Our study aimed to assess the population-wide effects of expanded HIV treatment and screening on the HIV epidemic in the U.S. Although prior studies have addressed the effectiveness and cost-effectiveness of either expanded HIV screening (21
) or treatment (92
), those analyses were not designed to fully evaluate how such programs would influence HIV transmission in the overall population or the course of the epidemic. Our study is the first to evaluate the population-wide effects (new HIV infections and other health outcomes) and the cost-effectiveness of alternative combinations of HIV screening and treatment in the U.S.
Our study has several key findings. First, we find that expanding HIV screening and treatment could prevent 200,000 to 300,000 infections over 20 years, or approximately 17% to 24% of new infections, adding up to 6.8 million quality-adjusted life years to the population. To prevent 24% of new infections, routine HIV screening would need to occur annually, with antiretroviral treatment available for essentially all symptomatic patients. Our analysis assumes that people identified as HIV-infected reduce risk behaviors by 20%; even with modest reductions in risk behavior, expansion of screening and treatment would provide enormous health benefit. If HIV-infected people reduce risk behavior further as some studies suggest (7
), the health benefit could be substantially higher than we have estimated. Annual HIV screening and counseling leading to 50% behavior reduction in infected individuals, along with 90% ART initiation in symptomatic patients, reduces new infections to fewer than 35,000 per year. Even under such optimistic assumptions, the U.S. HIV epidemic is unlikely to be completely eliminated without additional preventive measures.
Second, our analysis highlights the importance of emphasizing risk behavior reduction as HIV screening and treatment becomes increasingly available. For example, in addition to expanded screening and treatment, a 50% reduction in sexual risk behaviors among MSM and needle-sharing among IDUs could prevent 65% of new infections, reducing HIV incidence to approximately 20,000 cases per year. This suggests that programs to reduce risk behavior among high-risk individuals are likely to be a key component of a successful prevention program. If, however, uninfected individuals increase risk behavior post-screening, some of the benefits would be attenuated.
Our third finding is that the net benefit of implementing both interventions is greater than the sum from implementing each program individually. A substantial increase in HIV screening or treatment could prevent 95,000 or 198,000 new infections, respectively, whereas a combination program could avert 300,000 infections (a gain of 7,000 infections prevented or 2%). Programs to expand screening and treatment will be most effective if they are implemented together, because this realizes the complementary
effects of each program. Essential to achieving these levels of infections averted is patient receipt of test results following diagnosis, as well as linkage to care, which has been shown to improve with nurse-initiated counseling (95
) and health worker follow-up interviews with newly diagnosed patients (96
The effectiveness of screening and counseling in reducing sexual activity will likely vary within healthcare settings due to differences in risk behavior, and the length, content, and intensity of counseling services. Even with no reduction in risk behavior, one-time screening of low-risk groups and annual screening of high-risk groups could prevent nearly 4% of new infections, by identifying infected individuals and linking them to treatment programs. Augmenting this strategy with expanded ART prevents 16% of new infections. This suggests that preventing future infections through increased ART becomes increasingly important as the effectiveness of screening and counseling diminishes. In settings where counseling is unavailable or ineffectual, increased utilization of ART can help ensure that expanded screening will lead to reductions in HIV transmission.
Finally, we find that expanded utilization of ART (to 75% or 90% initiating ART at CD4 <350 cells/mL) is very cost-effective, as is one-time screening of low-risk groups and annual screening of high-risk groups. Combination strategies prevent more HIV infections and increase QALYs more than either individual strategy. As noted, our analysis specifically accounts for the effect of combination screening and treatment on population-wide HIV transmission, which is a strength of our modeling framework. As routine HIV screening for adults increases across healthcare settings due to recently revised CDC guidelines (3
), it is important to ensure that ART utilization increases at a concomitant rate. Further expanding HIV screening and counseling services, without expanding the proportion of infected individuals on ART, fails to realize the potential benefits of implementing these two complementary interventions.
Compared to other disease screening programs in the United States, one-time HIV screening of low-risk individuals and annual screening of high-risk individuals is economically attractive with a cost-effectiveness ratio less than $23,000/QALY gained. This compares favorably with other accepted interventions, including screening for type-2 diabetes (97
) and breast cancer mammography (98
Our study has several limitations. First, we assumed proportional mixing among sexual partners and needle-sharing contacts, which simplifies the complex network structure of partnership formation and dissolution. Second, although we stratified the population according to gender and risk behavior, we did not include variations by race or ethnicity. To fully account for such granularity, we would need to accurately estimate sexual and needle-sharing behavior within and between races, which would be difficult to obtain. Moreover, significant disparities in treatment rates, background mortality, and co-morbidities exist, and our model cannot account for these additional factors. A third limitation of our study is the omission of acute HIV screening, which would require a different model structure and specific assumptions about the benefits and costs associated with identification and treatment of acute HIV infection. The degree to which acute infection contributes to transmission is uncertain and estimates vary (47
). Fourth, we used a simplified HIV treatment model that does not include the intricacies of individual HIV disease management, drug toxicities, CD4 monitoring, or the presence of co-morbidities such as coronary heart disease, diabetes, and various cancers. Our results, however, are broadly consistent with those from more complicated models of HIV disease progression (21
). Finally, we did not explicitly model development of resistance to ART, although we believe our assumptions about the benefits of ART are conservative given the introduction of new classes of antiretroviral therapy, such as integrase inhibitors and entry inhibitors, and we evaluated scenarios that included resistance in sensitivity analyses.
Expanded HIV screening and counseling in the United States can prevent a substantial number of new HIV infections, adding millions of QALYs to the population. Programs that simultaneously expand antiretroviral therapy utilization can prevent more HIV infections than expanding either intervention alone. Our analysis indicates that over the next two decades, HIV incidence in the U.S. could be reduced by 24% with a comprehensive expansion of screening and treatment. If these programs are accompanied by additional interventions that halve risky sexual and needle-sharing behavior, the epidemic could be reduced by 65%, suggesting the need for a comprehensive portfolio of HIV prevention, screening, and treatment.