High Levels of Prevention-effective Adherence to HIV PrEP

An Analysis of Substudy Data From the EPIC-NSW Trial

Benjamin R. Bavinton, PhD; Stefanie Vaccher, PhD; Fengyi Jin, PhD; Garrett P. Prestage, PhD; Martin Holt, PhD; Iryna B. Zablotska-Manos, PhD; Rebecca Guy, PhD; Janaki Amin, PhD; David J. Templeton, PhD; Barbara Yeung, MPH; Mohamed A. Hammoud, PhD; David Lewis, PhD; David Baker, MBBS; Nila Dharan, MD; Anna M. McNulty, MBBS; Andrew E. Grulich, PhD


J Acquir Immune Defic Syndr. 2021;87(4):1040-1047. 

In This Article


Study Design and Participants

The EPIC-NSW study protocol and primary results have been published.[12,14] Briefly, EPIC-NSW was a single-arm, prospective implementation trial of daily, single-dose, oral, coformulated tenofovir disoproxil fumarate (TDF)/emtricitabine (FTC) in individuals at high risk of HIV infection in which the primary analyses were to calculate cohort-level HIV incidence and state-wide reductions in new HIV notifications after rapid scale-up of PrEP. EPIC-NSW participants were enrolled at 31 clinical sites between March 2016 and April 2018 and followed up until March 31, 2019. Ethical approval was obtained from St Vincent's Hospital, Sydney. Participants were 18 years or older, lived in NSW or Australian Capital Territory (or visited to attend follow-up visits), had a negative fourth-generation HIV antibody/antigen test within 7 days before commencing PrEP, were at high and ongoing risk of HIV infection, and had an estimated glomerular filtration rate (eGFR) of more than 60 mL/min per 1.73 m2. As the primary at-risk population, the study was targeted toward GBM, although cisgender women and heterosexuals at high risk of infection were also able to enroll. At enrollment, all participants gave written consent to participate in the main study and were offered the opportunity to consent to participate in an optional behavioral substudy. The substudy was designed to measure additional demographics, PrEP use, sexual behavior, and drug use throughout follow-up.


Participants who consented to participate in the behavioral substudy received an email after their baseline EPIC-NSW visit and quarterly thereafter with a link to an online survey hosted on the SurveyGizmo platform. For each survey round, they received up to 2 reminder emails, 7 days apart for each survey. Participants could miss a survey but complete the next. Surveys took approximately 5–10 minutes to complete. There were 7 survey rounds including baseline.


Participants were asked about their sex assigned at birth, current gender identity, and current sexual identity; these data were used to classify participants as gay-identified cisgender men or not, and this determined which sexual behavior questions were asked. Country of birth was categorized into 6 regions: Australia, high-income English-speaking countries (United Kingdom, Ireland, New Zealand, Canada, and United States of America), Europe, Asia, Latin America and the Caribbean, and other. Postcode of residence was classified into 4 groups using a previously published method estimating the proportion of men residing in each Australian postcode who identified as gay[16] ("≥20% gay," "10%–19% gay," "5%–9% gay," and "<5% gay"); all postcodes classified as "≥20% gay" were in inner-city suburbs of Sydney, NSW. These data were collected at enrollment as part of the main study.

Quarterly MPR was calculated for each participant from PrEP dispensing logs at each clinic as part of the main study. These were reviewed to determine the date and number of pills dispensed. Each participant's time in the study was divided into 90-day intervals from the date of first dispensing, and for each period, MPR was calculated as the number of pills in the participant's possession divided by the number of days in the period, resulting in an MPR ranging from 0 to 1.

The following data were collected from behavioral substudy surveys. Participants were asked about PrEP use in the previous 3 months and the "last full week" (defined as Monday to Sunday). "Adequate PrEP protection" in the last full week was defined as ≥4 pills for participants assigned male sex at birth and ≥6 pills for participants assigned female sex at birth (including transgender men). Condomless sex (CLS) questions were presented to participants based on their sex assigned at birth, current gender identity, and current sexual identity; gay-identified cisgender men were asked about insertive and receptive condomless anal intercourse in the last full week, whereas other participants were asked about receptive or insertive anal and/or vaginal/front hole CLS in the last full week. If participants reported CLS on any day of the last full week, we termed this a "CLS-week." Both groups were asked to indicate whether they had CLS in the last full week with HIV-positive men with undetectable viral loads, HIV-positive men with detectable or unknown viral loads, HIV-negative men on PrEP, HIV-negative men not on PrEP (or unknown to be on PrEP), and unknown HIV status men. "Higher risk CLS" was defined as CLS with HIV-positive men with detectable/unknown viral loads or with unknown HIV status men. Although at the time of the study, only daily PrEP was recommended, participants were also asked about the ideal preference for PrEP dosing: "every day" (termed "daily PrEP"), "for periods of time when I feel I am at high risk of getting HIV" ("periodic PrEP"[17]), "only on specific occasions when I am at high risk of getting HIV" ("event-driven PrEP"[5]), and "other."

Statistical Analysis

This analysis aimed (1) to determine the proportion of weeks in which higher-risk CLS occurred but were inadequately protected by PrEP and (2) among weeks in which higher-risk CLS occurred, to determine factors associated with inadequate PrEP protection (defined as weeks in which <4 pills for participants assigned male sex at birth and <6 pills for participants assigned female sex at birth were taken). Participants were included if they completed the baseline survey and at least 1 follow-up behavioral survey. For each follow-up survey completed, only the last full week was examined; participants could complete multiple follow-up surveys (maximum = 6). If participants missed follow-up surveys, these data points were not included in the analysis. The last full week data from each survey were compared with the quarterly MPR of the period that the date of survey completion fell in. We describe the characteristics of the last full weeks by the following: total cumulative number of last full weeks reported by all participants, occurrence of any sex, occurrence of any CLS, occurrence of any higher risk CLS, and adequate or inadequate PrEP protection. Within the higher risk CLS weeks, we examined factors associated with inadequate PrEP protection using bivariate and multivariate generalized estimating equation modeling to control for inter- and intrasubject variability. Variables with P values of <0.1 in bivariate models were included in the multivariate model. We report odds ratios (OR), adjusted ORs (aOR), 95% CIs, and 2-sided P values for these associations. All analyses were conducted with Stata version 14.2 (StataCorp, College Station, TX).