Daily Oral HIV Pre-exposure Prophylaxis Among Young Men Who Have Sex With Men in the United States: Cost-saving at Generic Drug Price.
Feb 17, 2024DOI: 10.1093/cid/ciad566Journal ArticleCID
Written by: Alyssa K AmickG Ege EskibozkurtSybil G HosekClare F FlanaganRaphael J LandovitzElena Y JinCraig M WilsonKenneth A FreedbergMilton C WeinsteinPooyan KazemianA David PaltielAndrea L CiaranelloAnne M Neilan
Find these authors on Roon
RL
Dr. Raphael J. Landovitz
HIV/AIDS Medicine · UCLA
AC
Dr. Andrea Ciaranello
HIV/AIDS Medicine · MGH
AN
Dr. Anne Neilan
HIV/AIDS Medicine · MGH
METHODS
We used the Cost-Effectiveness of Preventing AIDS Complications (CEPAC)–Adolescent microsimulation model to simulate YMSM at increased risk of HIV acquisition (as defined by ATN 110/113 eligibility) in the United States. Criteria for increased risk of infection in these trials include recent history of condomless anal sexual intercourse, sexually transmitted infection, or multiple sexual partners. We modeled 2 strategies: (1) “Screening” (annual HIV screening, as recommended by national guidelines [14]) and (2) “PrEP” (TDF-FTC–based daily oral PrEP with HIV screening every 3 months). We projected HIV transmissions (percent of cohort infected), quality-adjusted life-years (QALYs), and costs. Clinical and cost outcomes per person are reported undiscounted and discounted at 3%/year [15, 16]. Costs are scaled to 1000 YMSM to highlight differences between strategies. We report incremental cost-effectiveness ratios (ICERs; $/QALY), calculated by dividing the difference in strategy cost by the difference in QALYs. We examined a range of willingness-to-pay thresholds ($100 000/QALY–$200 000/QALY) [17]. Model outcomes are projected over 10-year and lifetime horizons.Young MSM enter the model without HIV. Their progress through the model is tracked in monthly cycles over both a fixed 10-year horizon and a lifetime horizon, where the simulation continues until the individual's death. Each month, individuals face age-stratified risks of HIV incidence. HIV diagnosis can occur via (1) status quo HIV testing (ie, “background detection”), (2) a regularly scheduled screening program (eg, annual or every 3 months), or (3) testing after developing an opportunistic infection (an infection that often occurs in people with weakened immune systems) [18].In the PrEP strategy, all individuals enroll in the PrEP program at the beginning of the simulation and face monthly probabilities of dropout through the program end at 10 years. They also face reduced monthly probabilities of acquiring HIV when on PrEP, depending on PrEP adherence and efficacy. Simulated individuals are screened for HIV at PrEP program visits.The CEPAC model simulates disease progression for individuals with HIV regardless of whether they are diagnosed. Diagnosed individuals can link to HIV care and start antiretroviral therapy (ART). The probability of linking to care is user-specified and differs by detection mode (ie, detected by background screening, a screening program, or after presenting with an HIV-related illness). Simulated individuals who are not diagnosed or do not link to care do not receive ART; they experience a monthly decline in CD4 count resulting in a higher probability of opportunistic infections and HIV-associated mortality. Those who become virologically suppressed on ART face a monthly probability of failing therapy and developing viremia, which can be detected through routine viral load testing and may result in a switch in ART regimen. Individuals in care face a monthly probability of being lost to follow-up and stopping ART, and then of returning to care thereafter.
HIV transmission within the cohort is also simulated in the model [19]. Primary transmission rates vary by viral load, stage of HIV infection (acute and chronic), and response to ART. Pre-exposure prophylaxis confers both direct and indirect transmission benefits. Individuals receiving PrEP receive the direct benefit of reduced personal susceptibility to infection. All uninfected persons—even those not receiving PrEP—receive the indirect transmission benefits: reduction in the overall prevalence of infection in the population and quicker detection/linkage of breakthrough cases, leading to quicker virologic suppression and reduced community viral load (Supplementary Methods). Results are shown both with and without accounting for indirect benefits of averting transmissions.
All simulated individuals are assigned male sex at birth and enter the model between ages 15 and 24 years (mean: 19.2, SD: 2.1) (Table 1) [5, 6]. To reflect the skew of these age bounds, we used 100 000 simulations of a normal distribution and fit these data to approximate the initial age model parameters; each simulated individual is assigned an age from this distribution at model start.
Input Parameters for a Model of PrEP for Young Men Who Have Sex With Men at Increased Risk of HIV Acquisition in the United States
Additional details of inputs may be found in the Supplementary Methods and Supplementary Table 1.
Abbreviations: ART, antiretroviral therapy; HIV, human immunodeficiency virus; PrEP, pre-exposure prophylaxis; PY, person-years; TDF-FTC, tenofovir disoproxil fumarate/emtricitabine; USD, US dollar.
aCorresponds to 28% of the initial modeled cohort retained in the PrEP program at 6 years (range in sensitivity analysis of 6–100% retention at 6 years).
bCosts are entered into the model as monthly costs.
Monthly on-PrEP HIV incidence was estimated from ATN 110/113 data (Supplementary Tables 2 and 3). Monthly off-PrEP HIV incidence was estimated from on-PrEP HIV incidence using published PrEP efficacy data (4 doses/wk is 96% effective; 7 doses/wk is 99% effective) accounting for PrEP adherence as reported in ATN 110/113 (33–56%) (Supplementary Table 1, Supplementary Figure 1) [2, 5, 6]. PrEP program retention is 28% at 6 years [22]. In the screening and PrEP interventions, both of which continue for 10 years, individuals not known to have HIV are screened for HIV annually and every 3 months, respectively [13, 14]. Based on 2021 national guidelines, individuals receive both an HIV Ag/Ab and HIV RNA test [13]. These screening tests are in addition to “background testing,” which reflects routine testing in the medical setting (approximately once every 3.4 years) [11].
Linkage to care probability varies by mode of diagnosis (after background testing: 77%; through the Screening or PrEP strategy: 100%) [24]. Individuals linked to care are prescribed a once-daily integrase strand transfer inhibitor–based ART regimen [32, 33]. ART efficacy is stratified by adherence level, with ongoing probabilities of developing viremia and then resuppression, as well as regimen changes [32, 33]. Efficacy is based on aggregated suppression data at 48 weeks [32, 33]. Similar efficacy was observed between dolutegravir- and bictegravir-based regimens in more recent studies [34, 35]. While in HIV care, individuals face an adherence-stratified probability of becoming lost to follow-up, ranging from 0.01% to 6.8% monthly, with higher rates among those with lower adherence [25].
Transmission rates range from 0.0 to 78.4 transmissions/100 person-years (PY), stratified by HIV RNA level, with the highest rates for those with acute HIV infection and not yet on ART (Supplementary Table 1) [36].
Annual PrEP program costs are $900, including the following: $104/visit (4 times/y), $5/serum creatinine (2 times/y), $2/urinalysis (2 times/y), and $118/HIV test (4 times/y) [27, 37]. HIV testing includes both an HIV antigen/antibody ($33/test) and an HIV viral load test ($85/test) [37]. For reactive screening tests, counseling costs of $24 are also incurred. The generic PrEP drug price is from Red Book at $360/year based on wholesale acquisition costs (Supplementary Table 4) [12]. ART costs range from $32 000 to $69 000 annually by ART line; we averaged costs from multiple regimens recommended in US guidelines [12, 27, 28].
In 1-way sensitivity analyses, we varied single parameters including PrEP costs, ART costs, HIV care costs, linkage to care rates, and PrEP retention, efficacy, and adherence (Table 1). In 2-way sensitivity analyses, we varied influential parameters together—HIV incidence and PrEP drug price, and HIV incidence and PrEP program retention. Additional 2-way analyses included the following: HIV incidence and ART price, PrEP price and PrEP retention, and PrEP price and ART price. Because of the more inclusive 2021 PrEP guidelines, we varied off-PrEP HIV incidence from a high of 15.2/100 PY to a low of 0.1/100 PY (base-case value: 10.1/100 PY). In additional scenario analyses, we assessed the branded oral PrEP drug price of $9000 annually [27], the incremental benefit of PrEP with 3-monthly screening compared with 3-monthly screening alone, and the “worst case” of adverse events associated with TDF-FTC [38].
RESULTS
Over 10 years, 37% of the cohort in the Screening strategy would acquire HIV compared to 30% in the PrEP strategy (Table 2). The PrEP strategy would increase undiscounted quality-adjusted life expectancy compared with Screening by 0.09 and 0.42 QALYs over 10-year and lifetime horizons, respectively (Table 2). Individuals would be on PrEP for an average of 2.8 years. For individuals who acquire HIV, the PrEP strategy would delay acquisition by 3.5 months compared with Screening (Supplementary Table 5).
Model-Projected Outcomes for HIV Pre-exposure Prophylaxis and Annual HIV Screening Strategies
Strategies are listed in order of increasing cost per cost-effectiveness analysis convention. Cost-saving strategies are less expensive but more effective than the comparator strategy. Results are shown for the generic PrEP drug price. Some individuals developed HIV but were not detected: 10-year horizon: Screening—3.17%, PrEP—2.68%; lifetime horizon: Screening—1.78%, PrEP—1.75%.
Abbreviations: HIV, human immunodeficiency virus; ICER, incremental cost-effectiveness ratio; PrEP, pre-exposure prophylaxis; QALY, quality-adjusted life-year; $, 2020 US dollars.
aIncremental cost-effectiveness ratios are the difference in total healthcare-related costs divided by the difference in total QALYs compared with the next most expensive strategy. Costs and QALYs are discounted at 3%/year.
In year 1, compared with the Screening strategy, PrEP would have higher costs primarily due to PrEP drug price, in addition to increased HIV screening and PrEP program costs (Figure 1). Over time, ART costs in the Screening strategy would increase more steeply than in PrEP, as more individuals acquire HIV. After year 3, the cumulative costs of PrEP would become less than Screening because fewer people would incur the costs of HIV, including ART. Over 10 years, the PrEP strategy would decrease total costs by $5000/person, undiscounted, compared with Screening; at longer horizons, the difference in costs would increase, with PrEP decreasing per-person costs by $35 000, undiscounted, at the lifetime horizon (Table 2).
Ten-year annual component costs of Screening and PrEP strategies per 1000 YMSM at increased risk of HIV infection. The figure depicts the projected annual component costs of strategies (left bar: Screening; right bar: PrEP). The vertical axis shows the annual total cost in 2020 million USD. The horizontal axis shows select years from the model start. Annual component costs include ART drug, HIV-related healthcare, HIV test, PrEP program, and PrEP drug. Costs are scaled to 1000 YMSM at increased risk of HIV infection in the United States. Abbreviations: ART, antiretroviral therapy; HIV, human immunodeficiency virus; PrEP, pre-exposure prophylaxis; USD, US dollars (2020); YMSM, young men who have sex with men.
At the generic drug price, the PrEP strategy would be cost-saving at 10-year and lifetime horizons compared with Screening (Table 2). Additional clinical, cost, and cost-effectiveness outcomes stratified by HIV status are shown in Supplementary Table 6.
At the generic drug price, the PrEP strategy would remain cost-saving when varying the following parameters over the ranges examined: background HIV detection rate, linkage to care, HIV care cost, and PrEP adherence, efficacy, and program retention. Likewise, when PrEP uptake was varied, per-person life expectancy and cost changed minimally (Supplementary Table 7). At base-case incidence, the PrEP strategy would remain cost-saving over 10 years if first-line ART were $16 000 or more per year. The PrEP strategy would not exceed a $100 000/QALY willingness-to-pay threshold at any ART cost (Supplementary Figure 2A). Over a lifetime, PrEP would be cost-saving at base-case incidence even if ART were free due to high non–ART-related HIV care costs (Supplementary Figure 2B). At the base-case ART price, the PrEP strategy would be cost-saving at incidences of 2.5/100 PY or greater at the 10-year horizon and 0.5/100 PY at a lifetime horizon (Supplementary Figure 2). Removing the additional cost reductions from averted transmissions does not substantially change the conclusions (Supplementary Table 8).
As PrEP drug price decreases, the PrEP strategy would be more likely to be cost-saving compared with Screening at higher HIV incidences: over 10 years, PrEP would be cost-saving for off-PrEP incidences of 5.1/100 PY or greater at or below generic PrEP drug prices (≤$360/y) (Figure 2A). At a lifetime horizon, the PrEP strategy would be cost-saving for off-PrEP incidences of 1.0/100 PY or greater at or below generic PrEP drug prices (≤$360/y) (Figure 2B). At an incidence of 0.1/100 PY, however, ICERs for the PrEP strategy would exceed $100 000/QALY, even at a PrEP drug price of $0/year (lifetime ICER = $610 000/QALY) (Figure 2B).
Two-way sensitivity analysis varying off-PrEP HIV incidence and PrEP drug price. Two-way sensitivity analysis simultaneously varying PrEP drug price (vertical axis) from $0 to $9000 and HIV off-PrEP incidence (horizontal axis), from 0.1 to 10.1/100 PY at 10-year (A) and lifetime (B) horizons. Each cell reflects the ICER of the PrEP compared with Screening strategies, ranging from cost-saving to $0–$100 000/QALY, $100 001–$200 000/QALY, and >$200 000/QALY. HIV incidence (off-PrEP) is stratified by age. Values along the x-axis reflect the highest incidence rate for the 15–17-year-old age group. Base-case values for both branded and generic PrEP are labeled as “Branded” or “Generic” in bold. Abbreviations: HIV, human immunodeficiency virus; ICER, incremental cost-effectiveness ratio; PrEP, pre-exposure prophylaxis; PY, person-years; QALY, quality-adjusted life-year.
Varying retention levels with HIV incidence had minimal effect on the ICER of the PrEP strategy compared with Screening at the 10-year horizon (Figure 3A) and no effect at the lifetime horizon (Figure 3B). A similar pattern was seen at varying retention levels with annual PrEP drug price (Supplementary Figure 3). Two-way sensitivity analyses demonstrated that the PrEP strategy could be cost-saving or cost-effective over a range of PrEP and ART costs (Supplementary Figure 4).
Two-way sensitivity analysis varying monthly probability of PrEP program retention and off-PrEP HIV incidence. Two-way sensitivity analysis simultaneously varying 6-year retention in PrEP program (vertical axis) from 6% to 100% and HIV off-PrEP incidence (horizontal axis) from 0.1 to 10.1/100 PY at 10-year (A) and lifetime (B) horizons. Each cell reflects the ICER of the PrEP compared with Screening strategies, ranging from cost-saving to $0–100 000/QALY, $100 001–200 000/QALY, and >$200 000/QALY. HIV incidence (off-PrEP) is stratified by age. Values along the x-axis reflect the highest incidence rate for the 15–17-year-old age group. The base case is labeled with an X. Results are shown for PrEP at generic prices. Abbreviations: HIV, human immunodeficiency virus; ICER, incremental cost-effectiveness ratio; PrEP, pre-exposure prophylaxis; PY, person-years; QALY, quality-adjusted life-year.
At branded PrEP drug price, over 10 years, the PrEP strategy would increase costs by $26 000, discounted, compared with Screening, resulting in an ICER of $425 000/QALY. At the lifetime horizon, however, the PrEP strategy would lead to an ICER of $62 000/QALY (Supplementary Table 9). At branded PrEP drug price, cost and cost-effectiveness results are most sensitive to changes in HIV incidence, PrEP retention, and PrEP efficacy (Supplementary Figure 5).
We also compared the PrEP strategy to an every-3-month testing strategy to assess the incremental benefit of the PrEP strategy. Compared with every-3-month testing alone, at generic PrEP price, the PrEP strategy would be cost-saving over all time horizons. At branded PrEP price, the PrEP strategy would exceed commonly accepted willingness to pay thresholds over 10 years (ICER = $384 000/QALY) but not over a lifetime (ICER = $49 000/QALY) (Supplementary Table 10).
Finally, we examined the cost and clinical implications of the “worst case” of TDF-FTC–related adverse events for the PrEP strategy, over 10 years (ICER = $19 000/QALY) and over a lifetime (cost-saving) (Supplementary Table 11).
DISCUSSION
In this adolescent-focused simulation modeling analysis, we projected the clinical, cost, and cost-effectiveness outcomes of daily oral PrEP compared with every-3-month screening among a population of YMSM at increased risk of HIV acquisition. We had 3 major findings. First, at the generic PrEP drug price, the PrEP strategy was projected to provide more clinical benefit, including longer quality-adjusted life expectancy and fewer HIV infections, at a lower cost compared with Screening. Second, in sensitivity analyses, we found that these findings were not dependent on retention. In fact, retention in the PrEP program could decrease to 6% at 6 years (base-case value: 28%) without altering our main conclusions. This suggests that even a short amount of time on PrEP during this increased-risk period for YMSM can positively impact clinical outcomes and life expectancy in a cost-effective manner. Finally, at the generic price point, oral PrEP would be cost-saving compared with every-3-month HIV screening. This underscores the added benefit of the oral PrEP regimen in addition to increased screening, through the prevention of HIV infection and its resulting care costs. Taken together, these findings indicate strong support for the cost-effectiveness of PrEP in YMSM at increased risk of HIV infection in the United States.
Despite these large potential benefits, expanding PrEP prescription beyond specialty clinics to primary care settings as well as nontraditional delivery options will be critical to improve accessibility and equity [39–41]. Interventions addressing inequities, particularly by geography and race, are needed: the US Southeast lags in percentage of YMSM indicated for PrEP who are on PrEP, and across the United States, PrEP usage among Black and Hispanic MSM is lower compared with White MSM [8, 42]. Because PrEP would be cost-saving, among the highest risk population, such as those in ATN 110/113, additional resources could be directed towards improving PrEP uptake, retention, and adherence, particularly in populations that lack access, while remaining cost-effective compared with every-3-month screening.
There are a few notable exceptions. First, if HIV incidence decreased to 0.1/100 PY, an incidence far lower than what we would expect within this population, PrEP would not be cost-effective, even at the generic drug price. New Centers for Disease Control and Prevention PrEP guidelines expand PrEP eligibility criteria; our findings support even broader implementation, including in populations with incidences as low as 0.5/100 PY. Our conclusions, however, set a lower bound on the implementation of PrEP based on the current guidelines for PrEP since PrEP would not be cost-effective in a population with an incidence of 0.1/100 PY or lower due to PrEP program costs. Second, at the branded drug price, PrEP's cost-effectiveness varied with changes in HIV incidence, PrEP adherence, and PrEP efficacy, among others. The PrEP drug price was highly impactful on cost-effectiveness conclusions.
Since 2008, a number of analyses have assessed the cost-effectiveness of daily oral PrEP in MSM in the United States [43–52]. Projected ICERs ranged from cost-saving to $1.4 million/QALY. Cost-effectiveness conclusions varied, depending largely on HIV incidence and PrEP adherence within the target population, PrEP efficacy, and PrEP drug price. Our analysis builds upon this existing literature and is unique in 3 key ways: (1) a specific focus on a YMSM population at increased risk of HIV infection, (2) consideration for the 2021 PrEP guidelines, and (3) incorporation of the generic PrEP drug price. The reviewed analyses used branded PrEP prices; the lowest monthly PrEP price considered in these analyses was in Ross et al [48] at $500 per month, significantly higher than the generic PrEP price of $30 per month that became available in late 2020. Indeed, most studies underscored the sensitivity of their findings to the PrEP drug price, acknowledging that conclusions should be reconsidered in the context of lower PrEP prices.
This analysis has several limitations. First, we assumed that there is no added benefit from regular PrEP visits that may impact HIV incidence, such as counseling on sexually transmitted infections and harm-reduction strategies. However, in reality, these visits may provide additional benefit to the PrEP intervention, which would likely strengthen our conclusions. Additionally, we assume 100% uptake for those on PrEP, mirroring the ATN 110/113 trials; however, current PrEP uptake in this population is much lower [53]. Additional scenarios varying uptake as low as 50% did not substantially impact modeled outcomes. Second, we did not explicitly model dynamic HIV transmission as a function of time-varying exposures to HIV. We did, however, examine the results both with and without the estimated impact of reduced HIV acquisition among PrEP users on onward transmission risk to the entire cohort, with little change in our overall conclusions; incorporating the secondary benefits of PrEP would render all scenarios more favorable both clinically and financially. Third, we chose to limit the PrEP program to 10 years because of uncertainty around the population being modeled as well as future changes to the PrEP program that may arise. Additional analyses with longer PrEP programs did not alter conclusions; however, modeled individuals could not restart PrEP once stopping, which could affect long-term outcomes. We elected to restrict our attention in the present analysis to HIV-related outcomes. However, we acknowledge that screening for non-HIV sexually transmitted infections is an important treatment and prevention adjunct of most PrEP programs. Future research might account for the additional costs of screening and care for such infections and weigh these against the benefits of both treatment and downstream secondary transmissions averted. Finally, we did not evaluate clinical differences between generic and branded PrEP, although worst-case-scenario analyses around generic PrEP toxicities did not change policy recommendations, nor did we evaluate long-acting injectable PrEP, for which data in the youngest adolescents are only just emerging.
In conclusion, we used a microsimulation model to project the clinical outcomes, costs, and cost-effectiveness of annual HIV screening versus daily oral PrEP with every-3-month screening in a population of YMSM. At the generic drug price, daily oral PrEP is consistently cost-saving compared with a Screening strategy, even at incidences lower than those reported in ATN 110/113. This clinical benefit is significant enough to mitigate the price of the drug. Taken together, these findings emphasize the clinical and economic value of expanding and promoting the broad use of daily oral PrEP in YMSM in the United States.
Supplementary Data
Supplementary materials are available at Clinical 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.
Supplementary Material
Click here for additional data file.
AN
Dr. Anne Neilan
Infectious Disease · HIV & AIDS MedicineMassachusetts General Hospitalannen