Projections of the course of the acquired immunodeficiency syndrome (AIDS) epidemic are crucial for developing public health policies to control the spread of human immunodeficiency virus (HIV) infection and to meet future health care needs. The first Public Health Service projection1 was 270 000 cumulative AIDS cases in the United States by the end of 1991. This projection was made at the 1986 Coolfont Planning Conference, at which time fewer than 30 000 AIDS cases had been reported, and, although this projection appears to have overestimated the incidence of AIDS somewhat, it has proved to be a good guide for health planning.
The original successful Coolfont projection was based on simple extrapolation of the observed incidence of AIDS.2 Successful extrapolation depends on a fortunate choice of the mathematical curve used, and certain specialized mathematical functions can produce surprising results. For example, in this issue of The Journal, Bregman