Indirect methods for estimating prevalent HIV infections: adults in England and Wales at the end of 1993.

Two indirect methods were used to estimate the point prevalence of HIV infection in England and Wales at the end of 1993 using data on diagnosed HIV infections, AIDS cases, HIV-related deaths and HIV testing behaviour from unlinked anonymous surveys. The methods estimated the proportion of all prevalent HIV infections that diagnosed infections represented. Most of those exposed to HIV infection through injecting drug use or sexual intercourse between men had had their infections diagnosed compared to less than half of those exposed through heterosexual intercourse. The total estimated number of prevalent infections was 22,350 for the diagnosis interval method and 20,540 for the test history method, and about 56-57% of these were in homo/bisexual men. These indirect methods are cheap and simple applications of surveillance data which provide estimates that compare favourably with those produced by more complex methods.