CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

908 PROFILES OF HIV CARE DISRUPTIONS IN ZAMBIA: A LATENT CLASS ANALYSIS Aaloke Mody 1 , Kombatende Sikombe 2 , Sheree Schwartz 3 , Laura K. Beres 3 , Ingrid Eshun-Wilson 4 , Sandra Simbeza 2 , Njekwa Mukamba 2 , Carolyn Bolton Moore 2 , Izukanji Sikazwe 2 , Charles B. Holmes 5 , Nancy Padian 6 , Elvin Geng 1 1 University of California San Francisco, San Francisco, CA, USA, 2 Centre for Infectious Disease Research in Zambia, Lusaka, Zambia, 3 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA, 4 Stellenbosch University, Cape Town, South Africa, 5 Georgetown University, Washington, DC, USA, 6 University of California Berkeley, Berkeley, CA, USA Background: Beyond observed traits (e.g., sex, age), there may also be unobserved (i.e., “latent”) traits–each leading to distinct profiles of barriers to care–that influence retention in care. We used latent class analysis (LCA) of patient-reported reasons for HIV care disruptions in Zambia to identify these patient profiles and examine their associations with engagement in care. Methods: We traced a probability sample of patients lost to follow-up (LTFU, >90 days late for last visit) as of July 31, 2015 from 64 clinics in Zambia. Among those found alive, we used a semi-structured instrument to identify patient-reported reasons for care disruptions. We performed LCA–incorporating sampling weights–to identify patient subgroups based on the number and types of patient-reported reasons for care disruptions. We characterized patient characteristics for each class and used logistic regression to assess the association between class membership and updated care status (disengaged vs. silent transfer to a new site). Results: We successfully traced 642 patients LTFU (59.2% female; median age 35y [IQR 30-41]; median enrollment CD4 236 cells/μl [IQR 124-368]). We identified five classes of care disruptions (Table): 1) the “livelihood and mobility” class (29.9% of sample) reported work obligations and mobility/ travel as reasons for their care disruptions; 2) the “mobility and family” class (27.5%) were likely to report mobility/travel, family obligations, and transport; 3) the “doubting need for HIV care” class (8.4%) reported care disruptions due to beliefs about their needs for HIV care; 4) the “clinic accessibility” class (25.1%) were likely to report transport-, clinic-, and disclosure-related care challenges; and 5) the “multidimensional barriers” class (9.2%) reported multiple reasons (mean 5.5) across categories. The “mobility and family” class was least likely to be disengaged (19.2% disengaged vs. 80.8% silent transfer), followed by the “livelihood and mobility” (44.1%), “clinic accessibility” (48.8%), and “multidimensional barriers” (57.2%) classes, with the “doubting need for HIV care” class most likely to be disengaged (100%). Conclusion: There are distinct profiles for HIV care disruptions that are associated with whether a patient disengages or silently transfers their care. Strategies to target these unique patient profiles by concurrently addressing multiple barriers, rather than individual barriers, may be a more effective way to design and implement interventions to improve retention in care.

907 ANTIRETROVIRAL ADHERENCE AND HIV-1 DRUG RESISTANCE IN THE US Carmela Benson 1 , Laura Mesana 2 , Keith J. Dunn 1 , Xiaotian Wu 3 , Nicole Li 2 , Johnny Lai 4 , Eric Y. Wong 1 , Kimberley Brown 1 1 Janssen Scientific Affairs, LLC, Titusville, NJ, USA, 2 Amaris, Jersey City, NJ, USA, 3 Brown University, Providence, RI, USA, 4 Monogram BioSciences, San Francisco, CA, USA Background: Adherence to antiretroviral therapy (ART) is critical to achieving viral suppression. However, social determinants of health (SDoH) can undermine patient adherence to ART, which can result in drug resistance that compromises future treatment options. We investigated SDoH factors and their impact on ART adherence, resistance and other HIV-related measures (prevalence, mortality, viral suppression). Methods: Rates of HIV-related measures and SDoH (age, gender, education, poverty, employment) from publicly available databases during the period of 2014-2016 reported for each state in the US (N=50) were collected. ART adherence was measured by the average proportion of days covered (PDC) of all patients per state using the Symphony Health Solutions claims database (N=165K). Poor adherence was defined as PDC < 80%. Separately, isolates submitted to Monogram Biosciences for routine clinical testing from 2015-2017 (N=95,956) were used to determine rates of resistance, defined as proportion of isolates with a genotypic assessment for resistance to any commercially available NRTI, NNRTI, PI, and/or INI. Exploratory inferential analyses were performed to investigate associations between SDoH, HIV-related measures, adherence, and resistance, using correlation analysis. Results: Rates of poor adherence ranged from 26% to 55% [median=44%] and resistance rates ranged from 20% to 54% [median=30%]. Adherence and resistance were both significantly correlated with gender and HIV prevalence (p≤0.05). States where poor adherence was more prevalent had a higher percentage of low education level, households living below poverty level, and unemployment rates; states with higher prevalence of poor adherence were also those with higher HIV prevalence and mortality rates and lower viral suppression rates (Table). States with higher resistance rates had higher HIV prevalence (Table). Conclusion: Nationally, poor adherence and resistance rates exceeded 20%. State-level data showed gender, race, education level, poverty, and employment were associated with poor adherence to ART, and gender was associated with resistance to ART. Adherence was also correlated to HIV mortality and prevalence rates, whereas resistance was correlated to prevalence rates. Based on these results, patients could benefit from HIV treatment that is simple, convenient, and has a high genetic barrier to resistance.

Poster Abstracts

CROI 2019 354