Figure 5. The relationship between the long-term false-recent rate (ε) and the pre-set optical density cut-off (C) in the ZVITAMBO Trial.

The value of ε was estimated as the proportion of cases with a BED OD<C among women tested at T = 12-months postpartum, given that they had previously provided a positive HIV test at baseline. Error bars indicate the 95% confidence intervals.