These data show the importance of distinguishing between total numbers of people living with HIV/AIDS, many of whom will have been infected many years ago, and new infections. (The technical terms that you may come across, if you read more on HIV statistics, are prevalence for numbers living with the virus at a given time and incidence for new cases in a given period.)

But how do the HPA know how many people have HIV infections? As in most developed countries, the information comes from an elaborate system of disease surveillance.

The HPA and other official agencies have set up a detailed system for the systematic recording of cases of HIV and AIDS that come to the attention of various agencies. This information comes from clinicians (on new diagnoses and when people living with HIV access treatment), from laboratory reports, and from the death registration system (when a death from AIDS-related cause is recorded).

In addition, data come from a system of unlinked anonymous surveillance. The aim of this system is to measure the extent of undiagnosed HIV infection, by testing samples of (for example) blood that were principally provided for another purpose, for instance by people attending genito-urinary medicine clinics or specialist treatment centres for injecting drug users. Samples from pregnant women are also tested.

It would clearly be unethical to carry out such testing on such a wide range of individuals who had not specifically requested an HIV test, if the results of the tests could be linked back to the individuals involved., since there can be serious social and legal consequences of having volunteered for such a test. Therefore careful precautions are taken so that the individual test results cannot be linked back to the details of the patients who provided the samples. (Patients are made aware of the possibility that their samples may be anonymously tested, and are allowed to refuse permission for such use, but in fact there are few refusals.)

Results from unlinked anonymous surveillance play a key role in the HPA’s estimates of the total number of undiagnosed cases of HIV infection in the UK. Unlinked anonymous surveillance provides estimates of the percentage of people with undiagnosed infection in various population groups (divided up by age, sexual behaviour and so on), but they do not give the actual numbers of infected people in these groups. However, the HPA also have reasonably good estimates of the total numbers of people in these groups from the Census and other Government population statistics, and from a major survey of sexual behaviour (NATSAL, the National study of Sexual Attitudes and Lifestyles). Putting all these sources of information together provides reasonably accurate estimates of numbers of undiagnosed HIV infections.

It takes considerable time, money and effort to produce and process all this information on HIV and AIDS, and the HPA’s aims in doing so are not simply to produce the fodder for scare stories in the media. HPA data are used for many crucial purposes. They help with the identification of patterns in infection that can be used to target health promotion advice more effectively. They allow the effects of antiviral treatments to be monitored on a large scale. They vastly improve the forecasting of future health care needs for people with HIV. They allow changes in patterns of infection to be identified quickly, so that they can be acted on.

  < previous   Page 2 of 2

Content last updated: 14/07/2004

About our expert

Kevin McConway is a Senior Lecturer in the Department of Mathematics and Statistics at the Open University, where he teaches statistics and health studies, and researches in several areas including statistical theory, health service organization, ecology and evolution.

He has degrees in mathematics, statistics, psychology and business from the Universities of Cambridge and London and the Open University. Kevin originally comes from rural Northumberland but is now a long-term Milton Keynes resident.

Comments