Monte Carlo simulation of the heterosexual selective spread of the human immunodeficiency virus

The spread of human immunodeficiency virus (HIV) by heterosexual intercourse, during the first 2 years following the introduction of the virus among a sexually active and unprotected group of men and women, is modelled by Monte Carlo simulation. A beta distribution of the infectee's risks of infection per infected partner-month is assumed, with the same coefficient of variation as used in previous studies for risk of conception (natural fecundability), but with a mean of 0.04 per infected partner-month, after scaling the distribution. The number of sexual partners that one sex has (here, the women), is assumed to be more variable than for the other, with a mean of 2 partners each. The individual infection risks per infected partner-month are generated initially and do not change, but some random gains and losses of partners occur each month. Infections are updated each month. It is found in this simple model that women who become infected by 2 years had a mean risk of infection (not counting the original infector) only about 13% higher than the others. Some implications of the low selection are noted. Very great variability in the number of infections subsequently due to an index seropositive is found, which prevents easy discrimination in the characteristics of "at-risk" persons.