Fig 4. Weighted risk of early mother to child transmission among mothers with HIV seroconversions during pregnancy compared to mothers with reported HIV-positive status prior to the current pregnancy.

This figure illustrates the estimate (diamonds) and 95% confidence intervals (vertical lines) for the proportion of HIV-infected mothers who had transmitted HIV to their infants by 4–8 weeks postpartum. Proportions are adjusted for the survey sampling design including clustering, design effects and nonresponse, and weighted to the population distribution of live-births for South Africa (SA) in 2011. The left panel of the graph shows a significantly higher proportion of mothers who seroconverted during pregnancy transmitted HIV to their infants than did mothers whose HIV infection was known to them before or during this pregnancy. The right panel subsets the women who seroconverted during pregnancy by the estimated week of gestation in which they seroconverted. Estimates of time of seroconversion were imputed using a uniform distribution to assign an equal probability of seroconversion to each day between delivery and the last reported negative HIV test during pregnancy. The graph shows results based on combined analysis of 20 imputations of this variable, and shows that the confidence intervals for the estimated proportion of mothers transmitting to their infants did not differ whether the mother was estimated to have seroconverted before or after 32 weeks gestation.