Estimating the cumulative incidence of HIV infection among persons with haemophilia in the United States of America.

We used complementary approaches to estimate the cumulative incidence of infection with the human immunodeficiency virus (HIV) among persons with haemophilia in the United States of America. One approach, ratio estimation, divided the cumulative number of haemophiliacs diagnosed with acquired immunodeficiency syndrome (AIDS) in the United States by the corresponding cumulative proportion with AIDS among HIV-positive subjects in the Multicenter Hemophilia Cohort Study (MHCS). The other approach, back-calculation, reconstructed past HIV incidence from national surveillance of AIDS using essentially non-parametric estimates of the hazard functions for AIDS and for pre-AIDS death. We derived confidence intervals that fully incorporated uncertainty about the hazard functions. Results from the two approaches were consistent. Around 9200 haemophiliacs became infected during the course of the epidemic. Of them, around 7000 were living with HIV or AIDS as of 31 December 1992 and at least 5630 as of 31 December 1995. Credible calculations for this group must account for those who die before AIDS and for the significantly longer incubation times in those infected as children or adolescents. The consistency of back-calculation and cohort data in haemophiliacs supports the use of back-calculation to estimate prevalence in other populations.