An estimated 25% of HIV-infected Americans are unaware of their HIV status  and recent Centers for Disease Control and Prevention (CDC) HIV prevention initiatives have emphasized case-finding and treatment [2–4]. This emphasis reflects the importance of providing effective medical care to persons infected with HIV, as well as a strategy to diminish ongoing transmission. While there is an urgent need to assure that HIV-infected men who have sex with men (MSM) are aware of their infection, traditional approaches to case-finding, particularly partner notification, may be less effective among MSM than among other high-risk groups [5–8].
New approaches are needed to identify MSM with unrecognized HIV/sexually transmitted disease (STD). In 1998, Jordan reported successfully employing patients with HIV as peer recruiters for HIV counseling and testing; 31 HIV positive patients referred 79 peers for testing and counseling, of whom 37 tested HIV positive [9,10] More recently, a preliminary evaluation of nine small, CDC-funded peer referral programs conducted by community-based organizations (CBOs) reported that 46 (5.7%) of 814 peers referred by 133 recruiters had previously undiagnosed HIV infection . We describe the case-finding effectiveness and cost-effectiveness of a public health peer referral program for HIV and STDs among MSM in King County, Washington.
Between September 2002 and February 2004, we instituted a peer referral program for HIV/STD testing as a CDC funded study. As the study demonstrated that peer referral effectively identified new cases of HIV, Public Health – Seattle & King County (PHSKC) continued peer referral as a public health program after the study ended. This paper presents data from the study period, as well as data on peer referral conducted as a public health program from March 2004 to January 2005.
Peer recruiter study population
We enrolled potential peer recruiters from the following populations of MSM: persons with bacterial sexually transmitted infections (STI) (gonorrhea, chlamydial infection, or infectious syphilis) receiving partner notification services through the PHSKC STD Program; STD Clinic patients; and Harborview Medical Center HIV Clinic patients. Staff offered all MSM receiving partner notification services participation in the study, excluding MSM with newly diagnosed HIV. STD clinic protocol during the study period was to offer all MSM patients referral to learn about ongoing studies, including the peer recruiter study; we did not collect data on the total number of MSM offered study participation. In the HIV clinic, we restricted referral for enrollment to persons our study nurse believed would be appropriate participants who had signed a registry indicating their willingness to hear about studies. Additional peer recruiters were enrolled through advertisements in local newspapers, by offering men referred as peers an opportunity to become recruiters (snowball enrollment), and via outreach to three CBOs that serve MSM.
Informed consent and peer referral training
The University of Washington Human Subjects Review Committee approved all study procedures. During the study period, interested potential participants underwent informed consent. (Informed consent was not guided by human subjects study procedures after September 2004 when the study ended and peer referral was continued as a public health program.) During both the study and the program periods, one of two public health workers who typically perform partner notification activities and HIV testing and counseling trained recruiters. The training included three modules that addressed the purpose of the project, how to approach peers safely, and how to establish a commitment to refer peers and attend a follow-up interview. While recruiters were not explicitly told not to refer their sex partners, the focus of the program was to encourage recruiters to refer their social contacts. Recruiters were instructed to not ask peers their HIV status, but were encouraged to refer MSM they believed to be at risk for HIV. We informed recruiters that they would be compensated only for referring MSM, but that compensation was not contingent on referred peers' previous or current HIV status. The training lasted approximately 40 min.
We paid peer recruiters US$ 20 when they completed their initial training and audio-computer assisted self-interview (ACASI), and US$ 20 for each person they referred who attended the clinic for HIV/STD testing. At the initial enrollment visit, the study agreed to pay each recruiter for referring up to three peers. Based on a recruiter's success enrolling MSM at risk for HIV, he was permitted to refer additional peers at the discretion of project staff. Peers referred for testing were paid US$ 10 for testing, and an additional US$ 10 for receiving their test results. Persons referred for testing who stated they already knew that they were HIV positive were tested for other STDs and were still compensated for participating in the program. Recruiters gave peers numbered cards to help the program identify which recruiter referred the peer for testing. When peers presented without a card, we identified the referring recruiter by asking the peer for their recruiter's name and physical description.
Data collection, HIV/STD testing and analysis
During the study period, all peer recruiters completed an ACASI that elicited information on the individual's demographics, HIV testing history and serostatus, substance abuse, and sexual behavior. No ACASI data were collected after the study period ended. We continued to collect demographic data on both recruiters and peers after peer referral became a public health program.
Referred peers were offered counseling and serological testing for HIV, syphilis, and hepatitis A, B and C. Starting in 2004, peers were offered rapid serum HIV tests. Staff routinely asked peers if they knew they already had HIV, and whether they were infected, immune or vaccinated for hepatitis A, B and C. We omitted tests for viral hepatitis if peers acknowledged already having these infections, and did not perform tests for hepatitis A and B on persons who reported previous vaccination. The PHSKC HIV data manager checked clinic registration information of persons testing HIV positive against a list of persons previously diagnosed with HIV through PHSKC; the list included date of birth and a non-name code identifier. We excluded persons identified as having previously tested HIV positive from the analysis of newly identified HIV. Staff also offered participants testing for rectal, pharyngeal and urethral gonorrhea, and for rectal and urethral chlamydial infection. We used Aptima Combo 2 (Gen-Probe Incorporated, San Diego, California, USA) to test urine specimens for gononrrhea and Chlamydia trachomatis. Rectal and pharyngeal infections were detected by culture. During the study period, we asked referred peers to complete the same ACASI interview as peer recruiters; peers did not need to complete the ACASI to be compensated for testing.
We used generalized estimating equations to assess the association of recruiter and peer demographic and behavioral characteristics, and chi-squared and Wilcoxon/Mann–Whitney tests to compare HIV cases identified through peer referral to those identified through other means. The SAS system (SAS Institute, Cary, North Carolina, USA) was used for all analysis.
We collected cost data for the study period only. Project personnel tracked the amount of time required to conduct the intervention for a sample of 18 enrolled peer recruiters, potential peer recruiters who declined study enrollment, and referred peers. For enrolled peer recruiters, this included the time required to initiate contact, describe the study and conduct the training, and follow-up with recruiters. For those not enrolled, time spent included time to initiate contact and describe the study. We compared per-patient time estimates with information staff recorded in time diaries tracking all of their daily activities to ensure that the study did not underestimate staff time requirements. We multiplied time estimates by representative wages to determine staff costs, and by commercial leasing rates to determine the attributable facility costs. Examination costs for referred peers included costs for staff time and testing. We defined test costs using the Centers for Medicare & Medicaid Services schedule, and treatment costs from PHSKC records and the Red Book [12–17]. Table 1 presents costs used in the analysis.
The overall program cost for the study included staff time, incentives, database management and materials used for HIV/STD testing. The primary outcome for the cost-effectiveness analysis was the cost per newly-identified HIV-infected person who received his test results. For comparison, we calculated the costs of outreach HIV testing conducted by PHSKC in local bathhouses and through a contract with a local CBO. The cost of bathhouse-based HIV testing and counseling included staff wages and benefits, part-time use of a vehicle attributed to the program in PHSKC budgets, database maintenance and data entry costs, and laboratory and administrative costs. The cost of CBO HIV testing was based on PHSKC's total contract cost for the program plus laboratory costs. We adjusted costs to 2004 dollars using the consumer price index (CPI) for all urban consumers and, for test and treatment costs, the medical care component of the CPI. Our sensitivity analysis varied staff, overhead, and testing costs by 25% above and below the baseline estimates, and HIV prevalence among tested peers by 75% above and below the rate found during the study period.
Two hundred and eighty-three men agreed to act as peer recruiters, including 186 who enrolled during the study period between September 2002 and February 2004, and 97 who became peer recruiters after the project became a public health program. The 283 recruiters referred 498 peers for HIV/STD testing, including 310 peers referred during the study period and 188 referred during the public health program. A total of 142 recruiters (51%) referred at least one peer for HIV/STD testing (median, 2; range, 1–26 among those referring one or more peers). Recruiters enrolled from the STD clinic more frequently referred one or more peers than recruiters enrolled from other sources (69 versus 32%; P < 0.0001). Success in referring one or more peers was not significantly associated with age, race/ethnicity, or participation during the research period. Among recruiters completing the ACASI, success referring peers was significantly associated with having a college education or more (62 versus 39%; P = 0.02) and not using injection drugs in the preceding 2 months (50 versus 25%; P = 0.02), but not with HIV status, use of methamphetamine, number of anal sex partners, or income.
Table 2 presents sources of enrollment, demographic characteristics, and sexual risk and substance abuse behaviors of recruiters and peers. The majority of both recruiters and peers were white, with incomes below US$ 10 000 per year. Thirty-five percent of recruiters and 51% of peers reported a history of injection drug use; 33% of recruiters and 36% of peers reported using methamphetamine in the preceding 2 months. Only 8% of peers had never HIV tested, and 57% had tested in the preceding year.
Table 3 presents the prevalence of HIV, other STDs and viral hepatitis among peers. The project identified 22 new cases of HIV among 438 tested peers and 18 (82%) of the 22 men with newly diagnosed HIV received their test results. Twelve (54%) of the 22 peers diagnosed with HIV were non-white. All recruitment sources yielded at least one peer with previously undiagnosed HIV infection, including the STD clinic (n = 8), snowball referral (n = 7), the HIV clinic (n = 4), CBOs (n = 2), and media advertisements (n = 1). Seven (8%) of the 87 men referred by recruiters who reported being HIV infected on their ACASI and five (4%) of 122 men recruited by recruiters who reported being HIV negative were newly diagnosed with HIV (P = 0.16). The success of peer referral in identifying new cases was relatively stable over time, with 12 new cases identified during the 18-month research study period, and 10 cases identified in the 11 months after peer referral became a public health program. Six peers who we did not classify as having newly diagnosed HIV tested HIV positive, four of whom acknowledged at post-test counseling that they had previously tested HIV positive. Registration information from two additional men matched records with non-name identifiers and dates of birth, suggesting that they had previously tested HIV positive. Among the 498 peers, 145 (29%) had at least one new diagnosis of an STD, including HIV, or viral hepatitis. Fifty-six (11%) had at least one newly diagnosed bacterial STD or HIV. Sixty-one men were newly diagnosed with hepatitis C.
Recruiters tended to refer peers who were like themselves in terms of demographic characteristics and substance abuse behaviors (Table 4). For example, whereas 60% of men referred by recruiters under age 30 years were also aged under 30 years, only 16% of peers referred by men aged 30 years or more were less than 30 years of age (P < 0.0001). Recruiters who reported higher levels of sexual risk were not more likely to refer peers who engaged in higher levels of sexual risk.
Between September 2002 and January 2005, PHSKC-supported HIV testing outside of the peer referral program identified 259 cases of previously unrecognized HIV infection, including 144 cases diagnosed in our STD clinic, 29 cases diagnosed through screening conducted in three local bathhouses (eight 4-h shifts weekly), and 86 cases diagnosed by CBOs supported by PHSKC or at public health clinics other than the STD clinic. In comparison with persons with HIV identified through other mechanisms, persons identified through peer referral were somewhat more likely to be nonwhite (52 versus 32%, P = 0.06) and to have never previously HIV tested (17 versus 11%; P = 0.37), and had a somewhat longer median time since last HIV test (565 versus 422 days; P = 0.60).
Table 5 presents data on the peer referral program's cost as well as comparative data on costs associated with PHSKC's bathhouse-based testing program and the county's largest publicly-supported CBO testing program. Excluding costs associated with testing for STDs other than HIV, the cost per newly diagnosed case of HIV detected and cost per HIV-infected person who received test results were lower for peer referral than for either the bathhouse or the CBO testing program. In sensitivity analyses, the total program costs shown in Table 5 varied from the baseline costs by 9% when testing and treatment costs were varied by 25%, and by 11% when counseling and data management costs were varied by 25%. Considering only HIV testing in the program, the cost per newly-detected HIV-infected peer who received test results changed 2–3% when varying testing costs, and by 14–15% when varying counseling and data management costs. The most important variable influencing cost per newly-identified HIV-infected peer receiving test results was HIV prevalence. During the study period, 4.4% (12 of 270) of peers were newly diagnosed with HIV infection. Varying the prevalence of newly-diagnosed HIV among peers 75% above and below the observed rate (1.1–7.8%) and considering only HIV-related costs at baseline values shown in Table 1, the cost per newly-identified HIV-infected person receiving test results ranged from US$ 3076 to US$ 21 479. Varying HIV prevalence and program costs simultaneously, the cost per newly diagnosed case of HIV detected ranged from US$ 2576 to US$ 25 191. With STD-specific costs included, total program costs ranged from US$ 4301 to US$ 45 318 per newly-identified HIV-infected person receiving test results.
We found that paying MSM a small incentive to refer peers for HIV, hepatitis and STD testing was an effective means to identify new cases of HIV and viral hepatitis. Our case-finding yield was substantially lower than that observed by Jordan in a report on peer referral conducted in Los Angeles in the 1990s [9,10], but was similar to preliminary data on peer referral conducted by CBOs as part of a CDC-funded multi-center demonstration project . While we cannot be certain why Jordan's program identified so many more cases that either our program or those conducted through CDC, it seems likely that much of this difference is attributable to differences in the populations enrolled and the time periods during which the studies were conducted. Our findings extend those of previous studies by demonstrating that peer referral is cost-effective relative to at least some other publicly supported HIV testing programs. Furthermore, our experience shows that peer referral can be integrated into a public health HIV/STD program, sustained after a period of research funding, and be used to target priority MSM subpopulations.
Although relatively few new cases of HIV were identified through peer referral, this number needs to be interpreted in comparison to other publicly financed HIV testing and counseling efforts, and other recently proposed strategies to increase HIV case-finding. The 22 cases identified through peer referral represent an 8% increase in our health department's total HIV case-finding, and was roughly as effective as our bathhouse-based HIV testing program. This increase is also comparable in magnitude to the 4–9% increase in case detection observed when HIV RNA testing was added to conventional HIV testing in North Carolina , King County, Washington , and San Francisco . Moreover, peer referral was especially effective at identifying non-white MSM with HIV infection, increasing our health department's total case-finding in this group by 19%.
Because cost and cost-effectiveness analysis of HIV testing and counseling programs have used widely variable cost estimates, direct comparison of different interventions across studies is difficult. Economic evaluations of HIV counseling and testing programs have reported costs per new case identified as low as US$ 3120 for testing in prisons (US$ 31 per test with 1% prevalence)  and as high as US$ 56 000 for testing a general US population (US$ 56 per test at 0.1% prevalence) , with most studies reporting costs per case detected between US$ 3500 and US$ 6500, with costs per HIV test varying from US$ 10 to US$ 95 [23–25]. Cost-effectiveness studies of partner notification have reported roughly similar costs per case detected (US$ 2962 to US$ 4632) [26,27], but have been based on estimates of program success that are probably higher than those currently observed in the US . Despite the fact that we included many program costs that most previous analyses excluded (e.g. database management), our findings suggest that the cost of HIV case-finding through peer referral is similar to that of other, widely employed, publicly funded approaches to HIV case-finding. In places like King County, Washington, where the prevalence of HIV among persons tested through publicly financed testing is less than 1% and HIV partner notification identifies very few HIV cases, peer referral appears to be more cost-effective than other outreach approaches to HIV case-finding.
We observed that recruiters referred in men who were demographically similar to them, and that substance using MSM referred other substance using MSM. This finding has operational significance, suggesting that public health programs can use peer referral as a means to identify HIV cases in specific high-priority populations, such as minority MSM and methamphetamine using MSM.
Although we believe our findings are promising, our results may not be generalizable. In particular, our conclusion that peer referral is cost-effective will not be true everywhere. Some public health programs almost certainly have opportunities to expand case-finding programs that have lower marginal costs than peer referral, and other programs might be expanded in which a portion of the costs could be borne by organizations other than public health programs themselves. In such instances, peer-referral may not be the best use of limited public resources. Furthermore, as in other evaluations of HIV case-finding interventions, we cannot estimate when persons identified through peer referral would have been diagnosed in the absence of the program. In comparison with other MSM diagnosed with HIV through our health department during the same period, men diagnosed via peer referral were somewhat more likely to have never HIV tested and had somewhat longer median time periods since their last HIV test, suggesting that peer referral may identify MSM who HIV test less frequently. Finally, although we sought to minimize the number of peers misclassified as being newly diagnosed with HIV by asking them about previous HIV testing, compensating them regardless of their past HIV status, and checking their identities against a list of previously identified cases, some of the men we classified as being newly diagnosed with HIV may have previously tested HIV positive.
In conclusion, our findings suggest that peer referral can be a cost-effective addition to ongoing HIV case-finding efforts among MSM. Wider application of this approach should be considered in other areas.
The authors gratefully acknowledge the assistance of the Harborview Madison Clinic, Lifelong AIDS Alliance, and Gay City in conducting this study; Bill DeYoung for working with participants; Carol Glenn for help recruiting participants; Barbara Krekeler and Frank Chaffee for assistance administering the program and evaluating costs; and Dr. Jane Simoni for assistance with the project's initial design.
Sponsorship: M.R.G. was supported by NIH K23 AI01846. The research was funded by the Division of STD Prevention, National Center for HIV, STD, and TB Prevention, Centers for Disease Control and Prevention.
1. Glynn M, Rhodes P. Estimated HIV prevlaence in the United States at the end of 2003
. Paper presented at: National HIV
Prevention Conference, Atlanta, Georgia, June, 2005.
2. Janssen RS, Holtgrave DR, Valdiserri RO, Shepherd M, Gayle HD, De Cock KM. The Serostatus Approach to Fighting the HIV
Epidemic: prevention strategies for infected individuals. Am J Public Health 2001; 91:1019–1024.
3. CDC. Advancing HIV prevention: New strategies for a changing epidemic
4. CDC. The National Plan to Eliminate Syphilis from the United States.
Atlanta: Center for Disease Control and Prevention; 1999.
5. Pavia AT, Benyo M, Niler L, Risk I. Partner notification
for control of HIV
: results after 2 years of a statewide program in Utah. Am J Public Health 1993; 83:1418–1424.
6. Spencer NE, Hoffman RE, Raevsky CA, Wolf FC, Vernon TM. Partner notification
for human immunodeficiency virus infection in Colorado: results across index case groups and costs. Int J STD AIDS
7. Tomnay JE, Hatch BA, Pitts MK, Carter TR, Fairley CK. HIV partner notification
: a 2002 Victorian audit. Int J STD AIDS
8. Golden MR, Hogben M, Potterat JJ, Handsfield HH. HIV partner notification
in the United States: a national survey of program coverage and outcomes. Sex Transm Dis 2004; 31:709–712.
9. Jordan WC, Tolbert L, Smith R. Partner notification
and focused intervention as a means of identifying HIV
-positive patients. J Natl Med Assoc 1998; 90:542–546.
10. Jordan W, Spring M. HIV focused intervention program.
Paper presented at Infectious Disease Society of America, Philadelphia, November 1999.
11. CDC. Use of social networks to identify persons with undiagnosed HIV infection – seven U.S. cities, October 2003–September 2004
12. Ekwueme DU, Pinkerton SD, Holtgrave DR, Branson BM. Cost comparison of three HIV
counseling and testing technologies. Am J Prev Med 2003; 25:112–121.
13. Laboratories AoPH. 2001 Sexually transmitted diseases laboratory test methods survey [report].
Washington: Association of Public Health Laboratories; 2002.
14. Ciemins EL, Borenstein LA, Dyer IE, Cordero E, Courtney JG, Harvey SM, Richwald GA. Comparisons of cost and accuracy of DNA probe test and culture for the detection of Neisseria gonorrhoeae
in patients attending public sexually transmitted disease clinics in Los Angeles County. Sex Transm Dis 1997; 24:422–428.
15. Howell MR, Quinn TC, Brathwaite W, Gaydos CA. Screening women for chlamydia trachomatis in family planning clinics: the cost-effectiveness of DNA amplification assays. Sex Transm Dis 1998; 25:108–117.
16. Centers for Medicaid and Medicare Services. Clinical and diagnostic laboratory fee (CLAB) schedule 2004 release public use file.
Washington DC: Department of Health and Human Services; 2003.
17. Medical Economics Company. 2003 Drug topics red book
. Montvale: Medical Economics Company, Inc; 2003.
18. Pilcher CD, Fiscus SA, Nguyen TQ, Foust E, Wolf L, Williams D, et al
. Detection of acute infections during HIV
testing in North Carolina. N Engl J Med 2005; 352:1873–1883.
19. Stekler J, Swenson PD, Wood RW, Handsfield HH, Golden MR. Targeted screening for primary HIV
infection through pooled HIV
-RNA testing in men who have sex with men. AIDS
20. Klausner JD, Grant RM, Kent CK. Detection of acute HIV
infections. N Engl J Med 2005; 353:631–633, author reply 631–633.
21. Varghese B, Peterman TA. Cost-effectiveness of HIV
counseling and testing in US prisons. J Urban Health 2001; 78:304–312.
22. Paltiel AD, Weinstein MC, Kimmel AD, Seage GR 3rd, Losina E, Zhang H, et al
. Expanded screening for HIV
in the United States – an analysis of cost-effectiveness. N Engl J Med 2005; 352:586–595.
23. Phillips KA, Fernyak S. The cost-effectiveness of expanded HIV
counselling and testing in primary care settings: a first look. AIDS
24. Walensky RP, Losina E, Malatesta L, Barton GE, O'Connor CA, Skolnik PR, et al
. Effective HIV
case identification through routine HIV
screening at urgent care centers in Massachusetts. Am J Public Health 2005; 95:71–73.
25. Kassler WJ, Dillon BA, Haley C, Jones WK, Goldman A. On-site, rapid HIV
testing with same-day results and counseling. AIDS
26. Toomey KE, Peterman TA, Dicker LW, Zaidi AA, Wroten JE, Carolina J. Human immunodeficiency virus partner notification
. Cost and effectiveness data from an attempted randomized controlled trial. Sex Transm Dis 1998; 25:310–316.
27. Varghese B, Peterman TA, Holtgrave DR. Cost-effectiveness of counseling and testing and partner notification
: a decision analysis. AIDS