Introduction Since the announcement of the results of HIV pre-exposure prophylaxis (PrEP) trials and the HPTN052 early treatment for prevention trial, there have been crucial policy discussions about the use of antiretroviral (ARV) drugs to prevent HIV acquisition or transmission. With regards to PrEP, encouraging results were first reported for men and transgender women who have sex with men in the iPrEX trial [1], which showed a 44% (95% CI 15–63) reduction in HIV acquisition with a daily dose of tenofovir/emtricitabine (TDF/FTC). In two large trials, the Partners PrEP [2] and TDF2 [3] studies, PrEP was found to be effective in reducing the risk of heterosexual HIV transmission using either TDF or TDF/FTC daily (Partners PrEP) and TDF/FTC daily (TDF2). However, FEM-PrEP [4], a trial recruiting heterosexual women in South Africa, Tanzania, and Kenya for daily TDF/FTC was closed prematurely in 2011 for futility as was the oral TDF arm of the VOICE trial [5] in women in South Africa, Uganda, and Zimbabwe. Two topical PrEP trials have tested the efficacy of 1% TDF gel and a third, FACTS001 [6], is currently recruiting women in South Africa. The CAPRISA 004 trial [7] in Kwa Zulu-Natal found that pre- and post-coital vaginal TDF gel reduced women's acquisition risk by 39% (95% CI 6–60) but the VOICE trial stopped its gel arm when it became evident that daily gel use was safe but not effective [8]. Clinical guidance on oral PrEP has already been offered by the US Centers for Disease Control and Prevention, the Southern African HIV Clinicians Society, World Health Organization (WHO), and the British Association for Sexual Health and HIV [9]–[13]. An advisory panel to the US Food and Drug Administration recently recommended oral TDF/FTC for preventive use among people at higher risk of HIV exposure [14]. As PrEP emerges as an option for inclusion in the HIV prevention toolbox, it is important for national policy and decision makers to identify where PrEP may fit best within already established HIV prevention programming (and budgets) and the potential implications of introducing such policy changes. In particular, decision makers need information translating the trial results into potential population-level impact and cost-effectiveness to ensure that any additional investment will have the maximum possible effect on the epidemic. Economic and mathematical models provide a framework to integrate information on efficacy, effectiveness, costs, and patient outcomes to support decision making and resource allocation [15]. However, due to their complexity, dependence on assumptions made, and inherent uncertainties, generalising results from these models can be difficult. In this review, we aim to assess published cost-effectiveness models that have evaluated the expected health gains and costs of PrEP interventions. Specifically, our objectives are: (1) to describe modelling approaches of cost-effectiveness analyses of PrEP; (2) to compare the effects of epidemiological and cost assumptions on cost-effectiveness results; and (3) to explore the potential impact on cost-effectiveness estimates of five issues raised by policy makers [16]–[18] when considering PrEP implementation: prioritisation, adherence, behaviour change, toxicity, and resistance. Methods We performed a systematic review of the published literature following the protocol available in Text S2 and adhering to the PRISMA guidelines for reporting of systematic reviews (Text S1: PRISMA checklist) [19] and guidelines for appraisal of economic evaluations [20]. Search Strategy, Inclusion Criteria, and Study Selection A broad strategy using both MeSH headings and free text, with no language limitations, was used to search PubMed/Medline, ISI Web of Knowledge (including Web of Science, Current Contents Connect, Derwent Innovations Index, CABI: CAB Abstracts, and Journal Citation Reports), Centre for Reviews and Dissemination databases (including DARE - Database of Abstracts of Reviews of Effects, NHS EED - NHS Economic Evaluation Database, and HTA database - health technology assessments), EconLIT, and region-specific databases (African Index Medicus, Eastern Mediterranean Literature (WHO), Index Medicus for South-East Asia Region, LILACS for Latin America). Our searches covered all published research up to the last search performed 14 January 2013 with no limitations on publication date. The following keywords were used: “cost” AND “tenofovir OR pre-exposure prophylaxis OR chemoprophylaxis OR PrEP” AND “HIV.” Citations and bibliographies of full text reports retrieved were reviewed for additional relevant articles. Abstracts from international conferences identified in the searches were also reviewed, as was the website of the International AIDS Economic Network. Experts were consulted for additional studies. We included all modelling studies reporting both cost and impact of a potential roll-out of a PrEP programme. We excluded those studies where costs were not assessed. No restrictions were made on the type of model, geography, mode of transmission, or impact (effectiveness) metric chosen. We included studies looking at both topical and systemic PrEP products. Full published papers were eligible, as well as abstracts from conferences providing sufficient information. Two authors (GBG and AB) screened titles and abstracts to identify potentially relevant articles. Full text reports of these articles were assessed independently for inclusion. Data Extraction and Analysis Data were extracted from selected studies by one reviewer (GBG) into prepared data sheets and independently cross-checked by a second assessor (AB). For conference abstracts selected for inclusion, we contacted the first author listed for further information. Extracted information on the study design included the type of study, viewpoint of analysis, timeframe, setting and population, background HIV prevalence or incidence, mode of HIV transmission, and a detailed description of alternative programmes compared in the studies (baseline scenario and PrEP scenario). We also tabulated data on the impact including risk heterogeneity, efficacy or effectiveness of PrEP, adherence (to programme or individual), behavioural change expected after introduction of PrEP, resistance, toxicity due to PrEP use, and disability-adjusted life year (DALY)/quality-adjusted life year (QALY) assumptions. A description of economic assumptions includes expected drug cost, other service costs, costs above service level, downstream antiretroviral treatment (ART) costs averted, discount rates, and, finally, cost-effectiveness results by metric and the conclusions presented in each publication. Prioritised scenarios were defined as those scenarios where PrEP was offered to specific sub-populations within the population modelled. While providing a critical assessment and narrative review of the studies included, we did not attempt to perform a meta-analysis due to the variability across the studies in reporting outcomes. Therefore, we adjusted estimates of cost-effectiveness for inflation to US$2012 to be able to compare studies from different years [21]. For those studies reporting cost/DALY averted, cost/QALY averted, or cost/life-year saved (LYS), we compared the estimates to a benchmark for cost-effectiveness [22] of one times the gross domestic product per capita (GDP/capita) per DALY averted, per QALY gained, or per LYS, depending on the unit of outcome used by each study. While DALYs, QALYs, and LYS are not equivalent, and decision rules vary by setting, this gives a broad indication of potential cost-effectiveness. The values for current GDP/capita were sourced from the World Bank databank for each country [23]. There is much controversy around decision rules [24], and while the comparison against GDP is the conventional approach, it should be noted that this may not represent the true opportunity cost in countries where less cost-effective health interventions are not being implemented at scale. Results We screened 961 titles and abstracts retrieved from 14 electronic databases. After performing web searches and consulting experts in the field, 36 full text articles were evaluated. We also reviewed the reference lists and citations of these articles. Of these 36, 13 studies were included in the review [25]–[37]: 11 peer-reviewed publications and two peer-reviewed conference abstracts (Figure 1). Articles excluded are listed in Table S1 and a summary of conclusions of the articles included are presented in Table S2. 10.1371/journal.pmed.1001401.g001 Figure 1 Flow diagram of study selection. Region-specific databases can be accessed as follows: African Index Medicus, http://indexmedicus.afro.who.int/; Eastern Mediterranean Literature, http://www.emro.who.int/; Index Medicus for South-East Asia Region, http://www.hellis.org/; LILACS, Latin America, http://www.bireme.br/iah2/homepagei.htm. We present in Tables 1 to 4 the data extracted from the studies reviewed by study design, description of alternative programmes compared, impact, and cost assumptions. All studies were published between 2007 and 2013 and modelled the impact and cost, from a health care provider perspective, of PrEP scale-up in diverse settings. These settings included: heterosexual transmission in generalised epidemics in sub-Saharan Africa—the Southern Africa region [25], South Africa [28],[30],[31],[32],[36],[37]), and other modes of transmission in concentrated epidemics—among people who inject drugs (PWID) in Ukraine [33]; and men who have sex with men (MSM) in the USA [26],[29],[34],[35] and in Peru [27]. Timeframes varied from 5 to 20 y. All studies focused the models on high prevalence/incidence populations (Table 1). 10.1371/journal.pmed.1001401.t001 Table 1 Study design. Reference Study Type Setting/MoT Population Timeframe HIV Incidence/Prevalence Generalised epidemics in southern Africa Abbas [25] Deterministic simulation; Risk heterogeneity by age, sex, sexual behaviour, and HIV drug resistance Southern Africa/Heterosexual 15–49 y; General population 10 y Prevalence: 20%a Pretorius [30] Deterministic simulation; Risk heterogeneity by age and sex South Africa/Heterosexual 15–49 y; General population 10 y (programme scale-up: 5 y) Prevalence: ±20% in 2008b Hallett [28] Microsimulation; Risk heterogeneity by age, sex, sexual behaviour, and conception intentions or pregnancyc South Africa/Heterosexual Serodiscordant couples Each person is tracked until his/her 50th y n/a Williams [32] Deterministic simulation; Risk heterogeneity not included South Africa/Heterosexual 15–49 y; General population From 2012 to 2020 (scale-up by 2015) Prevalence: approximately 16% in 2012b Walensky [31] Monte Carlo state simulation; Risk heterogeneity by age South Africa/Heterosexual Women at higher risk Each person is tracked until death Incidence: 25 y, 1.0% Alistar [37] Compartmental dynamic simulation Risk heterogeneity by sexual behaviour behaviour (number of partners and condom use) South Africa/Heterosexual 15–49 y; General population 20 y Initial prevalence in adults: 17.9% and initial incidence: 1.4% Cremin [36] Deterministic simulation;Risk heterogeneity by age, sex, male circumcision status, behavioural; risk (partner change rate, condom use) South Africa/Heterosexual 15–54 y; General population 10 y (programme scale-up: 5 y) Age- and sex-specific prevalence peaking at 30–44 y (women: >40% and 35–44 y men: >30%). Concentrated epidemics among MSM in high-income countries Desai [26] Stochastic simulation; Risk heterogeneity by age, sexual risk behaviourd USA (NYC)/MSM 13–40 y; High risk MSM 5 y Prevalence: 14.6% in 2008 Paltiel [34] Monte Carlo state simulation; Risk heterogeneity by age (assumed higher incidence by age group) USA/MSM Average 34 y; High risk MSM Each person is tracked until death Incidence: 1.6% annual Koppenhaver [29] Compartmental dynamic simulation Risk heterogeneity not included USA (urban)/MSM 13–40 y; All MSM 20 y Prevalence: 17.5% Juusola [35] Deterministic simulation; Risk heterogeneity by sexual behavioure USA/MSM 13–64 y; 20 y Prevalence: 12.3%; Incidence: 0.8% annual Concentrated epidemics among MSM in low- and middle-income countries Gomez [27] Deterministic simulation; Risk heterogeneity by sexual behaviour Peru (Lima)/MSM All MSM 10 y (programme scale-up: 5 y) Incidence: MMSM, 1%; MMSW, 2.5%; SW, 3.1%; Trans, 7.3% Concentrated epidemics among PWID in low- and middle-income countries Alistar [33] Compartmental dynamic simulation Risk heterogeneity by IDU behaviour Ukraine/IDU and heterosexual 15–49 y 20 y Initial prevalence: 41.2% PWID, 1% general population Study type refers to the type of model and the inclusion of risk heterogeneity in the population modelled. Setting/MoT refers to the geographical setting and the mode of transmission modelled. a Female∶male ratio 1.66, based on data from urban antenatal care attendees in Zambia. b Model initiated at a high prevalence then fitted to Department of Health data. c Two types of couples were defined: (1) lower risk couples based on reported data from the Partners in Prevention HSV/HIV Transmission Study [49], and (2) couples at a higher risk reflecting a higher incidence. “Partners in Prevention” assumptions: incidence low (1.8/100 person-years at risk, high condom use); “more typical couples” assumptions: 50% of serodiscordant couples involved HIV-1 infected men. Compared to the partners in prevention cohort: condom use within the stable partnership was reduced by 25%, 50% more of the HIV-1 uninfected partners in couples had external partners, and frequency of unprotected sex with external partners was doubled. d Very high risk was defined as a participant reporting unprotected sex in the last 6 mo or in exchange for money or drugs, anonymous sex, ≥5 sexual or needle sharing partners, and/or an STI diagnosis in the last 6 mo. e The authors run the model separately for low risk and high risk populations. Therefore PrEP use in one group does not have an impact on the other (the mixing is considered totally assortative). IDU, injection drug use; MMSM, men who mostly have sex with men; MMSW, men who mostly have sex with women; n/a, not applicable; SW, sex worker; Trans, transgender or trans-sexual; USA (NYC), United States of America (New York City). 10.1371/journal.pmed.1001401.t002 Table 2 Alternative programmes compared. Reference Base Comparison Scenario PrEP Intervention PrEP Regimen Prioritisation Coverage Generalised epidemics in southern Africa Abbas [25] No PrEP. ART was not modelled. Once daily oral dosing No prioritisation: general population. By sexual activity: two highest sexual activity groups prioritised. By age: 15–20 y group prioritised. Percent of the population using PrEP: Optimistic scenario, 75%; Neutral scenario, 50%; Pessimistic scenario, 25% Pretorius [30] No PrEP. ART coverage expands at its current rate. ART efficacy: 90% reduction in transmission probability. Once daily oral dosing No prioritisation: 15–35 y; By age: 15–25 y, or 25–35 y Percent of women using PrEP: 20%, dropout rate:1.5% Hallett [28] No PrEP. ART initiation for the infected partner when CD4 cell count fell below 200 cells/ml. In a separate scenario, expansion of eligibility criteria for ART initiation was included (below 350 CD4 cells/ml). Once daily oral dosing No prioritisation: Always use PrEP after diagnosis partner. By timing: Up to partner's ART init; up to partner's ART init+1 y; during conception/pregnancya Percent of the population using PrEP: see prioritisation Williams [32] No PrEP. The scale-up of ARV therapy was not modelled. Vaginal gel, two doses pericoitally PrEP used only by women Percent of sex acts protected: High: 90%, Medium: 50%, Low: 25% Walensky [31] No PrEP. Patients identified as HIV infected received ART as per guidelines. Vaginal gel, two doses pericoitally PrEP used only by women. By age: ≤25 y (high inc. group) Cohort-wide PrEP use continues until HIV infection or death. Alistar [37] No PrEP. 40% HIV infected patients received ART as per guidelines. ART efficacy: 95% reduction in transmission probability. Once daily oral dosing No prioritisation: general use; By sexual activity: groups of high number of partners and low condom use Rate of recruitment into the program: 25%, 50%, 75%, 100%. Included a rate of dropout from PrEP. Cremin [36] ART efficacy: 96% reduction in transmission probability. Baseline scenarios varied: from status quo with current scale-up of ART to counterfactual including MC and ART scale-up. All scenarios included a 7/100 PY dropout rate while on ART. Once daily oral dosing No prioritisation: 15–54 y; By age: 15–24 y Percent of the population group using PrEP: 40%, 80% Concentrated epidemics among MSM in high-income countries Desai [26] No PrEP. The scale-up of ARV therapy was not modelled. Once daily oral dosing No prioritisation—results not shown. Results for scenarios targeting high risk MSM only. 25% high riskb; (5.2% of all MSM)c; Discontinuation rate: 40% per year Paltiel [34] No PrEP. Patients identified as HIV infected received ART as per guidelines. Once daily oral dosing No prioritisation: all MSM. By age: US$15,000 per person-year in the USA) (Table 4). We present all cost-effectiveness estimates in Table 5 by epidemiological context and scenario modelled. 10.1371/journal.pmed.1001401.t005 Table 5 Cost-effectiveness estimates by scenario. Reference Scenario Description: Prioritisation Estimate Measure US$ in Publication 2012US$ Generalised epidemics in southern Africa Abbas [25] Pessimistic: high sexual activity group Cost/infection averted 2,949–9,923 3,450–11,609 Pessimistic: 15–20 y Cost/infection averted 20,202–67,970 23,636–79,525 Pessimistic: no prioritisation Cost/infection averted 20,164–67,842 23,591–79,375 Neutral: high sexual activity group Cost/infection averted 1,160–3,904 1,357–4,567 Neutral: 15–20 y Cost/infection averted 8,968–30,173 10,492–35,302 Neutral: no prioritisation Cost/infection averted 9,629–32,398 11,265–37,905 Optimistic: high sexual activity group Cost/infection averted 638–2,147 746–2,512 Optimistic: 15–20 y Cost/infection averted 5,723–19,254 6,695–22,527 Optimistic: no prioritisation Cost/infection averted 6,812–22,918 7,970–26,814 Pretorius [30] Optimistic: women 15–25 y, no behaviour change Cost/infection averted >25,000 >26,625 Optimistic: women 15–35 y, no behaviour change Cost/infection averted >22,500 >23,963 Optimistic: women 25–35 y, no behaviour change Cost/infection averted >20,000 >21,300 Medium efficacy: women 25–35 y, behaviour change Cost/infection averted >30,000 >31,950 Hallett [28] Efficacy range, high risk: conception or pregnancy use Cost/infection averted −6,000 to 8,000 −6,192 to 8,256 Efficacy range, low risk: conception or pregnancy use Cost/infection averted −2,000 to 12,000 −2,064 to 12,384 Efficacy range, high risk: up to ART initiation Cost/infection averted −2,200 to 21,000 −2,270.4 to 21,672 Efficacy range, high risk: always use PrEP Cost/infection averted 0–26,000 0–26,832 Efficacy range, low risk: always use PrEP Cost/infection averted 6,000–66,000 6,192–68,112 Optimistic, low risk, high ART cost: up to ART initiation Cost/infection averted 3,000 3,096 Optimistic, low risk, high ART cost: up to ART initiation +1 y Cost/infection averted 3,000 3,096 Optimistic, high risk: up to ART initiation Cost/QALY gained −200 to 500 −206 a to 516 a Optimistic, low risk: up to ART initiation Cost/QALY gained 260–1,600 268 a –1,651 a Pessimistic, high risk: up to ART initiation Cost/QALY gained 700–1,900 722 a –1,960 a Pessimistic, low risk: up to ART initiation Cost/QALY gained 2,500–4,900 2,580 a –5,056 a Williams [32] CAPRISA efficacy: high coverage Cost/infection averted 420–2,982 447–3,175 CAPRISA efficacy: low coverage Cost/infection averted 562–4,222 598–4,496 CAPRISA efficacy: high coverage Cost/DALY averted 18–130 19 a –138 a CAPRISA efficacy: low coverage Cost/DALY averted 27–181 28 a –193 a Walensky [31] CAPRISA efficacy, test freq 3 mo: high incidence women Cost/life year saved 1,600 1,704 a CAPRISA efficacy, test freq 1 mo: high incidence women Cost/life year saved 2,700 2,876 a Alistar [37] b PrEP: no prioritisation recruitment rate 25% to 100%, no ART expansion Cost/QALY gained 1,200 1,200 a PrEP: high risk group recruitment rate 50% to 100%, no ART expansion Cost/QALY gained CS CS a PrEP: no prioritisation recruitment rate 25% to 100%, ART +25% as per guidelines Cost/QALY gained 980–1,050 980 a –1,050 a PrEP: high risk group recruitment rate 100%, ART +25% as per guidelines Cost/QALY gained 50 50 a PrEP: no prioritisation recruitment rate 25% to 100%, ART +50% as per guidelines Cost/QALY gained 900–1,000 900 a –1,000 a PrEP: high risk group recruitment rate 100%, ART +50% as per guidelines Cost/QALY gained 160 160 a PrEP: no prioritisation recruitment rate 25% to 100%, ART +75% as per guidelines Cost/QALY gained 860–970 860 a –970 a PrEP: high risk group recruitment rate 100%, ART +75% as per guidelines Cost/QALY gained 210 210 a PrEP: no prioritisation recruitment rate 25% to 100%, ART +100% as per guidelines Cost/QALY gained 840–950 840 a –950 a PrEP: high risk group recruitment rate 100%, ART +100% as per guidelines Cost/QALY gained 230 230 a PrEP: no prioritisation recruitment rate 25% to 100%, universal ART +25% Cost/QALY gained 810–940 810 a –940 a PrEP: high risk group recruitment rate 100%, universal ART +25% Cost/QALY gained 220 220 a PrEP: no prioritisation recruitment rate 25% to 100%, universal ART +50% Cost/QALY gained 760–900 760 a –900 a PrEP: high risk group recruitment rate 100%, universal ART +50% Cost/QALY gained 280 280 a PrEP: no prioritisation recruitment rate 25% to 100%, universal ART +75% Cost/QALY gained 740–890 740 a –890 a PrEP: high risk group recruitment rate 100%, universal ART +75% Cost/QALY gained 290 290 a PrEP: no prioritisation recruitment rate 25% to 100%, universal ART +100% Cost/QALY gained 740–880 740 a –880 a PrEP: high risk group recruitment rate 100%, universal ART +100% Cost/QALY gained 300 300 a Cremin [36] c PrEP: no prioritisation, cov 4.4% of 15–54 y (baseline: status quo, current ART scale-up) Cost/infection averted 9,390 9,390 PrEP: prioritisation, cov 7.3% of 15–24 y (baseline: status quo, current ART scale-up) Cost/infection averted 10,540 10,540 No PrEP, 80% universal ART (baseline: 80% ART200 and 80% MC) Cost/infection averted 10,530 10,530 PrEP: 15–24 y cov 40%, 80% universal ART (baseline: 80% ART200, 80% MC, 80% ART350) Cost/infection averted 39,900 39,900 PrEP: 15–54 y cov 80%, 80% universal ART (baseline: 80% ART200, 80% MC) Cost/infection averted 20,500 20,500 Concentrated epidemics among MSM in high-income countries Desai [26] d Exposure, pessimistic: high adherence Cost/QALY gained 6,661–36,268 7,793 e –42,433 e Exposure, pessimistic: medium adherence Cost/QALY gained 55,167–84,774 64,545 f –99,185 f Exposure, pessimistic: low adherence Cost/QALY gained 113,601–143,208 132,913 f–167,553 Adherence, pessimistic: high adherence Cost/QALY gained CS–8,158 CS e –9,545 e Adherence, pessimistic: medium adherence Cost/QALY gained CS–10,327 CS e –12,082 e Adherence, pessimistic: low adherence Cost/QALY gained CS–13,499 CS e –15,793 e Basic, pessimistic: high adherence Cost/QALY gained CS–15,099 CS e –17,665 e Basic, pessimistic: medium adherence Cost/QALY gained 17,168–46,775 20,086 e –54,726 f Basic, pessimistic: low adherence Cost/QALY gained 66,896–96,502 78,268 f –112,907 Exposure, optimistic: high adherence Cost/QALY gained CS–9,925 CS e –11,612 e Exposure, optimistic: medium adherence Cost/QALY gained 13,307–42,914 15,569 e –50,209 f Exposure, optimistic: low adherence Cost/QALY gained 46,502–76,109 54,407 f –89,047 f Adherence, optimistic: high adherence Cost/QALY gained CS CS e Adherence, optimistic: medium adherence Cost/QALY gained CS CS e Adherence, optimistic: low adherence Cost/QALY gained CS CS e Basic, optimistic: high adherence Cost/QALY gained CS–1,009 CS e –1,180 e Basic, optimistic: low adherence Cost/QALY gained 37,947–67,553 44,398 e –79,037 f Basic, optimistic: medium adherence Cost/QALY gained CS–28,393 CS e –33,220 e Paltiel [34] Medium efficacy: no prioritisation Cost/QALY gained 298,000 359,984 High efficacy: no prioritisation Cost/QALY gained 107,000 129,256 f Medium efficacy, low cost Cost/QALY gained 114,000 137,712 f Medium efficacy: young Cost/QALY gained 189,000 228,312 Koppenhaver [29] High adherence: no prioritisation Cost/QALY gained 353,739 376,732 iPrEX adherence: no prioritisation Cost/QALY gained 570,273 607,341 Juusola [35] Cov 100%, PrEP cost US$26/d, no resistance: high risk MSM Cost/QALY gained 52,443 55,852 f Cov100%, PrEP cost US$26/d, no resistance: no prioritisation Cost/QALY gained 216,480 230,551 Cov 100%, high eff, PrEP cost US$26/d, no resistance: high risk MSM Cost/QALY gained 35,080 37,360 e Cov 100%, high eff, PrEP cost US$26/d, no resistance: no prioritisation Cost/QALY gained 146,228 155,733 Cov 100%, PrEP cost US$15/d, no resistance: no prioritisation Cost/QALY gained 131,277 139,810 f Cov 100%, PrEP cost US$50/d, no resistance: high risk MSM Cost/QALY gained 104,516 111,310 f Cov 100%, PrEP cost (50% ARV), no resistance: high risk MSM Cost/QALY gained 25,165 26,801 e Cov 100%, PrEP cost (75% ARV), no resistance: high risk MSM Cost/QALY gained 38,804 41,326 e Cov 100%, no resistance, 8% reduction QoL: high risk MSM. Cost/QALY gained 95,006 101,181 f Cov 100%, PrEP cost US$26/d, resistance: high risk MSM Cost/QALY gained 57,861 61,622 f Cov 100%, PrEP cost US$26/d, resistance: no prioritisation Cost/QALY gained 233,040 248,188 Cov 50%, PrEP cost US$26/d, no resistance: high risk MSM Cost/QALY gained 44,556 47,452 e Cov50%, PrEP cost US$26/d, no resistance: no prioritisation Cost/QALY gained 188,421 200,668 Cov 50%, high eff, PrEP cost US$26/d, no resistance: high risk MSM Cost/QALY gained 26,766 28,506 e Cov 50%, high eff, PrEP cost US$26/d, no resistance: no prioritisation Cost/QALY gained 120,080 127,885 f Cov 50%, PrEP cost US$15/d, no resistance: no prioritisation Cost/QALY gained 113,935 121,341 f Cov 50%, PrEP cost US$50/d, no resistance: high risk MSM Cost/QALY gained 89,658 95,486 f Cov 50%, PrEP cost (50% ARV), no resistance: high risk MSM Cost/QALY gained 20,930 22,290 e Cov 50%, PrEP cost (75% ARV), no resistance: high risk MSM Cost/QALY gained 32,743 34,871 e Cov 50%, no resistance, 8% reduction QoL: high risk MSM. Cost/QALY gained 72,762 77,492 f Cov 50%, PrEP cost US$26/d, resistance: high risk MSM Cost/QALY gained 56,492 60,164 f Cov 50%, PrEP cost US$26/d, resistance: no prioritisation Cost/QALY gained 226,325 241,036 Cov 20%, PrEP cost US$26/d, no resistance: high risk MSM Cost/QALY gained 40,279 42,897 e Cov20%, PrEP cost US$26/d, no resistance: no prioritisation Cost/QALY gained 172,091 183,277 Cov 20%, high eff, PrEP cost US$26/d, no resistance: high risk MSM Cost/QALY gained 22,374 23,828 e Cov 20%, high eff, PrEP cost US$26/d, no resistance: no prioritisation Cost/QALY gained 105,066 111,895 f Cov 20%, PrEP cost US$15/day, no resistance: no prioritisation Cost/QALY gained 103,841 110,591 f Cov 20%, PrEP cost US$50/d, no resistance: high risk MSM Cost/QALY gained 81,593 86,897 f Cov 20%, PrEP cost (50% ARV), no resistance: high risk MSM Cost/QALY gained 18,637 19,848 e Cov 20%, PrEP cost (75% ARV), no resistance: high risk MSM Cost/QALY gained 29,458 31,373 e Cov 20%, no resistance, 8% reduction QoL: high risk MSM Cost/QALY gained 62,431 66,489 f Cov 20%, PrEP cost US$26/d, resistance: high risk MSM Cost/QALY gained 78,884 84,011 f Cov 20%, PrEP cost US$26/d, resistance: no prioritisation Cost/QALY gained 303,091 322,792 Concentrated epidemics among MSM in low- and middle-income countries Gomez [27] Low coverage: high prioritisation Cost/DALY averted 403–637 415 g –657 g Low coverage: some prioritisation Cost/DALY averted 447–707 461 g –729 g Low coverage: no prioritisation Cost/DALY averted 1,076–1,702 1,110 g –1,756 g High coverage: high prioritisation Cost/DALY averted 665–1,052 686 g –1,085 g High coverage: some prioritisation Cost/DALY averted 886–1,400 914 g –1,445 g High coverage: no prioritisation Cost/DALY averted 1,125–1,779 1,161 g –1,835 g Concentrated epidemics among PWID in low- and middle-income countries Alistar [33] MMT 25%, no PrEP Cost/QALY gained 530 546 h MMT 25%, ART 80% (for IDU and general population), no PrEP Cost/QALY gained 870 896 h MMT 25%, ART 80% (for IDU and general population), PrEP 25% to 50% Cost/QALY gained 3,080–3,910 3,172 h –4,027 i PrEP 25% to 50% Cost/QALY gained 14,590–14,680 15,028–15,120 MMT 25%, PrEP 25% to 50% Cost/QALY gained 4,800–6,100 4,944 i –6,283 i ART 80% (for IDU and general population), PrEP 25% to 50% Cost/QALY gained 3,290–4,210 3,389 h –4,336 i Thresholds used to determine cost-effectiveness, based on World Bank database [23]. Bold-black signifies an estimate is cost-effective or very cost-effective with regards to the country-specific threshold. a For South Africa, an intervention is considered very cost-effective at a threshold of less than 1× GDP per capita, US$8,070. b In Alistar et al., several scenarios were considered for ART recruitment rates of 25%, 50%, 75%, and 100% in addition to the 40% status quo coverage as per guidelines and following universal access. c In Cremin et al., several scenarios were considered for ART coverage. ART200: coverage of ART in HIV-infected people starting at CD4 count of <200 cells/ml; ART350: coverage of ART in HIV-infected people starting at CD4 count of <350 cells/ml; universal ART: coverage of ART in HIV-infected people starting at any CD4 count level. d In Desai et al., the authors considered three effectiveness mechanisms: basic, adherence-dependent, and exposure-dependent. e For USA, an intervention is considered very cost-effective at a threshold of less than 1× GDP per capita, US$48,442. f For USA, an intervention is considered cost-effective between 1× GDP per capita, US$48,442 and 3× GDP per capita, US$145,326. g For Peru, an intervention is considered very cost-effective at a threshold of less than 1× GDP per capita, US$ US$6,009. h For Ukraine, an intervention is considered very cost-effective at a threshold of less than 1× GDP per capita, US$3,615. i For Ukraine, an intervention is considered cost-effective between 1× GDP per capita, US$3,615 and 3× GDP per capita, US$10,845. cov., coverage; CS, cost saving; freq, frequency; MC, male circumcision; MMT, methadone maintenance treatment; QoL, quality of life; resist., resistance. Generalised Epidemics in Southern Africa (n = 7) Studies on topical PrEP and two studies on oral PrEP suggest the intervention to be cost-effective (topical PrEP: <200 US$/DALY [32], <3,000 US$/LYS [31]; oral PrEP: <5,000 US$/QALY [28], <2,800 US$/QALY [37]) using benchmarks for cost-effectiveness specific to South Africa [22]. Three studies reported cost/infection averted only, estimates ranging from US$1,000 to 39,900 [25],[30],[36]. For topical PrEP, the two studies presented different estimates of cost-effectiveness: less cost-effective in Walensky et al. [31] (