHS networking

Justin Keen, fellow ^a, Jeremy Wyatt, reader and director, Knowledge Management Centre ^b. 

^a King's Fund, 11-13 Cavendish Square, London W1M 0AN, ^b School of Public Policy, University College London, London WC1E 7HN

Correspondence to: J Keen jkeen{at}kehf.org.uk

The success of the internet poses two challenges to thinking about electronic networking in health care. One is technological: if you do not want to use the internet you need to show that your chosen alternative is more appropriate and cost effective. The second challenge is more conceptual. The internet embodies a particular way of thinking about communications, emphasising open and sustainable solutions. Are these the key concepts that should drive our thinking in health care, or are others more appropriate?

At present the NHS has a dedicated private electronic network service called NHSnet, which has now been operational for four years. Throughout its history it has been dogged by negative publicity, stemming in part from unhappiness with the user charges levelled before 1999, a lack of useful resources to access, and detailed objections to policies for security and access to personal data.¹ However, the NHS Executive signalled changes in its policies for NHSnet in late 1999,² which might make it a more attractive option for clinicians. In addition, the NHSnet commercial contracts are due for renewal from 2002 to 2004, so there is merit in reviewing the current networking strategy to inform future decisions.

In this article we present data on NHSnet, and contrast it with the internet. These are by no means the only possible networking options available, but contrasting the two serves to highlight key evidence and arguments about NHS networking.

	What is the evidence?

Top What is the evidence? User requirements for... Conclusions References

The information available on the costs and benefits of networking via NHSnet is limited. Full business cases for NHSnet were not prepared at the time of the two major NHS strategies, the 1992 Information Management and Technology Strategy³ and Information for Health in 1998.⁴ The main NHSnet contracts are provided through the private finance initiative (see box), and they have not been published.

Costs
The NHS Executive is currently paying £3.8m a year for hospital trust and health authority links to NHSnet centrally, which includes unlimited access to the internet (data provided by the NHS Information Authority in response to an "open access code" request). The figure shows that the costs of use of NHSnet have grown linearly over the past five years. The NHS Executive is also meeting centrally the costs of messaging, which are estimated at £2.8m for 1999-2000. The crude mean cost to the NHS Executive of connection and messaging for 1999-2000 is therefore about £10 600 per organisation. The NHS Information Authority was unable to provide us with data on general practitioners' and other primary care use of NHSnet, but told us that about 70% of computerised practices were connected at the end of August 2000. 

View larger version (16K): [in this window] [in a new window]   Total charges for use of NHSnet between 1995-6 and 1999-2000

Benefits
For benefits, reliable figures are again elusive. The NHS Executive hopes that NHSnet will help save £100m a year in NHS running costs. ^{4 6} However, convincing evidence of cost savings is lacking,⁷ as is evidence of clinical or management benefits attributable to NHSnet. More generally, there is scant evidence about the costs and benefits of the NHSnet networking technology in non-health settings. ^{8 9} Similarly, the benefits attributable to the internet have not been quantified, but observation of the rapid growth of internet connections¹⁰ and the large volume of health related resources on the world wide web suggest that it is widely judged to be useful.

	User requirements for networking

Top What is the evidence? User requirements for... Conclusions References

It is commonly assumed that large scale electronic networks are inevitable, so people do not stop to ask basic questions about their purpose and value. But what are networks actually for? One possibility is that they offer a means of communicating information that is more cost effective, secure, and reliable than paper based media (see table). Another possibility is that networks can be used to support wider policy developments, such as "joined up" government.¹¹ A networking policy that underpins cross boundary working will need to have open membership, in the sense that it should allow many different individual organisations to join the network without prejudging the technology they use and the information they need to communicate to one another. Technical solutions also need to be sustainablethat is, allow for likely changes over time in the technology itself and in patterns of use.

These five criteriacost-effectiveness, security, reliability, open membership, and sustainabilitycan usefully be thought of as a general statement of user requirements for NHS networking. The 1992 and 1998 NHS information strategies focused on specifying technologies, and there is no published user requirement for networking. We have already considered costs and benefits in this article and elsewhere,¹² and so we focus next on the other four criteria.

Security
Security has been one of the main sources of controversy about NHSnet.¹ Security threats can be external, from people attempting to hack into the network, or internal, with authorised staff misusing their access to sensitive data. ^{13 14} A study by the Audit Commission of fraud and abuse of information technologies indicated that a half of all public sector organisations are now affected.¹⁵

Reliability
Reliability is a key goal for both NHSnet and the internet. NHSnet adopted a closed network approach, broadly similar to that used by banks and other institutions that manage large volumes of transactions every day. In contrast, the internet uses protocols developed for ARPAnet, the original US military network that spawned the internet, with arbitrary routing over open networks to avoid problems such as broken links or nodes. Proper comparison of reliability is difficult, but NHSnet was "down" for about 2.2% of the time in late 1999 (although we were told that this has improved this year), and Demon Internet business services were down for around 1% of the time (see table). This criterion seems to favour the internet, though not decisively.

Membership
There are also differences between NHSnet and the internet on the issue of membership. NHSnet was conceived as a dedicated NHS network, and almost all health authorities and NHS trusts are now connected. This excluded patients, all health care provided outside the NHS, and many statutory organisations that need to communicate with the NHS (see box). In retrospect it seems that the NHS was viewed as a self contained organisation rather than one inextricably linked to the world around it. This position is changing, and there are now plans to link social services and other organisations to NHSnet. But the logical outcome must be a network with many non-NHS users, which tends to undermine the argument for sole reliance on a private network.

Sustainability
A sustainable networking strategy is one that allows individuals and organisations to change their own working practices, and the ways in which they use a network, and yet be able to continue to use the network without serious impediment. In effect this means that the economics of the solution, as well as the technology itself, must continue to make sense. In the case of information technologies, exposure to dynamic markets may lead to both cost reductions and innovations, which suggests that the NHS should use systems and software that are "open" and have many suppliers.

	Conclusions

Top What is the evidence? User requirements for... Conclusions References

The elements of the user requirement outlined here are interlinked: the best technical solution, now and in the future, must make sense economically and provide the required security and reliability. The NHS Executive has recognised that NHSnet in its original form was neither "modern and dependable"²⁰ nor the most appropriate or cost effective solution available and now seems to be more open to discussion of the merits of alternative approaches. Perhaps we can now move to a more considered debate about future networking options.

	Acknowledgments

We thank the NHS Information Authority for the data on NHSnet and to Demon Internet for their data. We thank staff at the NHS Executive, NHS Information Authority, members of the BMJ editorial committee, and anonymous referees for comments on earlier drafts. The NHS Information Authority's response to the authors' request about open access government can be found at www.kingsfund.org.uk/ehealthsystems/html

	Footnotes

Competing interests: None declared.

	References

Top What is the evidence? User requirements for... Conclusions References

1.	Anderson R. Clinical system security: interim guidelines. BMJ 1996; 312: 109-111[Free Full Text].
2.	Mitchell P. Go with the flow [IT update]. Health Serv J 2000; Mar 16: 16.
3.	Department of Health. Information management and technology strategy. London: DoH, 1992.
4.	NHS Executive. Information for health. Leeds: NHS Executive, 1998.
5.	Denham J. House of Commons official report (Hansard). , 1999 Apr 30: col 279.
6.	National Audit Office. The 1992 and 1998 information management and technology strategies of the NHS Executive. London: Stationery Office, 1999. (HC 371, session 1998-99.)
7.	Committee of Public Accounts. The 1992 and 1998 information management and technology strategies of the NHS Executive. London: Stationery Office, 2000. (HC 406, session 1999-2000.)
8.	Barua A, Lee B. An economic analysis of the introduction of an electronic data interchange system. Information Syst Res 1997; 8: 398-422.
9.	Lee M. Internet-based financial EDI: towards a theory of its organisational adoption. Comput Networks ISDN Syst 1998; 30: 1579-1588[CrossRef].
10.	Center for Next Generation Internet. Internet Trends. www.ngi.org/trends.htm (updated 13 Feb 2000).
11.	Cabinet Office. Modernising government. London: Stationery Office, 1999. (Cm 4310.)
12.	Keen J. Rethinking NHS networking. BMJ 1998; 316: 1291-1293[Free Full Text].
13.	Anderson R. Clinical systems security interim guidelines. BMJ 1996; 312: 109-111.
14.	Roscoe T, Wells M. NHSnetlearning from academia. BMJ 1999; 318: 377-379[Free Full Text].
15.	Audit Commission. Ghost in the machine: an analysis of IT fraud and abuse. London: Stationery Office, 1998.
16.	NHS Executive. NHS code of connection. Version 1.0. NHS Executive: Leeds, 1999.
17.	Borowitz S, Wyatt J. The origin, content, and workload of e-mail consultations. JAMA 1998; 280: 1321-1324[Abstract/Free Full Text].
18.	Wood FB, Cid VH, Siegel ER. Evaluating internet end-to-end performance: overview of test methodology and results. J Am Med Inf Assoc 1998; 5: 528-545[Abstract/Free Full Text].
19.	Bingham CM, Higgins G, Coleman R, Van der Weyden MB. The Medical Journal of Australia internet peer-review study. Lancet 1998; 352: 441-445[CrossRef][Medline].
20.	Department of Health. The new NHS: modern, dependable. London: Stationery Office, 1998.

(Accepted 14 June 2000)