Table 1. Overview of successive addition of components and structures to each of the nine models.
Component | Model A | Sub-Model B1 | Sub-Model B2 | Model B | Sub-Model C1 | Sub-Model C2 | Sub-Model C3 | Model C | Model D |
Prevalence density functiona | X | X | X | X | X | ||||
Age-structured population | X | X | X | X | X | X | X | ||
Transmission probability by disease stage | X | X | X | X | X | X | X | ||
Heterogeneity in sexual behavior | X | X | X | ||||||
STI co-factors, male circumcision, and condom use | X | X | X | ||||||
Up-to-date ART effectiveness assumptions | X | X | X | ||||||
Current ART scale-up from 2003 onwards | X |
This artificial prevalence density function was introduced by Granich et al. [9] to mimic processes that result in the observed leveling off of the HIV epidemic. In the more comprehensive models of our analysis we replaced this prevalence density function by the actual processes that may be responsible for the leveling off.