@inproceedings{orne-gliemann_is_2017, address = {Paris}, title = {Is there an effect of universal {ART} on sexual behaviours in rural {KwaZulu}-{Natal}, {South} {Africa}? {ANRS} 12249 treatment-as-prevention ({TasP}) trial}, url = {http://programme.ias2017.org/Abstract/Abstract/2390}, abstract = {Background: There are concerns that the implementation of Universal Test and Treat (UTT) could increase populationlevel sexual risk behaviours. We analysed the effect of universal ART vs CD4-guided ART (start at CD4≥350 then ≥500) on sexual behaviours over time, in the context of the cluster-randomised TasP trial. Methods: As part of the 6-monthly home-based survey rounds conducted in 11x2 clusters, a sexual behaviour questionnaire was administered to all residents ≥16 years. We used GEE modelling stratified by gender, to compare reported condom use at last sex (CLS), and multi-partnership (≥2 sexual partners) among those sexually-active in the previous six months across trial arms. We tested whether the sexual behaviours changed over time differently in each arm by inclusion of an interaction term between survey round and arm, using the Quasi-likelihood Independence Criterion (QIC) statistic to compare models. Results: The analysis included 43,106 reports of partnerships (22,974 control, 20,132 intervention) across 7 survey rounds (SR), between 03/2012 and 06/2016. There were no consistent or substantive changes in CLS over time neither by gender nor by arm (fig 1a, 1b); inclusion of an interaction term improved the model fit, reflecting small differences between arms in CLS over time. Less than 1.5\% of women reported multiple partnerships at any SR, too few for modelling (fig 1d). The proportion of men reporting multiple partnerships decreased significantly during the study (aOR 0.79, 95\% CI 0.75, 0.83), p{\textless} 0.001), similarly for each arm (interaction not significant), with overall a small, but significant higher proportion reported in the universal ART arm (13.7\%) vs CD4-guided ART (12.1\%) (OR=1.15, 95\% CI (1.03, 1.27), p=0.02 (fig 1c). [Figure 1] Conclusions: There is no evidence of increased unprotected sex with universal ART in this South African population. Continued monitoring of population-level sexual behaviour indicators, in particular multiple partnerships, is needed as the UTT strategy is rolled out.}, publisher = {poster WEPEC0968}, author = {Orne-Gliemann, Joanna and Rolland, Matthieu and Tiendrebeogo, Thierry and Larmarange, Joseph and Pillay, Deenan and Dabis, François and McGrath, Nuala and {ANRS 12249 TasP Study Group}}, month = jul, year = {2017}, }