CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

901 USING SURVEILLANCE DATA TO MEASURE TRIAL HIV INCIDENCE OUTCOMES: A MODELLING STUDY

San Diego, CA, USA, 5 University of Washington, Seattle, WA, USA, 6 University of California San Francisco, San Francisco, CA, USA, 7 University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, 8 University of Alabama at Birmingham, Birmingham, AL, USA Background: Missed HIV care provider visits are associated with increased mortality beyond the core indicators of retention in care and are an immediately actionable event. Previous prediction models for missed visits have not incorporated data beyond the individual level. Methods: We developed prediction models for missed HIV care provider visits among adult people living with HIV (PLWH) with ≥1 visit in the Center for AIDS Research Network of Integrated Clinical Systems (CNICS) from 2010-2016. Potential predictors were identified at the individual-, community-, and structural-levels. Individual-level data included demographics, patient- reported outcomes (tobacco use, AUDIT-C, patient health questionnaire-9, EuroQOL Health Related Quality of Life-5D, HIV symptom index), insurance type, and prior visit adherence. Community-level data were obtained from the American Community Survey using ZIP Code tabulation area of residence. Structural-level data included HIV criminalization laws, Medicaid expansion, and proportion of budget dedicated to AIDS Drug Assistance Programs by state of residence. Variables were selected and models fit using random forests and 10-fold cross-validation; candidate models with highest area under the curve (AUC) were identified. Results: Data from 382,548 HIV care provider visits among 20,889 PLWH followed for a median of 3.7 years were included. Median age was 44 years, 81%were male, 37%were Black non-Hispanic, and 57% reported male-to-male sexual contact as HIV transmission risk factor. Prior visit adherence improved discrimination most in all models; AUC jumped from 0.68 to 0.75 with its addition alone in one candidate model. The highest AUC was 0.75 (Table); the strongest predictors in this model were prior visit adherence, follow-up time, age, and CD4+ count at the individual-level, along with proportion with Black race, proportion unemployed, and proportion living below the poverty line at the community-level. Conclusion: Prediction models validated using multi-level data in a population representative of US PLWH had a similar AUC to previous models developed using only individual-level data. Strongest predictors were individual-level variables, particularly prior visit adherence, though community-level variables were also predictive. Absent additional behavioral, social, structural, or clinical data, PLWH with previous visits should be targeted by interventions to improve visit adherence. 903 THE CD4 DEPLETION MODEL DOES NOT DIFFERENTIATE INCIDENT FROM CHRONIC INFECTION Michael E. Tang 1 , Sanjay R. Mehta 1 , Susan J. Little 1 , Christy M. Anderson 1 1 University of California San Diego, San Diego, CA, USA Background: The Ending the HIV Epidemic (EtHE) initiative targets a 75% decline in HIV incidence in 5 years and a 90% decline in 10 years. Available estimates for U.S. population incidence are derived from a CD4 depletion model developed by investigators at the Center for Disease Control. We evaluated this model in a cohort with well characterized estimated dates of incident infection. Methods: We evaluated 702 antiretroviral (ART)-naïve, newly HIV-1 diagnosed individuals with acute and early HIV infection enrolled to the San Diego Primary Infection Resource Consortium (PIRC) between June 1996 and July 2019. Clinical data, including CD4 counts were collected at baseline (Day 0), weeks 4, 12, 24, and every 24 weeks thereafter. Persons with acute infection (antibody neg/HIV nucleic acid test pos) had additional CD4 and VL measures at weeks 2 and 8. We calculated an estimated date of infection (EDI) using previously characterized serologic and virologic criteria. We compared this PIRC EDI with the EDI generated by the CD4 model. Results: Of the 702 newly HIV diagnosed individuals (age 16-71), 234 (33.3%) were diagnosed during acute infection, 468 (66.7%) during recent infection; 90.8% estimated by limiting-antigen (LAg) avidity assay in combination with viral load information (PIRC EDI model) and 9.2% by interval HIV seroconversion

Kate M. Mitchell 1 , Dobromir Dimitrov 2 , James Hughes 3 , Eric Vittinghoff 4 , Albert Y. Liu 4 , Chris Beyrer 5 , Deborah J. Donnell 2 , Marie-Claude Boily 1 1 Imperial College London, London, UK, 2 Fred Hutchinson Cancer Research Center, Seattle, WA, USA, 3 University of Washington, Seattle, WA, USA, 4 University of California San Francisco, San Francisco, CA, USA, 5 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA Background: Cluster-randomized trials (C-RCTs) are expensive to conduct. Using surveillance data on new HIV diagnoses instead of measuring incidence in the trial could reduce costs. We used mathematical models to evaluate when surveillance data can be used to estimate impact in HIV intervention C-RCTs. Methods: We used a model of HIV transmission among men who have sex with men in Baltimore, US, to simulate C-RCTs scaling up antiretroviral therapy (ART), pre-exposure prophylaxis (PrEP) and HIV testing in combination or alone. We tested whether modelled reductions in total cumulative HIV diagnoses predict model cumulative HIV incidence reduction over a 2-year trial. We also tested if reductions in diagnoses predict incidence reduction better over a longer trial duration (≤4 yrs) or when measured in later trial years. We explored if reductions in other surveillance measures – diagnoses with acute infection, diagnoses with early (CD4>500 cells/µl) infection, or diagnoses adjusted for testing volume – better predict incidence reduction. We used Pearson correlation to assess precision and report bias and sensitivity to detect a true incidence reduction. Results: Over a 2-year trial expanding ART+PrEP+testing, model results suggest total diagnosis reductions correlate poorly with incidence reduction(r=0.386), underestimate incidence reduction (by 97%), and have 52% sensitivity (Table). Precision and sensitivity were better in trials expanding ART(r=0.878; sens 100%) or PrEP(r=0.960; sens 88%) alone, but bias remained (-52% for ART, -55% for PrEP). In trials expanding testing alone, diagnoses increased with decreasing incidence(r=-0.915). Measuring impact in longer trials or over later years improved correlations between diagnosis and incidence reductions for trials expanding ART+PrEP+testing, up to r=0.795 over the 4th year, and reduced bias. For ART+PrEP+testing trials, reductions in acute, early or adjusted diagnoses correlate poorly(r<0.51) with incidence reduction. Reductions in acute or early diagnoses correlate sufficiently with incidence reduction only when ART alone is expanded(r=0.993, r=0.953, respectively), but are biased (-18%, -41%). Conclusion: Modelling results suggest that surveillance diagnoses data can only rarely be used to estimate C-RCT HIV incidence reductions. Reductions in acute/early or total diagnoses may be adequate predictors in C-RCTs expanding ART or PrEP alone if bias can be adjusted for. None of the diagnoses markers explored were appropriate for C-RCTs expanding HIV testing.

Poster Abstracts

902 PAST BEHAVIOR OUTPERFORMS DEMOGRAPHY AND GEOGRAPHY AS PREDICTOR OF MISSED HIV CARE April Pettit 1 , Aihua Bian 1 , Cassandra Oliver 1 , Peter F. Rebeiro 1 , Bryan E. Shepherd 1 , Jeanne C. Keruly 2 , Kenneth H. Mayer 3 , W. C. Mathews 4 , Richard D. Moore 2 , Heidi M. Crane 5 , Elvin Geng 6 , Sonia Napravnik 7 , Mari M. Kitahata 5 , Michael J. Mugavero 8 , for the Center for AIDS Research Network of Integrated Clinical Systems (CNICS) 1 Vanderbilt University, Nashville, TN, USA, 2 Johns Hopkins University, Baltimore, MD, USA, 3 Fenway Health, Boston, MA, USA, 4 University of California San Diego,

CROI 2020 338