Table 1. Main characteristics of agent-based model for combination HIV prevention among MSM in peri-urban South Africa (additional information and specific parameter values are in the Supporting Information S1).

Attributes assigned to each person at start

Frequency of sexual activity

HIV status at start

CD4 count at start if HIV +

Knowledge of HIV status at start (yes, no)

Sexual role preference (insertive, receptive, versatile)

HIV testing frequency (3 levels: moderate, low, never)

Some assigned a main partner

Proportion of sexual contacts that are UAI (2 levels)

Sexual networks of regular partners (allowance for sero-sorting)

Daily updates

Daily sexual contacts depends on type of partnership

Likelihood of contact (in decreasing order): main, regular, casual, have other main partners

HIV testing possible

UAI rate adjusted if learns knowledge of HIV status

CD4 levels updated for HIV positive

Infection status updated

Prevention Interventions

ART for eligible HIV positives

Eligible: HIV test within 6 months and CD4 <350

Considered varying levels of coverage (X 1)

PREP for eligible HIV negatives

Eligible: in last 6 months had both HIV test and >12 UAIs or had infected main partner

Varying levels of PREP acceptance (X 2) with two levels of adherence (low and high)

Reduction in UAI frequency (considered varying reduction levels (X 3))

Increase in HIV testing: convert 50% of the never testers to low frequency testers