Research Article: Estimation of the cost-effectiveness of HIV prevention portfolios for people who inject drugs in the United States: A model-based analysis

Date Published: May 24, 2017

Publisher: Public Library of Science

Author(s): Cora L. Bernard, Douglas K. Owens, Jeremy D. Goldhaber-Fiebert, Margaret L. Brandeau, Alexander C. Tsai

Abstract: BackgroundThe risks of HIV transmission associated with the opioid epidemic make cost-effective programs for people who inject drugs (PWID) a public health priority. Some of these programs have benefits beyond prevention of HIV—a critical consideration given that injection drug use is increasing across most United States demographic groups. To identify high-value HIV prevention program portfolios for US PWID, we consider combinations of four interventions with demonstrated efficacy: opioid agonist therapy (OAT), needle and syringe programs (NSPs), HIV testing and treatment (Test & Treat), and oral HIV pre-exposure prophylaxis (PrEP).Methods and findingsWe adapted an empirically calibrated dynamic compartmental model and used it to assess the discounted costs (in 2015 US dollars), health outcomes (HIV infections averted, change in HIV prevalence, and discounted quality-adjusted life years [QALYs]), and incremental cost-effectiveness ratios (ICERs) of the four prevention programs, considered singly and in combination over a 20-y time horizon. We obtained epidemiologic, economic, and health utility parameter estimates from the literature, previously published models, and expert opinion. We estimate that expansions of OAT, NSPs, and Test & Treat implemented singly up to 50% coverage levels can be cost-effective relative to the next highest coverage level (low, medium, and high at 40%, 45%, and 50%, respectively) and that OAT, which we assume to have immediate and direct health benefits for the individual, has the potential to be the highest value investment, even under scenarios where it prevents fewer infections than other programs. Although a model-based analysis can provide only estimates of health outcomes, we project that, over 20 y, 50% coverage with OAT could avert up to 22,000 (95% CI: 5,200, 46,000) infections and cost US$18,000 (95% CI: US$14,000, US$24,000) per QALY gained, 50% NSP coverage could avert up to 35,000 (95% CI: 8,900, 43,000) infections and cost US$25,000 (95% CI: US$7,000, US$76,000) per QALY gained, 50% Test & Treat coverage could avert up to 6,700 (95% CI: 1,200, 16,000) infections and cost US$27,000 (95% CI: US$15,000, US$48,000) per QALY gained, and 50% PrEP coverage could avert up to 37,000 (22,000, 58,000) infections and cost US$300,000 (95% CI: US$162,000, US$667,000) per QALY gained. When coverage expansions are allowed to include combined investment with other programs and are compared to the next best intervention, the model projects that scaling OAT coverage up to 50%, then scaling NSP coverage to 50%, then scaling Test & Treat coverage to 50% can be cost-effective, with each coverage expansion having the potential to cost less than US$50,000 per QALY gained relative to the next best portfolio. In probabilistic sensitivity analyses, 59% of portfolios prioritized the addition of OAT and 41% prioritized the addition of NSPs, while PrEP was not likely to be a priority nor a cost-effective addition. Our findings are intended to be illustrative, as data on achievable coverage are limited and, in practice, the expansion scenarios considered may exceed feasible levels. We assumed independence of interventions and constant returns to scale. Extensive sensitivity analyses allowed us to assess parameter sensitivity, but the use of a dynamic compartmental model limited the exploration of structural sensitivities.ConclusionsWe estimate that OAT, NSPs, and Test & Treat, implemented singly or in combination, have the potential to effectively and cost-effectively prevent HIV in US PWID. PrEP is not likely to be cost-effective in this population, based on the scenarios we evaluated. While local budgets or policy may constrain feasible coverage levels for the various interventions, our findings suggest that investments in combined prevention programs can substantially reduce HIV transmission and improve health outcomes among PWID.

Partial Text: Over the past decade, injection drug use, particularly heroin injection, has increased across most US demographic groups, making substance-abuse-related mortality and morbidity a public health crisis [1]. In 2014, there were 47,055 deaths from drug overdose in the US, with almost 30,000 due to opioid overdose [2]. Because HIV spreads relatively efficiently through the transfer of blood in shared injecting equipment [3], people who inject drugs (PWID) account for a disproportionate share of HIV prevalence and incidence in the US [4,5]. Although HIV prevalence and incidence among US PWID have been falling over the past decade [4,6,7], recent growth in the size of the injecting population has raised concerns that HIV risks could rise [8]. Programs targeted to PWID, which have the additional benefit of preventing downstream sexual transmission of HIV to others in the population, are therefore a public health priority. Given the current epidemic of injection drug use in the US, benefits that extend beyond HIV prevention are also a critical consideration [1,8].

The opioid epidemic is a global public health burden that has become particularly acute in the US [1,2,8]. In addition to the substantial mortality associated with substance abuse [2], high rates of HIV transmission among PWID make the successful prevention of HIV in this population a public health priority. To that end, we consider portfolios of HIV prevention programs that include OAT, NSPs, Test & Treat, and PrEP scaled to various coverage levels. Although model projections can only provide estimates of health benefits and costs, such analyses can provide intuition around critical mechanisms and assumptions to inform decision making. Our main finding is that, over 20 y, high coverage (enrollment of 50% of the eligible population) of OAT, NSPs, and Test & Treat in combination could avert nearly 43,400 (95% CI: 23,000, 74,000) HIV infections among PWID and reduce HIV prevalence among PWID by 27% (95% CI: 12%, 45%). The construction of such a portfolio has the potential to be cost-effective at each incremental expansion, with projected ICERs below US$50,000 per QALY gained. Moreover, our analysis suggests that the estimated benefit obtainable by PrEP alone (measured in QALYs) could potentially be achieved and even surpassed at substantially lower cost by combining other prevention interventions into high-value portfolios.



Leave a Reply

Your email address will not be published.