CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

HIV (PLWH) accurately perceive norms around antiretroviral treatment (ART) adherence and whether these perceptions influence their own propensities to adhere to ART. Methods: We recruited a population-based sample of PLWH on ART in Nyakabare Parish, a rural region of southwest Uganda. Self-reported ART non-adherence was defined as missing any ART doses in the past 7 days. We also elicited their perception about the extent to which most other adult PLWH in their community were non-adherent to ART. Actual ART non-adherence was calculated by aggregating responses across all PLWH. Non-adherence was classified as normative if reported non-adherence was present among more than 50% of PLWH in the village. We then compared individuals’ perception of the adherence norm to the actual adherence norm, and also assessed the relationship between perception and personal adherence. Results: Adherence was normative among 158 adult PLWH (response rate 95%): only 15% of HIV+men and 9% of HIV+ women reported missing any doses in the past 7 days. However, approximately one-half of study participants (45% of men and 54% of women) incorrectly believed that most PLWH in their communities were non-adherent to ART. In addition, approximately one-quarter (22% of men and 25% of women) did not know whether most people had missed any doses. Only about one-quarter of this population (33% of HIV+men and 21% of HIV+ women) accurately perceived that ART adherence was normative among PLWH. Overestimating the pervasiveness of ART non-adherence was not associated with age, education, time since diagnosis, or serostatus status disclosure. Finally, there were almost three times as many non-adherents among the participants who misperceived the norm as compared to non-adherents among the participants who accurately perceived the norm (14% vs. 5%), though this difference was not statistically significant. Conclusion: Many PLWH on ART believe that non-adherence to ART is present among most PLWH on ART in their community, despite adherence actually being normative among PLWH in this population-based study from rural Uganda. Because those who are non-adherent appear to also perceive poor adherence as a normative behavior, altering those misperceptions might represent an opportunity for novel ART adherence intervention development. 497 RANDOMIZED STUDY OF AN ART ADHERENCE INTERVENTION USING A SMART-PILL BOTTLE SERVICE Grant B. Ellsworth 1 , Leah Burke 2 , Martin T. Wells 3 , Satish Mishra 1 , Matthew Caffrey 3 , David W. Liddle 4 , Malika Madhava 5 , Arsalan K. Muhammad 1 , Curtis O'Neal 1 , Peter L. Anderson 6 , Lane R. Bushman 6 , Lucas Ellison 6 , Josh Stein 7 , Roy M. Gulick 1 1 Weill Cornell Medicine, New York, NY, USA, 2 Yale University, New Haven, CT, USA, 3 Cornell University, Ithaca, NY, USA, 4 Children's Research Institute, Children's National Health System, Washington, DC, USA, 5 Thomas Jefferson University, Philadelphia, PA, USA, 6 University of Colorado Anschutz Medical Campus, Aurora, CO, USA, 7 AdhereTech, New York, NY, USA Background: Adherence is critical to achieve the benefits of antiretroviral therapy (ART). Smart-pill bottles (AdhereTech) securely transmit real-time adherence information via cellular networks to a central service that sends prompts to non- adherent patients by phone call or text in addition to on-device visual and audio cues. The smart-pill bottle service may improve adherence to ART. Methods: Adults with HIV taking a tenofovir disoproxil fumarate (TDF)- containing regimen with suboptimal adherence (2 detectable HIV RNA assays during the prior year) were randomized to receive adherence counseling with or without the smart-pill bottle service for 12 weeks. Tenofovir diphosphate (TFV-DP) levels by dried blood spot, HIV RNA, CD4 levels, and self-reported adherence (using the AIDS Clinical Trials Group [ACTG] Adherence Questionnaire) were collected. Results: 63 participants (22%women; 48% black, 25% Latino) were randomized (30 bottle, 33 control). At baseline, 49% of participants had HIV RNA <20 copies/mL and 61% reported 100% adherence with antiretroviral medications over the prior 4 days. From baseline to week 12, median TFV-DP levels increased from 1230 to 1887 fmol/punch in the smart-bottle group compared to a decrease from 1108 to 1048 fmol/punch in controls (see figure; median change +252 versus -41 fmol/punch, respectively, p=0.101). Discontinuation rates were 5 of 30 (17%) in the smart-bottle group vs. 7 of 33 (22%) in the control group (p=0.89). The number of participants with HIV RNA >20 copies/mL at baseline who decreased to ≤20 copies/mL at 12 weeks was 3 of 24 in the smart-bottle group vs. 7 of 26 in the control group (OR for the intervention 0.4; 95% CI 0.1, 2.0). The median change in CD4 count from

Pierce, FL, USA, 6 Care Resource Community Health Center, Miami, FL, USA, 7 George Washington University, Washington, DC, USA Background: Regimen complexity can adversely affect adherence, leading to virologic failure. It is unknown whether this occurs with regimens that contain the second-generation integrase inhibitors (INSTIs), bictegravir and dolutegravir, both of which are recommended in the current DHHS Guidelines. Methods: EMR, prescription and dispensing data for 2,217 patients initiating BIC/FTC/TAF, DTG/ABC/3TC, DTG+TDF/FTC, or DTG+TAF/FTC between Aug 2013 - Aug 2019 were collected from 5 practices across 17 US states. Only those without prior documented treatment with DTG or BIC, respectively, were included. Adherence was defined as proportion of days covered through the first 6 months of regimen treatment. To determine treatment effects on adherence, we (1) used multiple imputation with predictive posterior matching to account for incomplete baseline measures, (2) used mixed effects logistic regression, using BIC/FTC/TAF vs DTG-regimens with random intercept for practice, to adjust for heterogeneity between practices, (3) adjusted models using demographics and relevant baseline clinical data (CD4 count, viral load, AST, ALT, lipids, eGFR, hemoglobin A1C) and year of regimen initiation, and (4) employed propensity score matching using imputed baseline labs and demographics, allowing for squares and first order interactions between all included predictors. In addition to adherence, we assessed viral suppression (<200 copies/mL) in a subset of 655 patients at 6 months (measured within 1 week prior and up to two months after). Results: In observed (unadjusted) data, adherence was significantly greater at 6 months to BIC/FTC/TAF compared to any dolutegravir-regimen and to DTG/ABC/3TC in comparison to DTG+TDF/FTC or DTG+TAF/FTC at the 80% level [TABLE]. After controlling for non-treatment effects, adherence was only significantly different for BIC/FTC/TAF compared to DTG+TDF/FTC or DTG+TAF/ FTC (p<0.01). Assessment of viral suppression at 6 months for patients with measures (n=655) was favorably impacted by adherence ≥80% (OR 2.27 [1.26- 4.07] p<0.01) and ≥95% (OR 2.63 [1.55-4.48] p<0.01). Conclusion: This study of bictegravir and dolutegravir-based regimens supports the notion that simplifying treatment to a single tablet aids in adherence, and that adherence yields improved virologic outcomes in clinical settings.

Poster Abstracts

496 SOCIAL NORMS AND ART ADHERENCE: POPULATION-BASED STUDY OF PERSONS WITH HIV IN UGANDA Jessica M. Perkins 1 , Bernard Kakuhikire 2 , Charles Baguma 2 , Justin D. Rasmussen 3 , Carolyn M. Audet 1 , Mark J. Siedner 4 , Jessica E. Haberer 4 , David R. Bangsberg 5 , Alexander C. Tsai 4 1 Vanderbilt University, Nashville, TN, USA, 2 Mbarara University of Science and Technology, Mbarara, Uganda, 3 Duke University, Durham, NC, USA, 4 Massachusetts General Hospital, Boston, MA, USA, 5 Oregon Health and Sciences University, Portland, OR, USA Background: The extent to which certain health behaviors are perceived as normative is known to be an important determinant of one’s own propensity to engage in such behaviors. It is unknown, however, whether people living with

CROI 2020 176