Secondary Logo

Journal Logo


The epidemiological impact and cost-effectiveness of HIV testing, antiretroviral treatment and harm reduction programs

Li, Jinghuaa,b; Gilmour, Stuarta; Zhang, Haiqiangc; Koyanagi, Aia; Shibuya, Kenjia

Author Information
doi: 10.1097/QAD.0b013e3283574e54



HIV/AIDS is a growing public health problem in China. A total of 326 000 people living with HIV/AIDS (PLWHA) had been reported in China by the end of 2009, of whom 107 000 had progressed to AIDS [1]. In 2009, HIV/AIDS was the leading cause of infectious disease death, with 50% of all deaths due to infectious diseases attributed to AIDS [2]. It is estimated that there were a total of 740 000 PLWHA in China at the end of 2009, including unidentified PLWHA, with an HIV population prevalence of 0.057% (uncertainty range/UR: 0.042–0.071%) [1]. The number of new infections is also increasing steadily in China [3]. In 2009, 48 000 (UR: 41 000–55 000) people were estimated to have become newly infected with HIV [1].

Significant variations in epidemic patterns between provinces and risk-groups are noticeable in China. The five provinces with the highest HIV prevalence (Yunnan, Guangxi, Guangdong, Xinjiang, and Henan) account for 53.4% of the total number of PLWHA [4]. Remarkable differences in prevalence also exist between different risk groups, especially among injecting drug users (IDUs), female sex workers (FSW), and men who have sex with men (MSM), with prevalences of 9.3% [1], 0.6% [4], and 5.0% [5], respectively.

After years of insufficient attention, the Chinese government has established important responses to the HIV/AIDS epidemic [3], including the Regulations on HIV/AIDS Prevention and Treatment, China's 5-year Action Plan for Reducing and Preventing the Spread of HIV/AIDS (2006–2010), and the Four Frees, One Care policy, to ensure a comprehensive intervention combining prevention, treatment and support with government taking the lead, multiple sectors meeting their individual responsibilities and full-society participation [1]. Although still at very low levels of coverage, China's harm reduction program has achieved significant results: since 2004, when the first eight methadone maintenance treatment (MMT) clinics were established, the number of MMT clinics has increased rapidly, reaching 633 by June 2009 [6].

Urgent attention needs to be given to changing the main modes of transmission of HIV. It has been estimated that nearly 50% of all PLWHA were originally infected through injecting drug use, whereas about 7% of all cases might be due to unprotected sex among MSM [7]. Significantly increasing HIV prevalence have been found among MSM in Jiangsu [8] and Shandong [9] provinces, the city of Beijing [5], and in Chongqing City [10]. Many studies have suggested that HIV infections among IDU can be prevented through comprehensive harm reduction programs, based around pharmacotherapy and needle/syringe programs [11], but it is more difficult to control HIV epidemics that are spread sexually through behavior change interventions alone [12]. As a form of secondary prevention, voluntary counseling and testing (VCT) and early initiation of treatment may work well among MSM and FSW in many countries [13,14].

Routine VCT can identify HIV infections early and link identified HIV-positive individuals to treatment and counseling, which could result in reduced risky sexual behavior among participants [15,16]. In addition, early initiation of antiretroviral treatment (ART) can lower patients’ viral load and reduce their infectivity [14,17]. Furthermore, treatment can reduce mortality and improve quality of life among PLWHA.

Nonetheless, HIV testing and ART coverage remains low in China – in 2005, only 25% of those in need received ART [18] – and testing rates in the past 12 months were only 37 and 2% in high-risk groups and low-risk groups, respectively, in 2009 [1].

Even though the current HIV epidemic in China remains at a very low level nationally, prevalence in IDU is still high, and the epidemic among MSM may be growing [1]. It is important to explore future epidemiological trends at the national level by risk groups, as well as to assess the impact of different interventions. In this study, we compared the epidemiological impact and cost-effectiveness of four programs in the Chinese context.


Model framework

We developed a deterministic compartmental mathematical model to capture the dynamics of HIV transmission through sexual contact and needle sharing. The compartment diagram can be found in Supplemental Digital Content 1, Susceptible individuals were assumed to become infected in three ways: heterosexual contact, homosexual contact among MSM, and needle-sharing among IDU. Heterosexual transmission was assumed to occur among MSM as well. Once infected, individuals were assumed to progress through the natural course of HIV infection. All individuals were screened and tested at every disease stage at specified testing/screening and counseling rates. This model is an extension of a pre-existing HIV transmission model [19] with appropriate modifications made to reflect the context in China (see Supplemental Digital Content 1, Thus, the fundamental compartmental structure is the same, but the nature of the risk groups and their interactions has been changed. People identified as HIV-positive were to receive ART treatment at a specified ART utilization rate, assumed to be 30% in the situation with no intervention (the base case), and when their CD4 cell count fell below 350 cells/μl, with an additional 5% annual entry rate thereafter. Parameters for epidemiological factors, behavior, the mortality rate, and the probability of disease transmission were drawn from previous publications (Table 1). Data sources are referenced in Table A-1, Supplemental Digital Content 1,

Table 1-a:
Key model parameters.
Table 1-b:
Key model parameters.

We coded this program using Matlab R2010a and then ran the model with a 30-year time horizon. The baseline year of this study is 2010. We also conducted numerical calibration analysis on the data for the number of PLWHA, annual deaths due to HIV/AIDS, and new infections, using values reported by the Joint United Nations Programme on HIV/AIDS (UNAIDS) for China. Detailed results of calibration analyses, the definition of all symbols, and details of the equations are presented in Supplemental Digital Content 1, In order to capture epidemiological trends in HIV, prevalence, incidence, and cumulative incidence were calculated.

Study population

The target population programmed in our model was adults aged 15–64 years. Based on current understanding of the HIV epidemic in China, we subdivided the target population into seven groups: MSM, male IDUs (MIDUs), clients of female sex workers (CSW), low-risk men, female IDUs (FIDUs), FSW, and low-risk women. Each risk group was then subdivided into 10 groups by their disease, diagnosis and treatment status.

Cost-effectiveness analysis

We compared four different intervention types, referred to in the text by the following titles:

  1. Expanded VCT only: expanded VCT was modeled as being offered to 50% of the high-risk group and 15% of the low-risk group, annually. Four substrategies were included under this strategy (Table 2) [20,21].
  2. T3-9
    Table 2:
    Benefits, cost and saving in healthcare cost of different strategies.
  3. Expanded ART treatment only: ART treatment among people whose CD4 cell count fell below 350 cells/μl was expanded with a utilization rate of 50%.
  4. Harm reduction program: MMT and needle and syringe programs (NSP) were expanded to cover the entire population of IDU at specified rates of uptake (see Supplemental Digital Content 1, and specified success rates. Twelve-month retention in MMT was assumed to be 50% [22,23].
  5. Combination strategies: combinations of any two or more of the above three strategies.

These were compared against a control situation, referred to as the ‘base case,’ that was defined to reflect the current situation in China. As harm reduction programs in China remain at a preliminary stage, the base case was modeled based on the assumption that there was essentially no harm reduction program. Assumed rates of testing, treatment, and risk behavior for the base case are summarized in Table 1.

For cost-effectiveness, we calculated incremental quality-adjusted life years (QALYs) and costs of each intervention relative to the base case, with an annual discount rate of 3%. We did not include cost savings in the cost-effectiveness calculation, but reported them separately to the direct costs of the interventions. All costs were reported in 2010 International Dollars (Int.$), calculated using the International Monetary Fund gross domestic product (GDP) deflator and implied purchasing power parity conversion rates. All strategies were also compared in terms of total new infections prevented compared to the base case. In addition, we also explored the saving from healthcare costs and compared it with the cost of the interventions to identify whether any of the interventions might lead to overall cost savings.

Multivariate sensitivity analysis

The model we developed included some hidden populations, such as MSM and IDU. Because some behavioral variables, such as number of sexual partners, condom use rates, annual injections, and needle-sharing rates, are not well understood in these groups, we conducted sensitivity analysis of both epidemiological trends and cost-effectiveness on these behavioral variables and cost values. We used Latin Hypercube sampling to randomly generate one thousand sets of values of these parameters [24]. We then ran the model a thousand times with these sets of parameters and calibrated each run against the number of PLWHA, reported by UNAIDS. Model fit was tested using the modeling efficiency statistic (EF) (Equation C-1, Supplemental Digital Content 1,, which is conceptually similar in meaning to the R-squared statistic from linear regression and indicates better fit for larger values of EF. We kept the 200 runs with the largest EF values and calculated sensitivity ranges for the percentage of HIV infections averted over 30 years and average cost-effectiveness ratios (CERs) relative to the base case using these 200 sets of parameters (Supplemental Digital Content 1,

Scenario-based sensitivity analysis

We also tested the robustness of our primary findings to our assumptions about the effectiveness of the different interventions and infectiousness of HIV during the asymptomatic stage using scenario-based sensitivity analysis. In this analysis, we calculated the cost-effectiveness and HIV infections averted under different values for the key parameters that describe the intervention outcomes (Supplemental Digital Content 1,


HIV epidemic projections

Without any additional intervention, the total number of PLWHAs will climb to over 1.5 million in 2040, and HIV prevalence in all groups will increase during the next 30 years, from an overall prevalence of 0.08% in year 1 to 0.14% by year 30. HIV prevalence for IDUs (men and women) and MSM will increase gradually to 25.0 and 14.0%, respectively, after 30 years (Fig. 1). Furthermore, the cumulative number of infections over three decades will total 3.41 million, of which 75% are in high-risk groups, especially MSM and IDU. Without any change in current government policies, it, therefore, appears likely that although the magnitude of the epidemic will grow it will remain concentrated in high-risk groups.

Fig. 1:
Prevalence by risk group under the base case.FSW, female sex workers; IDU, injecting drug user; MSM, men who have sex with men.

Cost-effectiveness analysis

The costs and health benefits (QALYs) for all strategies are shown in Table 2 and Fig. 2. The sensitivity range of the average CER for each intervention is also shown in Table 2. All average CERs in Table 2 are relative to the base case, and the incremental CERs (ICERs) in Fig. 2 are compared with the next best alternative in order to construct a cost-effectiveness frontier. Among the expanded VCT-only strategies, the most cost-effective strategy was one-time low-risk and annual high-risk VCT, which could prevent 6.6% (0.23 million) of the total new infections projected under the base case, with an average CER of 5810 Int.$/QALY. On the contrary, the ART-only strategy (ART utilization 50%) yields an average CER of 4840 Int.$/QALY and prevents 0.34 million new infections, 10.0% of the total new infections. The harm reduction program strategy can prevent 0.71 million infections, 21% of the total new infections, at a cost of 5090 Int.$ per QALY gained (Table 2 and Fig. 2).

Fig. 2:
Cost-effectiveness frontier for all strategies.ART, antiretroviral treatment; VCT, voluntary counseling and testing.

Compared to the base case, the combination strategy of any two of VCT (low-risk once and high-risk annually), ART (50% utilization), and harm reduction could prevent 0.56–1.01 million new infections, 16.5–29.3% of the total new infections, yielding an average CER of 5030–5310 Int.$/QALY. Under a combination of all three strategies, 1.2 million new infections could be prevented at a cost of 5130 Int.$ per QALY gained (Table 2 and Fig. 2). When resources are available, the optimal cost-effectiveness path is from ART to the combination strategy of ART and harm reduction, with an ICER of 5260 Int.$ per QALY gained, followed by a shift to the combination strategy of harm reduction, ART, and VCT, costing 5550 Int.$ per QALY gained (Fig. 2). Subsequent steps on the cost-effectiveness frontier represent increased frequency of VCT, but have very high incremental cost-effectiveness values.

The saving in healthcare cost for all strategies compared to the base case is also shown in Table 2. Screening low-risk groups once and high-risk groups annually was also the most effective VCT strategy in terms of healthcare cost (Table 2 and Fig. 2). With this VCT strategy, the saving in healthcare cost is over twice the incremental cost. Even though the harm reduction program strategy is not as cost-effective as the expanded ART-only strategy in terms of average CER, the savings in healthcare cost covers all of the incremental cost and it can prevent 0.3 million more infections (Table 2 and Fig. 2).

Sensitivity analysis

Sensitivity analysis on the cost-effectiveness

Figure 3 shows the incremental costs and QALYs of the different interventions relative to the base case, charted on the cost-effectiveness plane. Because all sets of values for all interventions lie entirely in one quadrant, it is possible to calculate meaningful sensitivity ranges for the ICER average CER (Table 2). The sensitivity ranges of percentage of HIV infections averted over 30 years are also shown in Table 2, along with the point estimates from the primary model. Expanded VCT-only (low-risk once and high-risk annually) would prevent 2.8–11.2% of the total new HIV infections, whereas the harm reduction program strategy would prevent 3.2–33.6% of the total new infections. Under a combination strategy of expanded VCT (low-risk once and high-risk annually), ART, and harm reduction programs, 21.1–42.5% of the total new infections over 30 years would be averted.

Fig. 3:
Sensitivity analysis for average cost-effectiveness ratio of selected interventions relative to the base case.QALY, quality-adjusted life year; VCT, voluntary counseling and testing.

All selected strategies, including expanded VCT-only (low-risk once and high-risk annually), expanded ART-only, harm reduction program, and the combination strategy of these three remain cost-effective relative to the base case under all ranges of sensitivity values (Table 2). Note that the harm reduction program shows a wide range of possible values for the average CER, possibly because it focuses on IDU, whose prevalence and incidence trends are highly sensitive to a single value (number of injections).

Scenario-based sensitivity analysis

The interventions remained cost-effective under a wide range of scenarios, in which the effectiveness of screening and/or treatment was varied (See detail in Supplemental Digital Content 1,


This study employed a compartmental mathematical model to estimate trends in the HIV epidemic in China and to evaluate the cost-effectiveness of four intervention programs. We estimated the course of the HIV epidemic for the next three decades without any further interventions and found that nationwide prevalence will keep growing over this period. Through expanded VCT, however, a certain fraction of new HIV infections can be prevented and millions of QALYs can be gained. Combining VCT and ART can prevent between 8 and 24% of HIV infections, making the combination strategy more effective than VCT alone. Behavioral change through harm reduction programs is also shown to be a cost-effective tool among IDUs. Furthermore, a combination of harm reduction programs and biomedical strategies is also cost-effective.

This is the first study to consider the cost-effectiveness of treatment, testing, harm reduction, and combinations of these strategies at a national level in China. Although some studies report the cost-effectiveness of some of these interventions, research to date has only considered the strategies in isolation and only in certain provinces or cities [25–30].

This study confirms findings from other studies [31,32] that biomedical interventions, such as VCT and treatment, can have a significant impact on HIV and are cost-effective under certain implementation strategies. However, it does not support their finding of HIV elimination through testing and treatment strategies alone, so is more consistent with recent criticism that the assumptions and expectations underlying these studies are too optimistic [33]. In addition, this study gives detailed evidence to support claims that combinations of interventions are perhaps more effective in connection with IDUs [34].

This study also confirms other findings about the cost-effectiveness and effectiveness of harm reduction programs [11,35,36] and extends these findings to a middle-income country in a direct comparison with testing and treatment. Although not modeled in this study, subsidiary social, public health, and public order benefits from MMT, such as IDU in MMT returning to employment and reduced overdose death rates [37], could lead to further cost-effectiveness gains from combinations of NSP and pharmacotherapy treatment. This study also suggests that the cost-effectiveness of harm reduction is preserved in conjunction with biomedical interventions.

We also calibrated our model against the estimated number of PLWHA (PLWHA), annual deaths due to HIV/AIDS, and annual new infections in China, and all of the predictions of these indicators based on our model were within the range of the values published by UNAIDS. However, there are several limitations in this model, which should be considered in interpreting our findings. First, we assumed random mixing among sexual and needle-sharing contacts, even though people from high-risk groups may have a higher probability of contact with high-risk people than with low-risk individuals. Second, acute HIV testing, ART adherence, resistance to ART, false-positive results and drug toxicities were not discussed in the model. Moreover, we only considered epidemiological variations between risk and sex groups and omitted differences in the epidemic between regions and provinces or the effects of migration. These simplifying assumptions may lead to underestimation of the number of new infections, but many of these assumptions are impossible to avoid given the difficulty of estimating parameters to describe these effects. Third, we also ignored mother-to-child transmission, which is relatively small, representing an estimated 1.1% of total HIV incidence in China in 2007 [4]. Fourth, we did not account for the change in MMT retention rates, which could have resulted from the expansion of the MMT programs: we used a constant retention rate of 50% [22,23] for the entire intervention period. However, in most cases, the retention rate could be expected to decrease as the program is scaled up, due to a lowering of eligibility requirements. This would result in a reduction of effectiveness of MMT as the program expanded, with a consequent reduction in cost-effectiveness. Finally, we did not allow maturation from IDU, although it is well known that IDU can achieve permanent abstinence through drug rehabilitation programs [37].

An additional important consideration in this research is the potential difficulty in identifying high-risk groups, such as MSM and IDU, and in obtaining their cooperation in population health programs such as those that were tested in our models. These groups are very hard to reach, and this may result in an additional substantial cost for the VCT strategies, NSP, and MMT. Besides significant funding increases, legal changes may also be needed to make these strategies work, as the strategies may still be ineffective if there is strong distrust of health providers and/or law enforcement agencies among IDU, MSM, or CSW. It is likely that an expanded VCT program could only be effective if accompanied by significant legal and social changes in China's approach toward people from marginalized groups, just as happened in some developed nations in the 1980 s. This suggests that it would be difficult to implement an expanded VCT program without simultaneously implementing expanded harm reduction programs and making legal changes. Thus, estimates of cost-effectiveness and total costs of the Expanded VCT-Only program may well be too optimistic in this article, or the program may be impossible to implement in isolation.

Our model also shows that with the harm-reduction strategy alone, HIV among IDUs cannot be eliminated within 30 years in China: HIV prevalence among IDUs will decline and new annual infections among this risk group will decrease gradually. However, as opposed to the experience of some developed nations, such as Australia, this result suggests that harm reduction interventions may not be effective in isolation in China, as they depend on health promotion activities among hard-to-reach groups [34] in a more challenging legal and human rights environment.

This study shows that it may be possible to significantly control HIV growth among high-risk populations in China through a combination of VCT, treatment, and harm reduction programs. These results raise the possibility that China will face a future largely free of this disease if it acts now to implement a widespread expansion of both biomedical and harm reduction strategies.


J.L. and S.G. are responsible for the design, analysis and drafting of this study. H.Z., J.L. and S.G. conducted the coding and analysis of the mathematical model. A.K. and K.S. provided scientific discussion and comments.

This work was supported by Research on HIV/AIDS in Health and Labour Sciences Research Grants from the Ministry of Health, Labour and Welfare, Japan (No.08150755).

Conflicts of interest

There are no conflicts of interest.


1. Ministry of Health of China. China 2010 UNGASS Country Progress Report (2008–2009). Beijing, China: Ministry of Health of China; 2010.
2. National Bureau of Statistics of China. 2010 China Health Statistical Yearbook. Beijing, China: National Bureau of Statistics of China; 2010.
3. UNAIDS. 2008 Report on the global aids epidemic. Geneva, Switzerland: UNAIDS; 2008.
4. Wang L, Wang N, Li D, Jia M, Gao X, Qu S, et al. The 2007 estimates for people at risk for and living with HIV in China: progress and challenges. J Acquir Immune Defic Syndr 2009; 50:414–418.
5. Ma X, Zhang Q, He X, Sun W, Yue H, Chen S, et al. Trends in prevalence of HIV, syphilis, hepatitis C, hepatitis B, and sexual risk behavior among men who have sex with men. Results of 3 consecutive respondent-driven sampling surveys in Beijing, 2004 through 2006. J Acquir Immune Defic Syndr 2007; 45:581–587.
6. Cao X, Yin W, Pang L, Zhang C, Xu J, Xiao Y, et al. Risk factors which were associated with heroin use during the methadone maintenance treatment among 1301 patients in 9 cities of China. Chin J Epidemiol 2010; 31:269–272.
7. Lu F, Wang N, Wu Z, Sun X, Rehnstrom J, Poundstone K, et al. Estimating the number of people at risk for and living with HIV in China in 2005: methods and results. Sex Transm Infect 2006; 82 (Suppl 3):iii87–iii91.
8. Guo H, Wei JF, Yang H, Huan X, Tsui SK, Zhang C. Rapidly increasing prevalence of HIV and syphilis and HIV-1 subtype characterization among men who have sex with men in Jiangsu, China. Sex Transm Dis 2009; 36:120–125.
9. Ruan S, Yang H, Zhu Y, Wang M, Ma Y, Zhao J, et al. Rising HIV prevalence among married and unmarried among men who have sex with men: Jinan, China. AIDS Behav 2009; 13:671–676.
10. Feng L, Ding X, Lu R, Liu J, Sy A, Ouyang L, et al. High HIV prevalence detected in 2006 and 2007 among men who have sex with men in China's largest municipality: an alarming epidemic in Chongqing, China. J Acquir Immune Defic Syndr 2009; 52:79–85.
11. NCHECR. Return on investment 2: evaluating the cost effectiveness of needle syringe programs in Australia. Darlinghurst, NSW, Australia: NCHECR, Department of Health and Ageing; 2009.
12. Coates TJ, Richter L, Caceres C. Behavioural strategies to reduce HIV transmission: how to make them work better. Lancet 2008; 372:669–684.
13. McClelland RS, Graham SM, Richardson BA, Peshu N, Masese LN, Wanje GH, et al. Treatment with antiretroviral therapy is not associated with increased sexual risk behavior in Kenyan female sex workers. AIDS 2010; 24:891–897.
14. Porco TC, Martin JN, Page-Shafer KA, Cheng A, Charlebois E, Grant RM, et al. Decline in HIV infectivity following the introduction of highly active antiretroviral therapy. AIDS 2004; 18:81–88.
15. Cleary PD, Van Devanter N, Rogers TF, Singer E, Shipton-Levy R, Steilen M, et al. Behavior changes after notification of HIV infection. Am J Public Health 1991; 81:1586–1590.
16. Higgins DL, Galavotti C, O’Reilly KR, Schnell DJ, Moore M, Rugg DL, et al. Evidence for the effects of HIV antibody counseling and testing on risk behaviors. JAMA 1991; 266:2419–2429.
17. Castilla J, Del Romero J, Hernando V, Marincovich B, Garcia S, Rodriguez C. Effectiveness of highly active antiretroviral therapy in reducing heterosexual transmission of HIV. J Acquir Immune Defic Syndr 2005; 40:96–101.
18. UNAIDS. 2006 Report on the global AIDS epidemic. Geneva, Switzerland: UNAIDS; 2006.
19. Long EF, Brandeau ML, Owens DK. The cost-effectiveness and population outcomes of expanded HIV screening and antiretroviral treatment in the United States. Ann Intern Med 2010; 153:778–789.
20. WHO. Making choices in health: WHO guide to cost-effectiveness analysis. Geneva, Switzerland: WHO; 2003.
21. WHO Commission on Macroeconomics and Health. Macroeconomics and health: investing in health for economic development. Geneva, Switzerland: WHO; 2001.
22. Li X, Tan H, Sun Z, Zhang H, Chen M, Ou Q. Study on the time of retention and related infuluencing factors of patients receiving methadone maintenance treatment in Hunan province. Chin J Epidemiol 2009; 30:672–675.
23. Luo W, Pang L, Wu Z, Mi G, Wang C, Li J. Enrollment and retention rate of patients in the first batch of methadone maintenance clinics in China. Chin Ment Health J 2007; 21:478–480.
24. Vickerman P, Watts C. The impact of an HIV prevention intervention for injecting drug users in Svetlogorsk, Belarus: model predictions. Int J Drug Policy 2002; 13:149–164.
25. Cheng G, Qian Z, Hu J. Longitudinal analysis of technical efficiency of voluntary counseling and testing of HIV in China. J Peking Univ (Health Sci) 2009; 41:135–140.
26. Guo Z. The direct medical costs analysis of national free antiretroviral treatment for HIV/AIDS patients. Beijing, China: 2008:69.
27. He Q, Yuan J, Xu Y, Lin P. The projection of HIV/AIDS medical expenses in Guangdong province. China J AIDS/STD 2004; 10:271–274.
28. Xi J. Comparison of working mechanism between methadone maintenance treatment and needle and syringe exchange in Yunnan Province. Beijing, China: China CDC; 2010.
29. Yang H, Li J, Wu Z, Xu L, Wang K. Study on the utilization of health services and costs of hospital-based medical care for 29 patients with HIV/AIDS in China. Chin J Epidemiol 2003; 24:393–396.
30. Zhou F, Li Y, Wang J, Guo Z, xu Y, Yang H, et al. An analysis of direct medical expenditures ralated with free ARV treatment for AIDS patients in part of Yunnan province. Chin J AIDS STD 2008; 14:450–453.
31. Granich RM, Gilks CF, Dye C, De Cock KM, Williams BG. Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model. Lancet 2009; 373:48–57.
32. Wang S, Moss JR, Hiller JE. The cost-effectiveness of HIV voluntary counseling and testing in China. Asia Pac J Public Health 2011; 23:620–633.
33. Kretzschmar ME, van der Loeff MF, Coutinho RA. Elimination of HIV by test and treat: a phantom of wishful thinking?. AIDS 2012; 26:247–248.
34. Degenhardt L, Mathers B, Vickerman P, Rhodes T, Latkin C, Hickman M. Prevention of HIV infection for people who inject drugs: why individual, structural, and combination approaches are needed. Lancet 2010; 376:285–301.
35. Vickerman P, Miners A, Williams J. Assessing the cost-effectiveness of interventions linked to needle and syringe programmes for injecting drug users: an economic modelling report. London, UK: NICE; 2008.
36. Zaric GS, Barnett PG, Brandeau ML. HIV transmission and the cost-effectiveness of methadone maintenance. Am J Public Health 2000; 90:1100–1111.
37. Ward J, Mattick RP, Hall W. Methadone maintenance treatment and other opioid replacement therapies. Sydney, Australia: Harwood Academic Publishers; 1998.

China; cost effectiveness analysis; drug users; HIV/AIDS; mathematical models

Supplemental Digital Content

© 2012 Lippincott Williams & Wilkins, Inc.