A model for estimating incremental benefits and costs of testing donated blood for human immunodeficiency virus antigen (HIV-Ag).

We propose a model for estimating the benefit of adding a test for the human immunodeficiency virus antigen (HIV-Ag) to current procedures for testing donated blood. Using this model, data on HIV infection from published studies, and certain assumptions about blood donor behavior, we estimate that the probability of detecting an additional HIV-infective blood component is approximately 1 in 4,860,000. If this estimate is correct, adding HIV-Ag testing would prevent approximately 4 cases of primary transfusion-transmitted HIV infection annually in the United States. After adjustments for the median incubation period for AIDS, and for mortality due to primary illnesses, this estimate represents prevention of approximately 1 case of AIDS per year, within the 4 years after transfusion. A primary advantage of this model is its adaptability for recalculating cost-benefit analyses if more sensitive tests for HIV infection become available. In addition, we propose that comparing the anticipated costs and benefits of HIV-Ag testing to other possible uses of these funds should be an important factor in assessing the desirability of HIV-Ag testing.