Abstract

FRLBC05 - Oral Abstract Session

Combination interventions for the prevention of HIV among injection drug users: a complex systems dynamics model

Presented by Brandon Marshall (United States).

B. Marshall¹, M. Paczkowski², B. Tempalski³, E. Pouget³, S. Friedman³, S. Galea<sup>2

1</sup>Brown University, Department of Epidemiology, Providence, United States, ²Columbia University, Department of Epidemiology, New York, United States, ³National Development and Research Institues. Inc., Institute for AIDS Research, New York, United States

Background: Although combination prevention strategies are receiving growing attention, there is little evidence to inform their implementation, particularly for injection drug-using (IDU) populations. We constructed a complex systems dynamic model to assess various strategies for reducing HIV transmission among IDU.
Methods: We modeled HIV transmission in a dynamic network of IDU and non-IDU over a thirty-year time period (1992-2021). In the model, “agents” engage in risk behavior and interact with simulated prevention interventions (i.e., needle and syringe exchange programs [NSP], HIV testing, antiretroviral treatment [ART], substance abuse treatment). The model was constructed to represent the adult New York metropolitan statistical area (MSA) population, and calibrated by comparing HIV prevalence and incidence against historical validated MSA-level data. We obtained annualized incidence estimates from Monte Carlo simulations to examine the consequences of different intervention scenarios on a hypothetical population of 150,000.
Results: The model closely approximated published 1992-2011 data for HIV prevalence and incidence among IDU. Under current scenarios, HIV incidence among IDU residing in the New York MSA is estimated to be 3.7 per 1000 (95% CI: 1.0-6.3 per 1000) in 2021. Scenarios in which coverage of only one intervention was increased resulted in decreased HIV incidence at 2021, with expanded NSPs showing the lowest incidence rate (2.4 per 1000), followed by increased substance abuse treatment availability (2.8 per 1000), earlier initiation of HAART and improved adherence (2.9 per 1000), and increased access to HIV testing (3.5 per 1000). Combining all scenarios resulted in the largest absolute reduction in HIV incidence (1.7 per 1000, 95%CI: 0.0-3.9 per 1000) by 2021.
Conclusions: Our results demonstrate that combination interventions have the greatest potential to reduce HIV transmission among IDU. Although further research is required to determine cost-effectiveness, combining and bringing to scale existing evidence-based interventions may well be a highly effective strategy to reduce new infections.