CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

applied the new testing criteria to the validation dataset to determine the number of tests performed, test positivity, proportion of positives missed, and costs averted. Costs were derived based on total budgets allocated to I-TECH to support HIV testing and estimated costs of test procurement. Results: The analysis included 262,230 tests of which 4.3%were HIV positive. Model derivation analysis identified ages 23-29, 30-39, and 40+, non- citizenship, and emergency department testing as significantly associated with positivity. Among 131,115 tests in the validation analysis, 5,580 (4.3%) were HIV positive. Restricting testing to persons age >30 years and other defined criteria would reduce testing volume by 23% and increase positivity to 4.9%; 649 (2.1%) of the 30,178 persons who would not be tested were HIV positive representing 11.6% of all positive tests in the validation dataset. Positives missed by the criteria had a median age of 25 years and were mostly female (67%) and tested in the general outpatient department (86%). Assuming no changes in staffing, implementing the new testing criteria would decrease total HIV testing costs by 13%, a savings of $18 per positive test missed. Conclusion: In Botswana, a targeted approach to HIV testing could reduce testing volumes by 23% and modestly increase HIV test positivity while missing 11.6% of positive tests. Cost saving would be modest unless implementation was accompanied by changes in staff costs.

255–569), the median number of tests was 2 (IQR 2–3), and the median interval between tests was 255 days (IQR 198–325). The diagnostic yield among MSM or TGW of color who were frequent testers was 1.2% (34/2846) and among non-frequent testers 1.0% (13/1281). Among all other THRIVE clients, the yield was 0.2% (12/6056). Conclusion: The diagnostic yield was similar for MSM and TGW of color who were tested frequently or non-frequently, but frequent testing was associated with a shorter time to diagnosis. These data support the CDC recommendation to test persons at risk of HIV more often than annually.

950 INDETERMINATE HIV RAPID-TEST RESULTS: OUTCOMES AND RISK FACTORS George Mwinnyaa 1 , Mary K. Grabowski 2 , Ronald H. Gray 1 , Maria Wawer 1 , Larry W. Chang 2 , Joseph Ssekasanvu 1 , Joseph Kagaayi 3 , Godfrey Kigozi 3 , Ronald M. Galiwango 3 , Anthony Ndyanabo 3 , David Serwadda 3 , Thomas C. Quinn 4 , Steven J. Reynolds 4 , Oliver Laeyendecker 5 , for the The Rakai Community Cohort Study 1 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA, 2 Johns Hopkins University School of Medicine, Baltimore, MD, USA, 3 Rakai Health Sciences Program, Kalisizo, Uganda, 4 NIH, Bethesda, MD, USA, 5NIAID, Baltimore, MD, USA Background: Little is known about the frequency, subsequent outcomes and factors associated with indeterminate HIV rapid results. We assessed final HIV serological outcomes for individuals with rapid indeterminate test results and associated risk factors in Rakai, Uganda. Methods: 54,469 HIV rapid test results, defined by two parallel rapid tests, among 31,413 participants aged 15-49 years in the Rakai Community Cohort Study were assessed. 8361 participants were tested on two separate visits and 7354 had three time points tested. Each visit was approximately 18 months apart. Indeterminate results were defined as contradictory rapid test results or inconclusive concordant rapid test results. The final HIV status for each indeterminate observation was determined using previous HIV status information and additional testing, including PCR, ELISA and Western blot when necessary. Generalized estimating equations together with modified Poisson regression models with robust variance were used to assess prevalence ratios (PRs) of subsequent HIV serological outcomes and factors associated with indeterminate rapid test results. Results: The prevalence of HIV rapid test indeterminate results was 2.7% (1490/54,469). Of the 1,490 rapid indeterminate observations, 26%were eventually classified as HIV positive. The proportions of persons with rapid indeterminate results progressing to HIV rapid positive, negative, or still indeterminate at the subsequent visit were 19%, 40% and 41%, respectively. For individuals with two consecutive indeterminate results who had a third follow-up visit (67 individuals), 21% (14/67) tested negative, 9% (6/67) were positive and 70% (47/67) were still indeterminate. Factors associated with higher risk of an indeterminate result were: women vs. men (adjPR 2.07, 95% CI 1.77,2.41); >44 vs. <20 years of age (adjPR 1.69, 95% CI 1.26,2.26); student vs. farmer (adjPR 0.62, 95%CI 0.46, 0.83); shopkeeper vs. farmer (adjPR 0.81, 95%CI 0.68, 0.96); ART vs. not (adjPR 1.29, 95% CI 1.10,1.51). In total, 4.4% of individuals on ART had indeterminate test results. Conclusion: The frequency of indeterminate rapid results was low (<3%), and a quarter ultimately tested HIV positive. 41% of individuals with an indeterminate

Poster Abstracts

949 FREQUENT HIV TESTING OF MSM AND TGW OF COLOR RESULTS IN EARLIER DIAGNOSIS Karen W. Hoover 1 , Weiming Zhu 1 , Kenneth L. Dominguez 1 , Kirk D. Henny 1 , Ya-Lin A. Huang 1 , Kashif Iqbal 1 , Mary Tanner 1 , Kevin P. Delaney 1 1 CDC, Atlanta, GA, USA Background: Few clinical studies exist to support recommendations for more frequent than annual HIV testing of persons at increased risk for HIV. Frequent testing provides more opportunities for PrEP counseling and initiation, and earlier diagnosis of HIV and initiation of ARV medications to preserve immune function and prevent HIV transmission. We studied the effect of HIV testing frequency on time to diagnosis and yield of testing among MSM and TGW of color in the THRIVE demonstration project. Methods: We analyzed a longitudinal database that included HIV tests and results for a cohort of persons enrolled in THRIVE from September 2016 to March 2019. All MSM and TGW of color in THRIVE were at increased risk for HIV. We excluded those who were PrEP users. Among persons who had an initial negative HIV test and at least one additional test, we estimated the median number of HIV tests and conducted Kaplan-Meier analyses to determine the time to diagnosis since an initial negative HIV test. We defined frequent testing as a mean interval between tests of ≤180 days and non-frequent testing as >180 days. We estimated the yield of HIV testing as the number of new diagnoses per tests performed. All results were stratified by testing frequency. Results: In THRIVE, 20,956 clients received an HIV test. Of these, 26% (5408) had an initial negative test and at least one additional test. Among these 5408 persons, 1338 were MSM or TGW of color who did not use PrEP and 47 (4%) had a subsequent positive test. Overall, the median time to diagnosis was 235 days (IQR 92–364). Frequent testers were diagnosed earlier than non-frequent testers (p<0.001) (Figure). Among 34 frequent testers, the median time to diagnosis was 120 days (IQR 83–278), the median number of tests was 3 (IQR 2–4), and the median interval between tests was 84 days (IQR 53–119). Among 13 non-frequent testers, the median time to diagnosis was 364 days (IQR

result had an indeterminate result on a follow-up visit and 0.64% of the population had continuous indeterminate results over a 3 year period.

CROI 2020 356