The Clinical and Economic Impact of Attaining National HIV/AIDS Strategy Treatment Targets in the United States

We project the clinical and economic outcomes of achieving National HIV/AIDS Strategy viral suppression targets for human immunodeficiency virus–infected US persons and black men who have sex with men, which would yield substantial clinical benefits and be very cost-effective.

Abstract

Background

The US National HIV/AIDS Strategy (NHAS) aims for 72% (90% diagnosed times 80% of those virally suppressed) viral suppression among persons with human immunodeficiency virus (HIV) by 2020. We examined the clinical and economic impact of reaching this target, in the general US population and among black men who have sex with men (MSM), the group with the highest HIV prevalence.

Methods

Using a mathematical simulation, we project the 5- and 20-year clinical outcomes, costs, and incremental cost-effectiveness ratios for (1) Current Pace of detection, linkage, retention, and virologic suppression and (2) NHAS investments in expanded testing ($24–$74 per test) and adherence ($400 per person-year), calibrated to achieve 72% suppression by 2020. We examined alternative rates of testing, retention, and suppression and the efficacy and cost of adherence interventions.

Results

Compared with Current Pace over 20 years, NHAS averted 280000 HIV transmissions (80000 in black MSM) and 199000 (45000) deaths and saved 2138000 (453000) years of life, while increasing costs by 23%. The incremental cost-effectiveness ratio for NHAS compared with Current Pace was $68900 per quality-adjusted life-year ($38300 for black MSM) and was most sensitive to antiretroviral therapy costs.

Conclusions

Reaching NHAS targets would yield substantial clinical benefits and be cost-effective in both the general US and black MSM populations.

(See the editorial commentary by Adimora, on pages 787–9.)

In July 2015, the White House released a revised National HIV/AIDS Strategy (NHAS) that outlines aspirational human immunodeficiency virus (HIV) treatment targets in the United States [1]. A central goal of the strategy is to align US policy with Joint United Nations Programme on HIV/AIDS (UNAIDS) initiatives and for 90% of persons with HIV to know their diagnosis and 80% of those with a diagnosis to be virologically suppressed by 2020. This would result in 72% overall virologic suppression among persons with HIV, a marked improvement from the current estimate of 49% [2, 3, 4]. Achieving the NHAS goal would require additional investments at every stage of the HIV care continuum, from testing and linkage to retention and increased antiretroviral therapy (ART) suppression.

NHAS also emphasizes reducing HIV-related disparities in communities at high risk for HIV infection, including black men who have sex with men (MSM), the group with the highest HIV incidence and prevalence. Black MSM face a 50% lifetime risk of HIV infection, contributing disproportionately to the incidence of HIV infection in the United States [5].

The proposed 2017 federal spending on domestic HIV/AIDS treatment and prevention is $21.7 billion, an increase of 3% from 2016, which would be dedicated to supplying lifesaving treatment and support services [6]. As payers are asked to scale up this support to achieve NHAS targets, our goal in this analysis is to provide quantitative estimates of the clinical and economic outcomes associated with successful implementation of the NHAS in the US and black MSM populations.

METHODS

Analytic Overview

We used the Cost-Effectiveness of Preventing AIDS Complications–US (CEPAC-US) microsimulation model to assess 2 strategies: (1) Current Pace of detection, linkage, retention, and viral suppression; and (2) NHAS aspirational rates of detection, linkage, and retention, resulting in 72% suppression (90% with HIV diagnosed and 80% of those virally suppressed) by 2020. Under both strategies, we projected 5- and 20-year clinical benefits and costs, as well as the incremental cost-effectiveness ratios (ICERs) and budgetary impact of reaching NHAS targets compared with Current Pace. We examined these outcomes for the general population of persons with HIV in the United States and for a high-incidence proper subset of this population (ie, black MSM with HIV) [5]. Costs and clinical outcomes were reported undiscounted; ICERs were discounted (numerator and denominator) at 3% per year and included health and economic benefits attributable to reduced transmissions. Quality-adjusted life-years (QALYs) were used in the calculation of ICERs. We considered strategies with ICERs <$100000/QALY as cost-effective [7].

Cohort Definitions

For both populations, we simulated prevalent HIV cases, as well as incident cases over 20 years. At model initiation, prevalent persons with HIV (general US and black MSM populations) were assigned to 1 of 4 states in the HIV care continuum: (1) undiagnosed, (2) receiving first-line ART, (3) receiving second-line or subsequent lines of ART, and (4) previously diagnosed but disengaged from care. The initial number of persons in each state was derived from Centers for Disease Control and Prevention (CDC) estimates [2]; however, after the start of the simulation, the process by which persons enter or leave one state or another was determined by the model itself, parameterized by rates of transmission, screening, linkage, treatment failure, and loss to follow-up. Every year, a cohort of newly infected persons was added to the population, using “incident” population attributes (Table 1). The size of this incident cohort was determined by model-based estimates of new transmissions from the prior year’s population of prevalent persons with HIV.

Table 1.

Selected Model Input Values for NHAS Analysis in the United States

Cohort	Age, Mean (SD), y	CD4 Cell Count, Mean (SD), Cells/µLa	Weighted Distribution, %	References
Incident				[8, 9]
US population	36 (13)	667 (134)	…
Black MSM	25 (4)
Prevalent, undiagnosed				[3, 8–10]
US population	43 (12)	436 (166)	14
Black MSM	32 (6)		23
Prevalent, receiving 1st-line ART				[3, 9, 11]
US population	46 (12)	600 (313)	15
Black MSM	35 (6)		13
Prevalent, receiving 2nd-line ART				[3, 9, 11]
US population	48 (12)	600 (313)	11
Black MSM	37 (6)		10
Prevalent, receiving 3rd-line ART				[3, 9, 11]
US population	48 (12)	600 (313)	11
Black MSM	37 (6)		9
Prevalent diagnosed, not in care				[3, 9, 10]
US population	48 (12)	325 (53)	49
Black MSM	37 (6)		45

Abbreviations: ART, antiretroviral therapy; MSM, men who have sex with men; NHAS, National HIV/AIDS Strategy; SD, standard deviation.

^aMean CD4 cell count at model initiation, which is assumed the same for both the US and black MSM populations. When reported in the literature as medians and interquartile ranges, CD4 cell counts were converted to means and SDs for reporting purposes.

CEPAC-US Model

Disease Model

CEPAC-US is a widely published and validated microsimulation model of HIV disease, detection, and treatment [12–14]. Simulated HIV-infected patients, in the absence of treatment, experience HIV RNA–dependent decline in CD4 cell counts and risks of opportunistic infections stratified by these counts. Individuals become eligible for care when their HIV infection is identified, through either routine HIV screening or diagnostic HIV testing at clinical presentation. On linkage to care, patients begin ART; if successfully suppressive, ART reduces HIV RNA, increases CD4 cell count, and decreases HIV-related mortality [15, 16]. ART success or failure is monitored with laboratory evaluation. Diagnosis of treatment failure prompts an opportunity to resuppress on the current regimen [16]. Repeated treatment failure leads to switching to a subsequent ART regimen. While receiving ART, patients experience a monthly risk of loss to follow-up and a subsequent probability that they return to care. ART suppression and rates of loss to follow-up and return to care are determined by an individual’s predetermined, general level of adherence [14, 17]. (Additional model detail is available at http://www.massgeneral.org/mpec/cepac/.)

HIV Transmission

The likelihood that a person will transmit HIV is modeled as a function of the HIV RNA level. For those receiving suppressive therapy, HIV RNA is undetectable; for all others, the HIV RNA level is adjusted for stage of infection (primary, chronic, or advanced).

Input Data

Cohort and Care Continuum Characteristics

Simulated patients were assigned cohort characteristics at the start of the model per their state in the HIV care continuum. US persons with HIV (general population) were, on average, older than black MSM with HIV (Table 1) [18]. Cohort characteristics were age and race adjusted for the black MSM population. Using CDC data, we defined the fractions at model outset as follows: undiagnosed HIV, 14% (black MSM, 23%); receiving first-line ART, 15% (13%); receiving second- or third-line ART, 22% (19%); and lost to follow-up or otherwise not in care, 49% (45%) (Table 1) [3, 8, 11, 19].

Testing and Treatment

In the Current Pace strategy, US persons with undiagnosed HIV received an HIV test, on average, every 9.8 years (black MSM, 12.8 years), a frequency calibrated such that persons presented to care with a mean CD4 cell count of 327/µL (301/µL) [10, 20]. Once linked, all persons received guideline-concordant care, with a first-line dolutegravir-based regimen (48-week efficacy, 87% [80% in black MSM]) [15, 21–24]. Adherence-stratified loss to follow-up rates ranged from 0.01% to 6.8% per month [25]. In the Current Pace, rates of loss to follow-up and return to care were calibrated in the model to published data such that 78% of persons (76% of black MSM) remained in care at 24 months [25, 26].

Transmission

HIV RNA–stratified transmission rates ranged from 0.16 to 9.03 transmissions per 100 PY, with higher rates for those within 3 months of infection (Table 2) [27, 28]. Transmission calculations in all strategies included a percentage reduction in HIV RNA–based transmission calculations informed by data on the frequency and population-based effectiveness of condom use (ie, reduction in transmission rate due to condom use, or condom transmission reduction, 40% for US persons and 10% for black MSM), assuming that the source data reflected negligible condom use [9, 29–31]. This reduction was always applied to prevalent cohorts and was applied to incident cohorts after 5 years to reflect a mean delay to diagnosis and self-awareness of status.

Table 2.

Selected Model Input Values for NHAS Analysis in the United States

Cohort Characteristics	Value	References
HIV RNA distribution after acute infection, %		[19]
>100000 Copies/mL	25
30001–100000 Copies/mL	42
10001–30000 Copies/mL	21
3001–10000 Copies/mL	6
≤3000 Copies/mL	6
ART efficacy, mean virologic suppression at 48 wk, %
1st-line ART (DTG), US population/black MSM	87/80	Derived from [15, 22, 23]
2nd- and 3rd-line ART (PI-based and salvage regimens)	80	Derived from [21, 24]
Resuppression (suppression 24 wk after diagnosed failure), %	72	Derived from [26]
Retention in care (retained at 24 mo), %		Derived from [25]
Current Pace strategy, US population/black MSM	78/76
NHAS strategy, US population/black MSM	89/86
Transmission rate by disease stage and viral load, transmissions/100 PY		[27, 28]
Incident infection (3 mo after infection)	65.47a
Late-stage disease (CD4 cell count <200/µL)	9.03
>100000 Copies/mL	9.03
10001–100000 Copies/mL	8.12
3001–10000 Copies/mL	4.17
501–3000 Copies/mL	2.06
21–500 Copies/mL	0.16
≤20 Copies/mL	0.16
Condom transmission reduction, %, US population/black MSM	40/10	[9, 29–31]

Abbreviations: ART, antiretroviral therapy; DTG, dolutegravir; HIV, human immunodeficiency virus; MSM, men who have sex with men; NHAS, National HIV/AIDS Strategy; PI, protease inhibitor; PY, person-years.

^aDuring the period of acute infection, the transmission rate is 7.25 times that of the highest viral load stratum (>100000 copies/mL) [27].

Costs

Included were ART costs ranging from $2040 to $3300 per month, HIV RNA costs of $116 per test, and CD4 cell count–stratified routine care costs ranging from $236 to $2572 per month (Table 3) [32, 35, 36]. Costs were updated to 2014 US dollars using the Bureau of Labor Statistics consumer price index for medical care [33, 34].

Table 3.

Selected Model Input Values for NHAS Analysis in the United States

Costs	Value, $a	Reference
ART costs, monthly (annual)		[32, 33]
1st-line ART	2040 (24480)
2nd-line ART	2500 (30012)
3rd-line ART	3300 (39648)
Monitoring costs per test		[34]
HIV RNA	116
CD4 cell count	64
Routine care cost by CD4 cell count, monthly (off/on ART)		[35, 36]
>500/µL	290/236
351–500/µL	353/485
201–350/µL	553/495
101–200/µL	1063/573
51–100/µL	1347/1141
≤50/µL	2572/1045
NHAS strategy only		[37, 38]
HIV test cost per negative test	24
HIV test cost per positive test	74
Annual adherence/retention intervention cost, per patient receiving ARTb	422	[39]

Abbreviations: ART, antiretroviral therapy; HIV, human immunodeficiency virus; NHAS, National HIV/AIDS Strategy.

^aCosts in 2014 US dollars

^bAdherence/retention intervention affects ART efficacy (% suppressed) and attenuates loss to follow-up, as described in Methods.

Input Parameters to Achieve the NHAS Goal

To reach NHAS targets, we calibrated the detection, linkage, and retention rates such that 72% overall suppression was achieved by 2020. We increased testing rates to a mean of once every 3.6 years. We conservatively estimated the number of HIV-uninfected persons needing to be tested to identify each person with HIV by dividing the number of persons with HIV linked to care for the first time each year by the undiagnosed HIV prevalence of the entire US (or black MSM) population; we assumed linkage rates after positive test result of 80% (50% for black MSM) [40]. Each negative HIV test cost $24 and each positive test $74, which included the cost of confirmation and post-test counseling [37, 38]. We assumed a high-impact intervention resulting in treatment adherence and 24-month retention rates of 89% for US persons (86% for black MSM); because of the intensity this intervention may require, we assumed that the intervention would cost a yearly average of $422 for every person receiving ART [39].

Sensitivity Analyses

In 1-way sensitivity analyses, we varied each parameter across its plausible range while holding all other parameters at their base values. The most influential parameters from 1-way sensitivity analyses were subjected to additional, multiway sensitivity analysis. We also considered an extreme scenario 1 (better NHAS outcomes), where we varied multiple parameters that worsened Current Pace outcomes compared with NHAS, as well as an extreme scenario 2 (worse NHAS outcomes), where we varied parameters that improved Current Pace relative to NHAS outcomes (Supplementary Table S1).

RESULTS

HIV Care Continuum

Projected results of the Current Pace strategy, after 1 year of model simulation, were validated with recent CDC continuum of care estimates (Figure 1A and 1B) [3]. Reflective of modest improvements over time, for the US (black MSM) population, Current Pace resulted in 52% (45%) overall virologic suppression at 5 years and 64% (55%) suppression at 20 years (in 2035) (Supplementary Figure S1). NHAS virologic suppression was projected to be 72% (68%) at 5 years (2020) and 80% (73%) at 20 years (Figure 1B, dashed and solid vertical lines).

Figure 1.

Human immunodeficiency virus (HIV) care continuum and transmission rates by strategy. A, B, Percentages of all persons with HIV who are unlinked to care (A) or suppressed (B) under the Current Pace levels of diagnosis, linkage, and retention are shown; dashed and solid vertical lines in B denote 5- and 20-year projections, respectively. Data for the general US population are shown in blue, and for black men who have sex with men (MSM) in red. Dramatic improvements in the HIV care continuum to achieve the National HIV/AIDS Strategy (NHAS) viral suppression target of 72% by 2020 result in the dotted NHAS care continuum lines. At year 1, the Current Pace model–projected care continuum for the general US population is validated with Centers for Disease Control and Prevention continuum of care estimates (stars). C, Transmissions (per 100 person-years [PY]) decrease over time in both the Current Pace and the NHAS strategies; black MSM have higher transmission rates than the general US population in both strategies. The initial increase in transmissions in the Current Pace strategies is due to a delay in viral suppression on antiretroviral therapy.

HIV Transmissions

Projected HIV transmissions under the Current Pace totaled 42000 (10000 for black MSM) at the end of year 1, consistent with CDC estimates [5]. Over the subsequent 20 years, the Current Pace resulted in 670000 new cases of HIV in the United States, of which 195000 (29%) were in black MSM (Table 4). Achieving NHAS targets decreased transmissions from 2.8/100PY to 1.7/100PY in the US population (4.3/100 PY to 2.7/100 PY among black MSM) and averted 280000 (80000) transmissions (Figure 1C and Table 4). Under NHAS over 20 years, black MSM contributed 19% of the total life-years lived but accounted for 28% of the transmissions averted (Table 4 and Supplementary Figure S2). Despite increased survival among persons with HIV, the number of US persons with HIV at 20 years decreased by 82000 in the NHAS strategy, owing to fewer transmissions.

Table 4.

Clinical and Economic Outcomes of Current Pace and NHAS Strategies at 20 Years^a

Outcomesb	US Population (n = 1200000)c	Black MSM (n = 180000)c
Total transmissions
Current Pace	670000	195000
NHAS	390000	115000
Transmissions averted	280000	80000
Cumulative deaths
Current pace	750000	110000
NHAS	551000	65000
Deaths averted	199000	45000
Years of life
Current Pace	23932000	4531000
NHAS	26070000	4984000
Years of life saved	2138000	453000
Total costs, $ million
Current Pace	525274	88658
NHAS	645650	102493
Additional costs	120376	13835
ART costs, $ million
Current Pace	390289	64524
NHAS	478951	80478
Additional costs	88662	15954
Care and monitoring costs, $ million
Current Pace	134984	24135
NHAS	89545	15604
Difference in costs	–45439	–8531
NHAS-only costs, $ million
Testing intervention	42705	117
Adherence intervention	7433	1257
ICER, NHAS, $/QALY d	$68900/QALY	$38300/QALY

Abbreviations: ART, antiretroviral therapy; ICER, incremental cost-effectiveness ratio; NHAS, National HIV/AIDS Strategy; QALY, quality-adjusted life-years.

^aFor each set of results, the third row highlights the difference between NHAS and Current Pace strategies

^bAt 1 year, Current Pace transmissions are 44000 in the general US population and 11000 in the black MSM population.

^cThese values are at model initiation only. Population sizes change throughout the analysis as deaths and transmissions occur.

^dClinical outcomes and total costs are reported undiscounted; ICERs are calculated as Δ$/ΔQALY for NHAS strategy compared with the Current Pace strategy. Discounted values at 3% per year are used for ICERs.

Deaths Averted and Life-Years Saved

Over 20 years, the Current Pace strategy resulted in 750000 deaths from all causes among US persons with HIV (110000 in black MSM). NHAS resulted in 551000 (65000) deaths, averting 199000 (45000) deaths over 20 years. The NHAS strategy saved a projected 2138000 (453000) years of life over 20 years (Table 4).

Costs and Cost-effectiveness

As a point of validation, model-based Current Pace 1-year costs were consistent with estimated annual per-person costs from the HIV Research Network [41]. Current Pace cumulative total costs at 20 years were $525.3 billion for US persons ($88.7 billion for black MSM), 74% (73%) of which were ART costs and 2% (2%) of which were monitoring costs. Over 20 years, the NHAS strategy increased cumulative costs to $645.7 ($102.5) billion, with $42.7 ($0.1) billion and $7.4 ($1.3) billion allocated for the NHAS-specific testing and adherence interventions (Table 4 and Supplementary Figure S3). The ICER for the NHAS compared with the Current Pace strategy at 20 years was $68900/QALY in the general US population and $38300/QALY in the black MSM population.

For each outcome examined, resources were more efficiently used when targeted to the black MSM population. For example, total NHAS scale-up for the United States cost $605000 per death averted and $429100 per transmission averted, compared with $309400 and $174400, respectively, for the black MSM population.

Sensitivity Analyses

Clinical outcomes, measured in transmissions averted, deaths averted, and years of life saved under NHAS compared with Current Pace, were most sensitive to variations in condom transmission reduction, NHAS scale-up pace, 24-month retention in the Current Pace strategy, and the adherence intervention efficacy (Figure 2A and Supplementary Figure S4). We highlight key sensitivity analyses here but present a summary of the results in Figure 2 and in Supplementary Table S2. When we increased the condom transmission reduction (thereby reducing total transmissions and improving both strategies), NHAS still averted transmissions compared with Current Pace, only fewer than in the base case: with a 60% reduction (15% for black MSM), NHAS averted 38% (10%) fewer transmissions. This finding is because the increased effectiveness of the condom transmission reduction partially masks the benefits of NHAS. Conversely, with a 20% reduction (5% for black MSM), compared with base-case reduction of 40% (10%), NHAS averted 43% (11%) more transmissions.

Figure 2.

One-way and multiway sensitivity analyses. The percentage change compared with base-case differences between National HIV/AIDS Strategy (NHAS) target outcomes and Current Pace outcomes are shown for a variety of sensitivity analyses. More beneficial NHAS outcomes compared with the NHAS base case are depicted by increasingly intense shades of green (more transmissions or deaths averted, more years of life saved, and lower costs), and worse NHAS outcomes by increasingly intense shades of red (fewer transmissions or deaths averted, fewer years of life saved, and higher costs). Although increased prevention due to condom use would be an improvement, it is included in the NHAS worst case because it improves the Current Pace strategy and partially masks the benefit of the NHAS. See Methods and the Supplementary Materials for details on the parameters varied in each sensitivity analysis. ART, antiretroviral therapy; BMSM, black men who have sex with men; ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life-year; US, general US population.

If the NHAS goals take 10 years to achieve, rather than 5, we projected 22% (20%) fewer transmissions averted, 20% (18%) fewer deaths averted, and 22% (19%) fewer years of life saved in the United States. When 24-month retention rates in Current Pace were increased to 83% (81%)—from 78% (76%) in the base case—the clinical benefits were reduced by 19%–31%. When the efficacy of the NHAS adherence intervention was 0.25 times that of the base case, clinical benefits of the NHAS strategy decreased by 34%–43% across the US and black MSM populations.

Cost outcomes were most sensitive to variations in ART costs (Figure 2B and Supplementary Figure S4). When ART costs were increased by 1.5 times, the NHAS additional cost increased by 37% in the US population (58% for black MSM). Halving ART costs—as might occur with increased generic use—led to a 37% (58%) reduction in the NHAS additional cost. In this scenario, overall costs for the NHAS were lower than the base-case cost of Current Pace (with base-case ART costs).

All sensitivity analyses in both populations resulted in an ICER of <$100000/QALY for the NHAS scale-up compared with Current Pace, except when the adherence intervention efficacy was 0.25 times and the cost 4 times the base case in the US population. ICERs remained <$100000/QALY for all parameter variations for the black MSM population. When the condom transmission adjustment for the black MSM population was increased to 40% (equal to the base-case adjustment in the US population), the ICER was $50500/QALY.

Extreme scenario 1 (better NHAS outcomes, with condom use reducing transmissions by 20% [5% in black MSM], 50% testing costs, and 70% ART costs) resulted in 43% (11%) more transmissions averted, 14% (4%) more deaths averted, 12% (3%) more years of life saved, and an ICER of $35200/QALY ($24500/QALY). Extreme scenario 2 (worse NHAS outcomes, with condom use reducing transmissions by 60% [15%], in both strategies, ART efficacy improving to 95% in Current Pace, and a slower pace of NHAS scale-up, all simultaneous) resulted in 52% (30%) fewer transmissions averted, 33% (24%) fewer deaths averted, 32% (26%) fewer years of life saved, and an ICER of $90900/QALY ($46400/QALY).

DISCUSSION

Achieving NHAS targets by 2020 would yield substantial clinical benefits in the United States, averting >250000 new infections and >200000 deaths, and saving >2 million years of life over the next 20 years. While requiring a 23% increase in HIV care costs (an additional $120 billion over the next 20 years), NHAS was cost-effective for the general US population ($68900/QALY). When NHAS was evaluated specifically for black MSM, similarly dramatic clinical benefits were realized, with an ICER of $38300/QALY. Further, though implementing NHAS was cost-effective and recommended for the general US population, we find that resources would be even more efficiently used in the subset of the black MSM population, specifically owing to comparatively higher incidence, higher undiagnosed prevalence, and lower retention rates among black MSM at Current Pace [3, 40]. The results were robust across variation in parameters and populations.

An overall viral suppression rate of 72% is an increasingly plausible target; even with the modest scale-up under the Current Pace strategy, viral suppression would reach 64% at 20 years. Major improvements in the management of HIV, including more efficacious, tolerable, and durable ART, have already reduced the incidence of opportunistic infections; persons with HIV who are adherent to therapy and do not have advanced immunosuppression are expected to live nearly as long as similar, HIV-uninfected persons [42]. With the added impact of treatment as an effective prevention strategy, ART expansion with reliable suppression would accelerate the decreasing HIV incidence in the United States, a phenomenon already observed in many communities [9]. These results demonstrate the economic impact of NHAS and suggest that investments in HIV detection, linkage, adherence, and retention prove an excellent value in the United States.

Ongoing, well-documented health disparities persist among black MSM at each stage of the care continuum [40, 43]. This population has a considerably higher observed incidence rate and faces disproportionately high socioeconomic barriers, structural barriers, and stigma that limit access to care and exacerbate HIV-related health outcomes [43]. Some states with high populations of black MSM have healthcare policies, such as stricter Medicaid eligibility criteria and nonmandatory sex education, that only exacerbate their HIV-related health disparities [44]. Considering these barriers, achieving NHAS treatment targets for all persons with HIV in the United States, including black MSM, will be difficult and will require novel, culturally grounded, and nonstigmatizing interventions that recognize and address the specific determinants of health in key populations [45].

A recent study that engaged black MSM found that whereas investigators were worried about adherence to visits and treatment, many participants were more concerned with more basic needs, such as food and housing security [46]. Future interventions to engage this population in HIV care should better align with these goals. Tailored adherence interventions, such as a mobile text messaging app coupled with a mentor from the community, as well as other creative uses of technology, may prove especially useful in targeting this population [47]. Despite the aforementioned challenges and critical need for novel and targeted adherence interventions in the black MSM community, our study shows that those intervention designs need not be perfect to have a substantial impact: modestly effective adherence/retention interventions targeted to the black MSM population are likely to be cost-effective.

Broad implementation of the NHAS strategy is affordable. Although the overall 20-year costs for NHAS expansion may seem high, the federal HIV budget would only need to grow at approximately 5% per year for the next 20 years to support all model-projected costs of the NHAS strategy in the United States [6]. If persons with HIV who are privately insured had similar costs as those receiving federal funding, federal allocations would need to increase by only 3% per year. Furthermore, the model-projected NHAS target budget is most substantially influenced by ART costs, with NHAS testing program and adherence intervention costs comprising only 8% of the NHAS budget. Any efforts to reduce ART costs (eg, generic drugs) would both improve cost-effectiveness and substantially decrease the funding needed to achieve NHAS targets. A decrease in ART costs of 30% would more than make up for the increased spending required to achieve NHAS targets. Although healthcare spending in the United States is under pressure, we conclude that the NHAS strategy is affordable, with a likely 3% yearly budget increase that is in line with last year’s 3% proposed increase.

Because policy makers are increasingly encouraged to consider equity criteria in addition to cost-effectiveness analysis when setting priorities, these results present an example in which equity-based and cost-effectiveness conclusions appear aligned [48]. The NHAS strategy in the black MSM community is not only more cost-effective than in the US population, but it also helps reduce disparities in a population disproportionately affected by the HIV epidemic. Allocating resources specifically for black MSM in the NHAS makes clinical, public health, ethical, and economic sense; policy makers should recognize this population as an initial focus, while wide-scale implementation of a strategy to attain NHAS targets is realized in the general US population.

We note several limitations. First, we assumed an adherence intervention in the current analysis that, though expensive, yielded dramatic improvement in the US and black MSM populations. Although short-term coordinated care programs have shown promise in enhancing engagement in care and viral suppression, we assumed a longer-term and larger-scale intervention than any validated intervention of which we are aware [49]. We tested this assumption in sensitivity analyses by increasing the cost and decreasing the effectiveness of the NHAS intervention. In addition, the condom adjustment is an oversimplification intended to reflect the effectiveness of a diverse menu of prevention efforts in these populations. As such, we assumed a lower transmission benefit due to condoms in the black MSM population compared with the general US population; this was consistent with CDC transmission data [5, 43]. Furthermore, the cost-effectiveness results remained robust to variations in this parameter, including when the adjustment was assumed to be equal between the US and black MSM populations.

We did not specifically examine other key populations critical to attaining NHAS targets in the United States, such as black and transgender women or white/Hispanic MSM, because of their smaller absolute numbers [50, 51]. To the extent that their HIV incidence is higher than that of the general US population, as well as other documented disparities in the care continuum, they probably also merit additional resources and attention. Finally, we performed a subpopulation analysis for black MSM and did not assess implementing NHAS in the black MSM compared with the general US population. However, given that the policy recommendation without such an incremental comparison herein is to implement NHAS in the entire US population, this comparison would not change our conclusions, because achieving NHAS targets in some but not other subgroups would probably be viewed as politically unacceptable and unethical.

Successful achievement of the NHAS goals in the United States, while increasing overall 20-year costs by 23%, would move the United States toward the global target of eliminating new HIV infections worldwide and would save millions of years of life, all while offering good value. A modest reduction in ART costs would more than offset the investments necessary to reach these goals.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Financial Support. This research was funded by the National Institutes of Health (grants T32AI007433, R01 AI042006, R37 AI093269, K01 HL123349, R01 DA015612, and R01 MH105203) and by the Steve and Deborah Gorlin MGH Research Scholars Award (Massachusetts General Hospital Executive Committee on Research to R. P. W.).

Potential conflict of interest. P. E. S. reports grant support for research to his institution from BMS, Gilead, and GlaxoSmithKline (GSK)/ViiV and reports consulting work or participation as a scientific advisory board member for AbbVie, Bristol-Myers Squibb, Gilead, GSK/ViiV, Merck, and Janssen. M. C. W. has served as a consultant for OptumInsight on topics unrelated to HIV. All other authors report no potential conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.