Commentary: Public access to surveillance data

Sentiweb would perhaps be easier to understand if the underlying Sentinel system was more fully described.1 This system was set up in 1984 jointly by the French Department of Health, and the French National Institute of Health and Medical Research to improve surveillance of communicable diseases. There are about 500 sentinel general practitioners distributed around France. At least once a week they feed an agreed simple set of information relating to the eight diseases under surveillance across the telephone network, often using the videotex Minitel terminal freely available to all telephone users, to a national database. Associated with this database is a powerful “real time” toolbox capable of displaying the resulting aggregated data in a variety of ways as described. By using statistical tools, epidemiologists are able to detect with considerable accuracy when an epidemic is imminent and can then make this information available to the health services, which may in turn be able to take anticipatory action. The maps and time series look impressive, and there is something very satisfying about watching a video simulation of the spread of an epidemic along the nation's main communication arteries.

Sentiweb makes all of this publicly available through interactive web pages along with background information about the project team and a list of references. This open approach contrasts greatly with the philosophy of the closed NHS-wide network being set up in Britain to carry clinical, administrative, and contractual data for the NHS. Web sites on this network will be accessible only to users of the NHS network and not to the Internet.

The authors rightly raise the issue of the conflict between free access to information and the dangers of misinterpretation by the untutored. Presumably, there is a danger that the public might inappropriately react to a false alert, but surely health care organisations would want to see confirmation that the alert was authentic before reacting. In any case, because the site is freely accessible, the centre would be expected soon to be made aware of any false alarm and then be able to counter it. There are also definite benefits to having this information freely available. For example, the news media would be able to check the facts more thoroughly before running stories of imminent epidemics.

Underlying these concerns there are some very important and as yet unresolved problems relating to information obtained across the Internet. Firstly, when little or nothing is known about the author of a piece of information it is difficult to know if it has any intrinsic value. Secondly, even when the author's credentials are known, there can be difficulty in ascertaining if the work is authentic unless it has been digitally signed. Thirdly, the proliferation of web sites and news servers is truly bewildering, and we are in danger of being overwhelmed by information overload. Even with increasingly sophisticated search tools, we have problems in finding what we need, and in interpreting it when we find it.