JAIDS Journal of Acquired Immune Deficiency Syndromes:
Epidemiology and Social Science
Comparing the Cost-Effectiveness of HIV Prevention Interventions
Cohen, Deborah A MD, MPH*; Wu, Shin-Yi PHD*; Farley, Thomas A MD, MPH†
From the *RAND Corporation, Santa Monica, California; and †Tulane University School of Public Health and Tropical Medicine, New Orleans, LA.
Received for publication May 23, 2003;
accepted February 6, 2004.
Supported in part by The Center for HIV Identification, Prevention, and Treatment Services (CHIPTS) and the Centers for Diseases Control, PA# 01158; #R18/CCR920939-01.
This work is the sole responsibility of the authors.
Reprints: Deborah A. Cohen, RAND Corporation, 1700 Main Street, Santa Monica, CA 90405 (e-mail: email@example.com).
Objective: Communities need to identify cost-effective interventions for HIV prevention to optimize limited resources.
Methods: The authors developed a spreadsheet tool using Bernoulli and proportionate change models to estimate the relative cost-effectiveness for 26 HIV prevention interventions including biomedical interventions, structural interventions, and interventions designed to change risk behaviors of individuals. They also conducted sensitivity analyses to assess patterns of the cost-effectiveness across different populations using various assumptions.
Results: The 2 factors most strongly determining the cost-effectiveness of the different interventions were the HIV prevalence of the population at risk and the cost per person reached. In low-prevalence populations (eg, heterosexuals) the most cost-effective interventions were structural interventions (eg, mass media, condom distribution), whereas in high-prevalence populations (eg, men who have sex with men) individually focused interventions to change risk behavior were also relatively cost-effective. Among the most cost-effective interventions overall were showing videos in STD clinics and raising alcohol taxes. School-based HIV prevention programs appeared to be the least cost-effective. Needle exchange and needle deregulation programs were relatively cost-effective only when injection drug users have a high HIV prevalence.
Conclusions: Comparing estimates of the cost-effectiveness of HIV interventions provides insight that can help local communities maximize the impact of their HIV prevention resources.
The Institute of Medicine has made a strong case that HIV prevention efforts should maximize the number of HIV infections prevented.1 As more interventions are developed, state and local health departments and community planning groups must choose among the available models to develop a comprehensive and effective plan to meet the needs of their unique communities. Therefore, tools are needed to help community planning groups prioritize HIV prevention interventions and allocate limited resources. Such tools should take into account the cost-effectiveness of the interventions and their applicability to the characteristics of their local epidemic.
Cost-effectiveness analyses are widely used to determine how to allocate health resources efficiently. They allow different types of interventions to be evaluated on a common metric, with the aim of maximizing health benefits at a given monetary cost. Published cost-effectiveness analyses of HIV prevention interventions express benefits in terms of the estimated number of HIV infections prevented. It is impossible to measure the number of persons who are not infected with HIV as a result of a prevention program. Therefore, cost-effectiveness calculations in HIV prevention planning use estimates of the number of prevented HIV infections produced by various mathematic models. These mathematic models rely on a variety of assumptions, including the per-act HIV transmission probability, the frequency of risk behaviors, the population HIV prevalence, and many other parameters that are often not measured in intervention trials. These assumptions make the validity of the absolute number of infections prevented suspect. However, when the same mathematic models and the same assumptions are used across interventions, the relative value of the costs of different interventions per infection prevented is substantially valid; thus, cost-effectiveness analyses can be extremely helpful in choosing among a variety of effective intervention programs.
HIV is a serious illness with high morbidity, mortality, and treatment costs. The most recent estimates of the lifetime costs of treating HIV-related illness take into account new generations of antiretroviral drugs, combination drug therapies, and improved techniques for monitoring disease progression and therapeutic effectiveness. The average lifetime cost of HIV care is estimated at $195,188 (1996 dollars).2 This figure has been discounted at 3%, as recommended by the U.S. Panel on Cost-Effectiveness in Health and Medicine, to account for the fact that prevention benefits and their associated cost savings occur primarily in the future.2 Because the cost of treating HIV is so high, any intervention that costs less than this amount per HIV infection prevented may be considered cost-effective.
We developed a tool for community groups to compare the relative cost-effectiveness of different HIV prevention interventions in their local populations. We then conducted sensitivity analyses to assess the general patterns of cost-effectiveness of different types of interventions, and evaluated which of the interventions were likely to be cost-effective (compared with HIV treatment) in different populations.
Interventions included in the tool were those that have evidence of effectiveness in reducing HIV incidence, sexually transmitted disease (STD) incidence, or sexual or needle-sharing behavior likely to impact HIV transmission, based on prospective trials or observational studies, or are commonly used to prevent HIV infection in the United States. The set of interventions was not limited to those conducted or funded by public health agencies for the specific purpose of preventing HIV infection. It also included interventions conducted for other purposes that may have an additional effect of reducing HIV transmission (eg, drug treatment programs, male circumcision). Interventions were grouped in 4 broad categories: individual interventions that take place through a direct interaction between a health worker and an individual person,3 community and social network interventions that employ diffusion of prevention tools or messages through individuals to others not directly contacted, structural interventions that influence risk behavior by changing the physical or social environments to which persons at risk are exposed, and biomedical interventions that reduce HIV transmission through biologic mechanisms rather than changes in risk behavior. Specific interventions included are listed in Table 1.
For each intervention, we selected 1 study from the scientific literature demonstrating its effectiveness in changing HIV incidence, STD incidence, or risk behavior (unprotected sex or needle sharing). For HIV counseling and testing we reviewed 2 models: a standard model assessed with a meta-analysis4 with respect to changing sexual behavior and a “client-centered” model used in the “Project Respect” trial,5 which used declines in STD incidence as a measure of effectiveness. We included several community mobilization/street outreach interventions because of substantially different program implementation methods.
General Approach to Cost-effectiveness Estimation
In the prevention of HIV transmission, the effectiveness that we estimated is the number of HIV infections prevented, and the cost is the program cost of implementing a particular intervention. The cost-effectiveness ratio is
Equation (Uncited)Image Tools
In the tool we developed, the cost-effectiveness is assessed from the perspective of the public health system. Thus the program cost includes all resources (purchased, donated, or volunteered) used to implement the intervention, but excludes any cost incurred by the participants, unless they are reimbursed. The total number of HIV infections prevented includes those directly prevented by the intervention (primary infections) and an estimate of the number of infections prevented in sex partners (secondary infections). The secondary infections prevented were estimated by considering the prevalence of HIV in the sex partner pool, multiplied by the number of sex partners and the risk of sexual transmission.6
Estimates of HIV Infections Prevented
The estimate of the number of primary infections prevented is based on subtracting an estimate of the number of HIV infections that would have happened if the prevention program had not been in place from an estimate of the number of HIV infections that would have happened even with the program in place. The most common mathematic model used and the model we employed whenever possible is called the Bernoulli process model.7,8 In the Bernoulli model, each sex act is treated as an independent event with a small, fixed probability that HIV is transmitted between members of a couple who are discordant in their HIV status. From this per-act probability, the model then estimates the cumulative probability that an uninfected individual with given sexual behaviors (number of partners, frequency of sex acts) would become infected during a specified time period. The number of new HIV cases is determined by the size of the population with given behaviors, the estimated number of discordant partnerships, and the cumulative probability of transmission within these partnerships. Parameters measuring the effectiveness of the interventions, such as changes in condom use or number of sex partners, were drawn from the selected studies. In the spreadsheet tool, these parameters serve as default values that local groups can change if they have reason to believe their intervention effectiveness differs from that of the interventions studied. Not all the interventions collected data that are needed to calculate cost-effectiveness, such as the frequency of sex or the number of sex partners. When data were not available from the studies directly, values from national sources were used as default values for sexual behavior and HIV transmission probabilities (Table 29–17). HIV prevalence estimates were taken directly from the study or, when not available, were taken from published estimates based on the location in which the study was conducted.18
In assessing the effectiveness of HIV partner notification and HIV counseling and testing, we calculated separately the effect on partners of HIV-positive persons and HIV-negative persons and then combined the two for an overall effect.
For assessing interventions that prevent HIV transmission through needle sharing, a similar Bernoulli process formula was adapted from that used by Weinstein et al.7 This formula calculates HIV infections prevented based on HIV prevalence in drug users, the number of injections, and frequency of needle sharing.
Other interventions may reduce the probability of HIV transmission by improving the diagnosis and treatment of STDs that facilitate HIV transmission.19 For interventions involving STD screening and treatment, we used a model developed by Chesson and Pinkerton,20 which is an extension of the Bernoulli model described earlier, to estimate the number of HIV infections that would be prevented by this STD treatment.
Additional interventions were evaluated in the selected scientific study through their reduction in the incidence of other STDs, (particularly gonorrhea) rather than their effect on self-reported sexual risk behavior. To estimate the number of HIV infections prevented by these interventions, we assumed that the reductions in HIV incidence were proportional to the reductions in gonorrhea incidence (a “proportionate change” approach).
Estimates of Costs
Program costs were considered as the total cost to the public health system to implement the intervention. These were either assessed as the cost to intervene with an individual person at risk or the cost for the entire community divided by the number of persons targeted by the intervention. In either case, the final parameter used was the program cost per person reached. For some interventions, these costs were available from published cost analyses or cost–benefit analyses.14,21–29 For other interventions, the costs included as defaults in the spreadsheet tool (our “base-case estimates”) were estimates of the cost of person-hours, supplies, and overhead needed to implement the intervention. For this we used salary and subcontract figures supplied by HIV prevention staff at the Louisiana Office of Public Health and/or the Los Angeles County Department of Health. The spreadsheet tool requires users in local communities to replace these estimates with their local implementation costs. Because of the uncertainty of the default estimates, we conducted sensitivity analyses of the cost parameter in this study.
Two interventions that are likely to reduce HIV transmission are policy changes that do not have any program costs: the raising of taxes on alcohol and needle/syringe deregulation (ie, allowing the nonprescription sale of sterile needles and syringes). For these interventions we estimated the cost of lobbying to enact these regulatory changes. This one-time cost does not continue once the policy is in place.
Comparisons of Cost-effectiveness and Sensitivity Analyses
We first calculated the cost-effectiveness of each intervention using the population in the selected published study. Thereafter, to have some comparability across the interventions, we standardized the duration of effect to 1 year. In these cases we assumed that the effect found at the study end point (if it were less than 1 year) would be sustained for 1 year. If the effect was only measured at a follow-up time greater than 12 months, we interpolated the benefit in a linear fashion to estimate the effect after 12 months.
To determine the robustness of the estimates, we then conducted 1-way and 2-way sensitivity analyses and systematically varied each parameter in the equation. We calculated the cost-effectiveness of interventions designed for high-prevalence populations (men who have sex with men or injection drug users in high-prevalence cities) across an HIV prevalence range of 0.01 to 0.20, and for interventions designed for general or lower prevalence populations (high-risk heterosexuals, adolescents at risk), across an HIV prevalence range from 0.001 to 0.01. We then varied the cost per person reached by each intervention from 50% of our base-case estimate to twice the base-case estimate. Finally, we altered the effectiveness of the intervention by halving or increasing by 50% the change in condom use, the number of sex partners, the frequency of sex and needle sharing, and the change in STD incidence. We viewed separately the four interventions in which cost-effectiveness was evaluated using the proportionate change model, because it is not completely comparable with the Bernoulli model.
Table 330–55 summarizes the interventions included in the tool, the measures of effectiveness from the published studies, and the time period after which the intervention effects were measured. Table 4 summarizes the effectiveness (in number of HIV infections prevented) and the cost-effectiveness of each intervention based on the published study, supplemented when necessary by standard parameters in Table 2 and/or other parameters as described in the methods section. In the last column, the cost-effectiveness ratio is adjusted from the time period of the published study to an effect after 12 months. The HIV prevalence of the populations studied or similar populations and estimates of the HIV prevalence in their partners is also included. Because of the different populations, the interventions target—and in particular, the different HIV prevalence values in these populations—this table should not be viewed as a fair and balanced comparison of the relative cost-effectiveness of these interventions.
Of all the parameters that we systematically studied, the 2 that had the greatest impact on cost-effectiveness were the prevalence of HIV infection in the target population and the cost per person reached by the intervention. Figures 1 to 3 provide relevant comparisons of the cost-effectiveness among interventions across a range of HIV prevalence. Interventions for which the cost-effectiveness ratio remains below the $200,000 threshold may be considered cost-effective (compared with the lifetime costs of treatment of HIV infection) in any population. Assuming an effect at 12 months, the most cost-effective interventions for the general population are mass media campaigns and condom availability programs. Among targeted interventions for groups with high HIV prevalence, the most cost-effective interventions are opinion leader programs and community mobilization using the Mpowerment model.42 Needle exchange and needle deregulation were cost-effective only where HIV prevalence was very high. Among the interventions evaluated using the proportionate change model, when we varied HIV incidence, both alcohol tax increases and showing videos in STD clinics remain cost-effective,30,53 whereas group counseling and client-centered HIV counseling and testing are not cost-effective for low-incidence populations.5,33 The cost-effectiveness of the client-centered counseling approach is superior to the standard approach across a broad range of scenarios. All 3 interventions designed to prevent HIV transmission from a known HIV-positive person (discordant couple counseling, HIV partner notification, and STD screening/treatment in HIV clinics) are highly cost-effective compared with treatment (Fig. 4D).
Programs that are very costly per HIV infection prevented, and likely not cost-effective regardless of the changes in any parameters, include school-based education or any other programs targeted at youth, primarily because the HIV prevalence in this group is very low. Drug treatment programs are less cost-effective for HIV prevention primarily because their cost per person reached is so high compared with many other interventions. Similarly HIV antiretroviral treatment is not cost-effective as a measure to prevent secondary HIV transmission because of the high cost per person reached.
Figure 4 shows how the cost-effectiveness ratios may change when the cost of the program is varied at a fixed HIV prevalence. Again, mass media programs and condom availability programs remain relatively cost-effective for general populations, whereas the Wendell model of street outreach40 is no longer as cost-effective when the cost of the program doubles (Fig. 4B). Even at half of the base-case cost, other methods of street outreach/general community mobilization and youth supervision programs are not relatively cost-effective. In comparison, all 3 interventions targeted at HIV-positive persons remain cost-effective even when the base-case cost per person reached is doubled (Fig. 4D).
The cost-effectiveness of varying other parameters, such as halving or increasing by 50% program effectiveness, number of sex partners, and frequency of sex, were within the same range of the results shown for the potential variations in HIV prevalence and cost (not shown), and have a similar pattern and ranking to the figures shown.
Several principles can be gleaned from the analysis of the relative cost-effectiveness of these 26 HIV prevention interventions. First, the factors that most strongly determine the cost-effectiveness of programs that are known to be effective are the HIV prevalence in the target population and the cost per person reached by the intervention. Second, programs that target populations that already have a high prevalence of HIV can be cost-effective even when the implementation cost is several hundred dollars per person reached. Third, interventions to reach low-prevalence populations can also be cost-effective compared with treatment, but they need to have an extremely low price per person reached (eg, less than $10). Effective interventions that are essentially cost free, such as alcohol tax increases or showing already prepared videos to patients who are waiting in a clinic, are always cost-effective and should be implemented because of the potential benefit. Finally, it is not always necessary to target directly or produce behavior change to reduce HIV transmission; biomedical interventions can be cost-effective options.
Some of the interventions that our analysis found to be the most cost-effective are not even considered by health agencies in the HIV prevention plans. Nonetheless, published studies argue that they are effective, and our analyses suggest that they may be a far better use of HIV prevention resources than some standard interventions. In particular, increasing the price of alcohol reduces alcohol consumption, which has been associated with risky sex. Price increases do not change attitudes or knowledge regarding the risks of sex, but simply alter the opportunity for high-risk behaviors. A review of the effect of alcohol tax increases on STD rates suggests that the effect of this intervention may be large.53 Mass media campaigns have the potential to be effective at changing behavior, as suggested by the Swiss STOP AIDS programs that specifically targeted condom use, and, if effective, are likely to be relatively cost-effective. Screening and treatment of curable STDs among HIV-positive persons are likely to prevent substantial HIV transmission to their partners because bacterial STDs are important cofactors in HIV transmission.
Limitations of Cost-effectiveness Estimations
There are many limitations to cost-effectiveness analyses such as these. First, the Bernoulli model we used assumes that people within a given population choose sex partners randomly, rather than preferentially choosing partners similar to themselves in risk. The selective assortment that takes place in actual populations would change the dynamics of HIV transmission, but the degree to which it would change is unknown. We were unable to use a single mathematic model for all our cost-effectiveness calculations, so comparisons across interventions using different models may be inaccurate.
In addition, the duration of intervention effectiveness is unknown, so we cannot be sure whether all our results actually do extend to 12 months or whether the effectiveness is delayed and only appears after several years (as might be the case in school-based interventions). When interventions are evaluated, their effect is usually measured after a single time interval. It may not be appropriate to compare interventions when effectiveness is measured after different time periods. We also assumed an immediate and constant effect, which may not be the case in practice. On the other hand, if the effects of interventions last beyond the point at which we measured the outcome, our methods will underestimate the impact.
Another concern is the strength of evidence of the effectiveness of the interventions we used to conduct the analysis. Some studies were conducted using a rigorous, randomized, controlled design, whereas others were natural experiments or were not randomized. Moreover, even when the strength of the evidence is high for a particular intervention, the effectiveness in changing behavior when implemented in local communities may vary substantially from the effectiveness under research conditions.
Nevertheless, despite these limitations, our methods do provide a means to understand the general patterns of relative cost-effectiveness of different interventions, and we found that this cost-effectiveness varied between interventions by several orders of magnitude. In current practice, it is virtually impossible to use cost-effectiveness in choosing among different HIV prevention interventions, so funding and other resources for HIV prevention are likely not used to maximize the number of HIV infections prevented. At the same time, although cost-effectiveness comparisons such as these are extremely useful, they should not be the sole decision-making criteria in determining the allocation of HIV prevention resources. Other critical factors to be considered include the intervention’s strength of evidence for effectiveness, feasibility, acceptability, and replicability in the local area. Although not shown, our spreadsheet tool includes a linked application that allows these and other factors to be considered and weighted along with cost-effectiveness in prioritizing interventions. Given that current decision making for HIV prevention usually does not usually consider cost-effectiveness, this tool could substantially improve community decisions regarding the portfolio of HIV prevention interventions and thereby optimize the use of limited HIV prevention resources.
1. IOM. No Time to Lose. Getting More from HIV Prevention
. Washington, DC: National Academy Press; 2000.
2. Holtgrave DR, Pinkerton SD. Updates of cost of illness and quality of life estimates for use in economic evaluations of HIV prevention programs. J Acquir Immune Defic Syndr Hum Retrovirol
3. Cohen D, Scribner R. An STD/HIV prevention framework. Aids Patient Care STDS
4. Weinhardt LS, Carey MP, Johnson BT, et al. Effects of HIV counseling and testing on sexual risk behavior: a meta-analytic review of published research, 1985–1997. Am J Public Health
5. Kamb ML, Fishbein M, Douglas JM, et al. Efficacy of risk-reduction counseling to prevent human immunodeficiency virus and sexually transmitted diseases. JAMA
6. Bedimo AL, Pinkerton SD, Cohen DA, et al. Condom distribution: a cost-utility analysis. Int J STD AIDS
7. Weinstein M, Graham J, Siegel J, et al. Cost-effectiveness analysis of AIDS prevention programs: concepts, complications, and illustration. In: Turner C, Miller H, Moses L, eds. AIDS: Sexual Behavior and Intravenous Drug Use
. Washington, DC: National Academy Press; 1989:471–499.
8. Pinkerton SD, Holtgrave DR, Leviton LC, et al. Model-based evaluation of HIV prevention interventions. Eval Rev
9. Mastro TD, de Vincenzi I. Probabilities of sexual HIV-1 transmission. AIDS
. 1996;10(Suppl A):S75–S85.
10. Mastro TD, Kitayaporn D. HIV type 1 transmission probabilities: estimates from epidemiological studies. AIDS R Hum Retrovirus
. 1998;14(Suppl 3):S223–S227.
11. Vittinghoff E, Douglas J, Judson F, et al. Per-contact risk of human immunodeficiency virus transmission between male sexual partners. Am J Epidemiol
12. Kaplan EH, Heimer R. A model-based estimate of HIV infectivity via needle sharing. J Acquir Immune Defic Syndr
13. Kaplan EH, Heimer R. A model-based estimate of HIV infectivity via needle sharing. J Acquir Immune Defic Syndr
14. Pinkerton SD, Holtgrave DR, DiFranceisco W, et al. Cost-threshold analyses of the national AIDS demonstration research HIV prevention interventions. AIDS
15. Allard R. A mathematical model to describe the risk of infection from sharing injection equipment. J Acquir Immune Defic Syndr
16. Laumann E, Gagnon J, Michael R, et al. The Social Organization of Sexuality: Sexual Practices in the United States
. Chicago: University of Chicago Press; 1994.
17. Kelly JA, Murphy DA, Sikkema KJ, et al. Randomised, controlled, community-level HIV-prevention intervention for sexual-risk behaviour among homosexual men in US cities. Community HIV Prevention Research Collaborative. Lancet
18. Holmberg SD. The estimated prevalence and incidence of HIV in 96 large US metropolitan areas. Am J Public Health
19. Anonymous. HIV prevention through early detection and treatment of other sexually transmitted diseases—United States Recommendations of the Advisory Committee for HIV and STD Prevention. MMWR Morbid Mortal Wk Rep
20. Chesson CW, Pinkerton SD. Sexually transmitted diseases and the increased risk for HIV transmission: implications for cost-effectiveness analyses of sexually transmitted disease prevention interventions. J Acquir Immune Defic Syndr Hum Retrovirol
21. Pinkerton SD, Holtgrave DR, DiFranceisco WJ, et al. Cost-effectiveness of a community-level HIV risk reduction intervention. Am J Public Health
22. Pinkerton SD, Holtgrave DR, Jemmott JB 3rd. Economic evaluation of HIV risk reduction intervention in African-American male adolescents. J Acquir Immune Defic Syndr
23. Pinkerton SD, Holtgrave DR, Valdiserri RO. Cost-effectiveness of HIV-prevention skills training for men who have sex with men. AIDS
24. Pinkerton SD, Johnson–Masotti AP, Holtgrave DR, et al. Using cost-effectiveness league tables to compare interventions to prevent sexual transmission of HIV. AIDS
25. Holtgrave DR, Pinkerton SD, Jones TS, et al. Cost and cost-effectiveness of increasing access to sterile syringes and needles as an HIV prevention intervention in the United States. J Acquir Immune Defic Syndr Hum Retrovirol
. 1998;18(Suppl 1):S133–S138.
26. Holtgrave DR, Valdiserri RO, Gerber AR, et al. Human immunodeficiency virus counseling, testing, referral, and partner notification services. A cost–benefit analysis. Arch Intern Med
27. Kahn JG, Kegeles SM, Hays R, et al. Cost-effectiveness of the Mpowerment Project, a community-level intervention for young gay men. J Acquir Immune Defic Syndr Hum Retrovirol
28. Wang LY, Davis M, Robin L, et al. Economic evaluation of safer choices: a school-based human immunodeficiency virus, other sexually transmitted diseases, and pregnancy prevention program. Arch Pediatr Adolesc Med
29. Varghese B, Peterman TA, Holtgrave DR. Cost-effectiveness of counseling and testing and partner notification: a decision analysis. AIDS
30. O’Donnell CR, O’Donnell L, San Doval A, et al. Reductions in STD infections subsequent to an STD clinic visit: using video-based patient education to supplement provider interactions. Sex Transm Dis
31. Kelly JA, Murphy DA, Washington CD, et al. The effects of HIV/AIDS intervention groups for high-risk women in urban clinics. Am J Public Health
32. Holtgrave DR, Kelly JA. Preventing HIV/AIDS among high-risk urban women: the cost effectiveness of a behavioral group intervention. Am J Public Health
33. Shain RN, Piper JM, Newton ER, et al. A randomized, controlled trial of a behavioral intervention to prevent sexually transmitted disease among minority women. N Engl J Med
34. Wykoff RF, Jones JL, Longshore ST, et al. Notification of the sex and needle-sharing partners of individuals with human immunodeficiency virus in rural South Carolina: 30-month experience. Sex Transm Dis
35. Toomey KE, Peterman TA, Dicker LW, et al. Human immunodeficiency virus partner notification. Cost and effectiveness data from an attempted randomized controlled trial. Sex Transm Dis
36. Coyle K, Basen–Engquist K, Kirby D, et al. Short-term impact of safer choices: a multicomponent, school-based HIV, other STD, and pregnancy prevention program. J Sch Health
37. St Lawrence JS, Brasfield TL, Jefferson KW, et al. Cognitive-behavioral intervention to reduce African American adolescents’ risk for HIV infection. J Consult Clin Psychol
38. Kelly JA, St Lawrence JS, Stevenson LY, et al. Community AIDS/HIV risk reduction: the effects of endorsements by popular people in three cities. Am J Public Health
39. Sikkema KJ, Kelly JA, Winett RA, et al. Outcomes of a randomized community-level HIV prevention intervention for women living in 18 low-income housing developments. Am J Public Health
40. Wendell DA, Cohen DA, LeSage D, et al. Street outreach for HIV prevention: effectiveness of a state-wide program. Int J STD AIDS
41. Lauby JL, Smith PJ, Stark M, et al. A community-level HIV prevention intervention for inner-city women: results of the Women and Infants Demonstration Projects. Am J Public Health
42. Kegeles SM, Hays RB, Coates TJ. The Mpowerment Project: a community-level HIV prevention intervention for young gay men. Am J Public Health
43. Prendergast ML, Podus D, Chang E. Program factors and treatment outcomes in drug dependence treatment: an examination using meta-analysis. Subst Use Misuse
44. Grosskurth H, Mosha F, Todd J, et al. Impact of improved treatment of sexually transmitted diseases on HIV infection in rural Tanzania: randomised controlled trial. Lancet
45. Farley T, Cohen D, Wu S, et al. STD screening in HIV clinics: value and implications. JAIDS
. 2003. In press.
46. Lepri AC, Miller V, Phillips AN, et al. The virological response to highly active antiretroviral therapy over the first 24 weeks of therapy according to the pre-therapy viral load and the weeks 4–8 viral load. AIDS
47. Weiss HA, Quigley MA, Hayes RJ. Male circumcision and risk of HIV infection in sub-Saharan Africa: a systematic review and meta-analysis. AIDS
48. Cohen DA, Farley TA, Bedimo–Etame JR, et al. Implementation of condom social marketing in Louisiana, 1993 to 1996. Am J Public Health
49. Heimer R, Khoshnood K, Bigg D, et al. Syringe use and reuse: effects of syringe exchange programs in four cities. J Acquir Immune Defic Syndr Hum Retrovirol
. 1998;18(Suppl 1):S37–S44.
50. Kaplan EH. Economic analysis of needle exchange. AIDS
51. Groseclose SL, Weinstein B, Jones TS, et al. Impact of increased legal access to needles and syringes on practices of injecting-drug users and police officers—Connecticut, 1992–1993. J Acquir Immune Defic Syndr Hum Retrovirol
52. Anonymous. Impact of new legislation on needle and syringe purchase and possession—Connecticut, 1992. MMWR Morbid Mortal Wk Rep
53. Anonymous. Alcohol policy and sexually transmitted disease rates—United States, 1981–1995. MMWR Morbid Mortal Wk Rep
54. O’Donnell L, Stueve A, San Doval A, et al. The effectiveness of the Reach for Health Community Youth Service Learning Program in reducing early and unprotected sex among urban middle school students. Am J Public Health
55. Dubois–Arber F, Jeannin A, Konings E, et al. Increased condom use without other major changes in sexual behavior among the general population in Switzerland. Am J Public Health
HIV prevention; cost-effectiveness; policy; community planning; economics
© 2004 Lippincott Williams & Wilkins, Inc.
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.