RESUMO
Peer advocacy can promote HIV protective behaviors, but little is known about the concordance on prevention advocacy(PA) reports between people living with HIV(PLWH) and their social network members. We examined prevalence and correlates of such concordance, and its association with the targeted HIV protective behavior of the social network member. Data were analyzed from 193 PLWH(index participants) and their 599 social network members(alters). Kappa statistics measured concordance between index and alter reports of PA in the past 3 months. Logistic and multinomial regressions evaluated the relationship between advocacy concordance and alter condom use and HIV testing behavior and correlates of PA concordance. Advocacy concordance was observed in 0.3% of index-alter dyads for PrEP discussion, 9% for condom use, 18% for HIV testing, 26% for care engagement, and 49% for antiretroviral use discussions. Fewer indexes reported condom use(23.5% vs. 28.1%;[Formula: see text]=3.7, p=0.05) and HIV testing(30.5% vs. 50.5%; [Formula: see text]=25.3, p<0.001) PA occurring. Condom advocacy concordance was higher if the index and alter were romantic partners(OR=3.50; p=0.02), and lower if the index was 10 years younger than the alter(OR=0.23; p = 0.02). Alters had higher odds of using condoms with their main partner when both reported condom advocacy compared to dyads where neither reported advocacy(OR=3.90; p<0.001) and compared to dyads where only the index reported such advocacy(OR = 3.71; p=0.01). Age difference and relationship status impact advocacy agreement, and concordant perceptions of advocacy are linked to increased HIV protective behaviors. Alters' perceptions may be crucial for behavior change, informing strategies for improving advocacy.
RESUMO
BACKGROUND: Delayed detection of ART failure in settings without access to viral load (VL) monitoring has been hypothesized to lead to suboptimal response to second-line therapy due to accumulated drug resistance mutations (DRMs). We tested this hypothesis in a program setting in rural Uganda. METHODS: From June 2012 to January 2014, we enrolled participants receiving nonnucleoside reverse transcriptase inhibitor-based first-line ART for ≥4 years, without access to VL monitoring. Participants who had a measured VL ≥ 1000 copies/mL on two occasions were switched to protease inhibitor-based regimens and followed every 6 months until September 2016. We measured VL at study exit. We conducted DRM testing at enrollment and study exit and examined factors associated with virologic failure. RESULTS: We enrolled 137 participants (64.3% female) with a median age of 44 years and a median duration on ART of 6.0 years. In a median of 2.8 years of follow-up, 7 (5%) died, 5 (3.6%) voluntarily withdrew, and 9 (6.6%) became lost to follow-up. Of 116 participants with a VL result at study exit, 20 (17%) had VL > 1000 copies/mL. Virologic failure was associated with reporting suboptimal adherence ( P = 0.028). Of patients with DRM data at enrollment, 103 of 105 (98%) had at least 1 DRM. Participants with thymidine analog mutations at enrollment were less likely to have virologic failure at study exit (11% vs. 36%; P = 0.007). No other DRMs were associated with failure. CONCLUSION: Even in the presence of multiple DRMs on first-line therapy, virologic failure after 3 years of protease inhibitor-based ART was infrequent. Suboptimal adherence to ART was associated with virologic failure.