Figure 2

Sensitivity analysis: percent difference of the Alternative strategy compared to the Status Quo for two health outcomes. Variation in health outcomes is shown on the horizontal axis and results from changes in select individual-level model parameters, which are listed on the vertical axis. To the left of the origin (i.e., 0%) is the percent difference of the Alternative strategy compared to the Status Quo regarding mean time on treatment among treated individuals (dark purple) and life expectancy of treated individuals (light purple). To the right of the origin is the percent difference of the Alternative strategy compared to the Status Quo regarding the mean number initiating antiretroviral therapy (ART) annually (dark green) and life expectancy of both treated and untreated individuals (light green). “+HIV RNA monitoring” refers to the addition of both HIV RNA monitoring to base case assumptions. “↑ Discordant response” indicates an increase in the fraction of discordant responses to ART (i.e., no immunologic response to ART among those virologically suppressed) from 5% to 19.1%. “ART effect” refers to the independent effect of ART on AIDS-related mortality. The percent difference in life expectancy among treated and untreated individuals for the Alternative strategy compared to the Status Quo is less than among treated individuals only, a sub-population in this analysis. Therefore, the percent difference in life expectancy at the population level serves as a conservative estimate of the public health benefit of the Alternative strategy. ART: antiretroviral therapy; LTFU: loss to follow-up; LE: life expectancy.