Scottish data

Scotland has a comprehensive system for recording both AIDS diagnoses and the numbers of HIV positive individuals under immunological monitoring who have low CD4 counts. The CD4 Collaborative Group (1992) provide data anually on new CD200 cases under monitoring in Scotland. A case becomes CD200 when two consecutive CD4 counts below 200 are obtained (the date of the first count defines the date of diagnosis) or when an AIDS defining diagnosis is reached.

Various investigations provide information on the likely incidence of new infections in Scotland since the start of the epidemic. They all point to an epidemic with a sharp peak of new infections in 1983 and 1984 and a steep decline thereafter. The total size of the HIV infected population can be estimated from the numbers of known HIV infections and the proportion of AIDS and CD200 cases who have their first HIV test around the time of their AIDS or CD200 diagnosis. Estimates in the range 1800–4000 were considered possible.

An Edinburgh study of a clinical cohort (McNeil 1993), along with results from elsewhere, allow us to define a range of possible incubation distributions to use in predicting new AIDS and CD200 cases. Incubation distributions of the Weibull and Gamma families were used. A treatment effect was incorporated from 1989 onwards, when pre-AIDS therapy was widely used.

Sensitivity analyses

Initial back-projections investigated the sensitivity of the predictions to the various features of the incubation distribution and the infection curve. These included back-projections where the infection curve was estimated for a fixed incubation distribution.