CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

1145 DURABILITY OF FINANCIAL INCENTIVES EFFECT ON VIRAL SUPPRESSION AND CONTINUITY IN CARE Wafaa M. El-Sadr 1 , Geetha Beauchamp 2 , Irene Hall 3 , Lucia V. Torian 4 , Barry S. Zingman 5 , Garret Lum 6 , Richard A. Elion 7 , Kate Buchacz 3 , David Burns 8 , Allison Zerbe 1 , Theresa Gamble 9 , Deborah J. Donnell 2 1 ICAP at Columbia University, New York, NY, USA, 2 Fred Hutchinson Cancer Research Center, Seattle, WA, USA, 3 CDC, Atlanta, GA, USA, 4 New York City Department of Health and Mental Hygiene, Long Island City, NY, USA, 5 Montefiore Medical Center, Bronx, NY, USA, 6 District of Columbia Department of Health, Washington, DC, USA, 7 George Washington University, Washington, DC, USA, 8 NIAID, Bethesda, MD, USA, 9 FHI 360, Durham, NC, USA Background: There is increased interest in use of financial incentives to achieve desired health, including HIV-related, behaviors. The HPTN 065 study demonstrated that financial incentives (FI) were associated with 3.8% [(0.7%- 6.8%), p=0.014) higher viral suppression (VS) and with 8.7% [(4.2%,13.2%), p=0.0001] higher continuity in care (CC) among patients at sites randomized to FI versus (vs) standard of care (SOC) in the Bronx, NY (BNY) and Washington, DC (DC). Whether the effects of FI are durable beyond withdrawal of FI is unclear. We assessed VS and CC at FI versus SOC sites post-intervention to determine durability of FI. Methods: A total of 37 (20 BNY/ 17 DC) care sites with 51,782 patients in care (28,439 BNY/23,343 DC), were site-randomized to FI or SOC. At FI sites, patients on ART could earn $70 gift card per quarter with VS. Lab data reported to HIV Surveillance were used for site-level outcomes: for VS, VL less than 400 copies/ ml in engaged patients (≥2 visits in last 15 months); for CC, CD4 or VL in 4 of prior 5 quarters. Post-intervention effects were assessed for the three quarters after discontinuation of FI (Apr-Dec 2013). GEE was used to compare FI and SOC site-level outcomes during the FI intervention and post-intervention (Table). Results: Post-intervention, a trend remained for an increase in VS by 2.7% (-0.3%, 5.6%, p=0.076) was noted at FI vs SOC sites. This difference in VS between FI and SOC sites was reduced from the 3.8% increase in VS to 2.7%, but was persistent nonetheless. Notably, in the subgroups of sites where FI achieved a significant increase in VS during the intervention, we noted a reduced but durable effect post-intervention at FI vs SOC sites: at DC sites 4.4% higher (p=0.057), at hospital-based sites 4.8% higher (p=0.003) and at sites with high baseline VS 3.2% higher (p=0.066). The significant increase in CC during FI intervention was sustained post-intervention with 7.5% (p=0.007) higher CC at FI vs SOC sites. A durable significant effect of FI post-intervention on CC persisted at sites randomized to FI vs SOC in BNY (p=0.010), at hospital-based sites (p=0.019) and at sites with higher baseline VS (p=0.014). Conclusion: Post discontinuation of FI, data from this large study showed evidence of durable effects of FI, both on VS and CC, at sites that were previously randomized to FI vs SOC. These findings suggest that behaviors motivated by FI may last beyond the provision of the FI, increasing the potential cost- effectiveness of FI strategies.

1144 DECREASED ALCOHOL USE (EVEN WITHOUT ABSTINENCE) IS ASSOCIATED WITH BETTER VIRAL LOAD Heidi M. Crane 1 , Robin M. Nance 1 , Andrew Hahn 1 , Judith I. Tsui 1 , Michael J. Mugavero 2 , Bryan Lau 3 , Geetanjali Chander 3 , Sonia Napravnik 4 , Karen Cropsey 2 , Katerina A. Christopoulos 5 , Rob Fredericksen 1 , Dennis Donovan 1 , Jane M. Simoni 1 , Mari Kitahata 1 , Joseph Delaney 1 1 University of Washington, Seattle, WA, USA, 2 University of Alabama at Birmingham, Birmingham, AL, USA, 3 Johns Hopkins University, Baltimore, MD, USA, 4 University of North Carolina Chapel Hill, Chapel Hill, NC, USA, 5 University of California San Francisco, San Francisco, CA, USA Background: Alcohol use is common among people living with HIV (PLWH) and associated with poor antiretroviral treatment (ART) adherence and detectable viral load (VL). Interventions for hazardous alcohol use exists; however, many PLWH may moderate their use but not abstain. We conducted this study to examine the potential impact of decreasing alcohol use on VL without abstinence and how this differs based on alcohol use patterns. Methods: We used data from 7 U.S. sites in the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) cohort. Eligible PLWH completed the CNICS clinical assessment including alcohol use on the AUDIT-C on or after 2010, reported alcohol use, and had 2 VL measures collected as part of clinical care. We examined frequency of alcohol use, frequency of binge drinking, and alcohol use severity (AUDIT-C score). Linear mixed models with time-updated alcohol use and VL were used to examine associations between changes in alcohol use and VL (log 10 transformed) adjusting for age, sex, race, frequency of illicit substance use by individual drug category, and calendar year. Models were repeated, stratified by Hepatitis C virus (HCV) status. Results: Among 7137 PLWH who drank alcohol there were 61,315 VL measures, mean baseline VL was 22,709 copies/mL (geometric mean 118) and 71%were undetectable (<100 copies/mL). Stopping alcohol use was associated with decreased VL for all alcohol measures (p values<0.05). Decreased alcohol use among those who continued to drink (not abstinent) was associated with lower VL for all 3 alcohol measures. Compared to those who did not decrease alcohol, those who decreased alcohol frequency had a mean 18% lower VL (95% confidence interval (CI) 11%-24%, p <0.001), those who decreased their binge drinking frequency had 26% lower VL (95% CI 15%-36%, p <0.001), and those who decreased their AUDIT-C score had 26% lower VL (95% CI 21%-31%, p <0.001). Even a 1-point AUDIT-C score decrease was significant. Impacts were attenuated among PLWH with HCV. Conclusion: We demonstrated alcohol cessation was associated with decreased VL. In addition, decreasing alcohol use without abstinence was associated with a lower VL, which could lead to improved health outcomes and public health benefits in terms of decreased transmissibility. The decreased VL could be via improved ART adherence or more direct biological effects of alcohol. This suggests that supporting decreased alcohol use could help patients achieve VL goals regardless of achieving abstinence.

Poster Abstracts

CROI 2018 443