In the United States, more than 60% of new HIV infections occur among men who have sex with men (MSM),1 although they are estimated to represent only 2% of the United States population.2 Behavioral surveillance among MSM is used to monitor the prevalence and correlates of HIV risk and preventive behaviors in advance of HIV disease outcomes.3 In 2003, the Centers for Disease Control and Prevention (CDC), in collaboration with state and local health departments in areas with high AIDS prevalence, established the National HIV Behavioral Surveillance System (NHBS) to collect behavioral surveillance data from populations at risk for HIV infection, including MSM.4 The objective of this surveillance system is to monitor the prevalence of and trends in HIV risk behaviors, HIV testing behaviors, and exposure to HIV prevention services.4
To enroll MSM, NHBS staff selected respondents using a venue-based sampling method.5 To ensure enrollment of a diverse population by race, ethnicity, and age, 12 types of venues were eligible for inclusion in NHBS sampling frames; gay bars, cafés and restaurants, dance clubs, fitness clubs or gymnasia, gay pride events, house parties, social organizations, parks and beaches, retail businesses, street locations, raves and circuit parties, and sex establishments, such as adult bookstores, adult theaters, and bathhouses. Using data from the 2002 Urban Men's Health Study in San Francisco, Pollack et al. evaluated the NHBS protocol before data collection.6 Their evaluation concluded that NHBS could be conducted without diminishing representativeness by limiting the sample to gay bar and club venues, and that it could be completed in 6 months instead of the 12 months allowed by the NHBS protocol. Further, the authors speculated that venue-based sampling would overestimate risk behaviors compared with the household probability sampling used in the Urban Men's Health Study. In response to this evaluation, CDC committed to conduct an evaluation of the NHBS data collection, including an evaluation of the sampling strategy, after the NHBS data were available.7
To evaluate the impact of restricting the sampling frames to bars and clubs in future NHBS cycles, we used data from the 2003–2005 NHBS MSM cycle to compare enrollment outcomes and the characteristics of respondents enrolled at bars or clubs with those enrolled at other types of venues.
The sampling methods of NHBS have been previously described.5 To implement the venue-based sampling scheme, investigators in each metropolitan area conducted formative research to create lists of venues where MSM congregate.8 Venues where men received HIV/sexually transmitted disease testing or HIV-related care, where the logistics of conducting surveys were infeasible (e.g., management disapproval) or that were considered unsafe were not eligible. For each eligible venue, staff ascertained all potential venue-day-time periods (VDTs) (e.g., a bar on Wednesday nights from 8:00 PM to 12:00 AM) that were estimated to produce at least 8 eligible men. Sampling frames were updated monthly to include new VDTs or exclude VDTs based on attendance, logistic, and safety criteria. On a monthly basis, approximately 15 VDTs were selected randomly in 2 stages (venues followed by day-time periods) and were scheduled on a calendar for sampling MSM.
During sampling events, all men who were at least 18 years of age who crossed a predetermined line or a defined area at the venue were counted. A sample of men was approached, enrolled, and interviewed to determine eligibility. Residents of the NHBS metropolitan area who were ≥18 years of age, born male, spoke either English or Spanish, and had not already participated in NHBS were eligible and were invited to participate in an interview lasting approximately 30 minutes. Interviews were typically conducted in a confidential space at the venue or in a large van. The minimum target sample size was 500 MSM for each NHBS metropolitan area and planned time line for completion of data collection was 12 months. Actual data collection occurred from November 2003 to April 2005, with varying initiation and completion dates by NHBS metropolitan area within this time frame.
The questionnaire assessed demographic characteristics, attendance at gay venues, risk behavior, and use of prevention services. Gay venue attendance was ascertained by response to the question, “In the past 12 months, how often have you gone to a place where gay men hang out, meet, or socialize? These could include bars, clubs, bookstores, sex clubs, social organizations, parks, gay-owned or -operated businesses, etc.” Unless otherwise stated, respondent behaviors refer to those that occurred in the 12 months preceding the interview date.
Enrollment location was categorized into 3 types of venues: bars and clubs, sex venues, and other venues. The bar and club category included gay bars and gay dance clubs; the sex venues category included sex establishments and environments, such as adult bookstores, adult theaters, and bathhouses; and the other venue category included cafés and restaurants, fitness clubs or gymnasia, gay pride events, social organizations, parks and beaches, retail businesses, street locations, raves and circuit parties, and any other remaining venue type. We chose sex venues and all other venues as reasonable comparison groups to assess how a bar/club sample of MSM might differ from other samples of MSM. Sex venues were used as a separate comparison group because of suspected differences in characteristics of MSM enrolled from these venues compared with MSM enrolled in the 9 types of venues included in the other category.
We restricted this analysis to male respondents who reported at least 1 male sex partner in the past 12 months using data submitted from the following 15 metropolitan areas: Atlanta, GA; Baltimore, MD; Boston, MA; Chicago, IL; Denver, CO; Fort Lauderdale, FL; Houston, TX; Los Angeles County, CA; Miami, FL; New York, NY; Newark, NJ; Philadelphia, PA; San Diego, CA; San Francisco, CA; and San Juan, PR.
All analyses were conducted using SAS, version 9.1 (SAS Institute, Cary, NC). We first evaluated enrollment outcomes by comparing the participation rate (number of enrolled divided by the number of eligible) and enrollment efficiency rate (number of enrolled divided by the number approached) among the 3 venue types. We next compared the characteristics of respondents enrolled from bars and clubs with those enrolled in sex venues and at all other venues in bivariate analyses. For these analyses, we report odds ratios and 95% confidence intervals separately for each comparison. We used logistic regression modeling to estimate odds ratios and 95% confidence intervals. Models included variables that were statistically significant (P < 0.05) in the bivariate analysis or were a priori considered to be potentially confounding factors (NHBS metropolitan area, age, and race/ethnicity). All NHBS areas included bars and clubs and at least 1 of the other venue types in their sampling frame. Five NHBS areas did not include sex venues in their sampling frame because they did not have permission to conduct interviews in these locations or for safety reasons. Therefore, we restricted the multivariable model comparing sex venues with bars and clubs to NHBS cities that included sex venues in their sampling frame.
Enrollment Comparison by Venue
As previously reported, 23,681 persons were approached for an NHBS interview.9 Of these, enrollment and eligibility outcome data were available for 23,452 (99%) persons, of whom 3964 (17%) refused to answer eligibility criteria questions (range: 13% at bars and clubs; 21% at other venues). Of the 19,488 persons who answered the eligibility criteria questions, 17,322 (89%) were eligible (bars and clubs: 89%; sex venues: 86%; other venues: 89%) (Fig. 1). Of eligible persons, 14,049 (81%) participated (bars and clubs: 82%, sex venues: 83%, other venues: 80%) (Fig. 1). The overall efficiency rate was 60% (bars and clubs: 63%; sex venues: 58%; other venues: 57%). We excluded 278 (2%) participants who did not identify as male, 1919 (14%) who did not report having sex with another man during the 12 months before the interview, and 381 (3%) with incomplete surveys, resulting in 11,471 participants included in this analysis.
The total number of respondents by NHBS metropolitan area ranged from 445 to 1419. The median length of data collection was 10 months (range by metropolitan area: 3–13 months). Of the 15 NHBS cities, 6 reached the protocol target number of 500 or more MSM in 6 months or less.
Of the 3 enrollment venue categories, 6419 (56%) respondents were enrolled at bars and clubs, 481 (4%) at sex venues, and 4571 (40%) at other venues (Table 1). The proportion of respondents enrolled at the 3 venue categories varied by NHBS metropolitan area (bars and clubs: 31%–89%, sex venues: 0%–9%, other venues: 7%–70%).
Comparison of Characteristics by Venue
Compared with MSM enrolled at bars and clubs, MSM enrolled at sex venues were more likely to be older than 30 years, be Hispanic or black, identify as heterosexual or bisexual, not have told anyone that they were attracted to or had sex with men, have indeterminate or unknown HIV infection status, have a female sex partner, have 5 or more male sex partners, and have a casual male sex partner (Table 2). MSM enrolled at sex venues were less likely than men enrolled at bars and clubs to have obtained at least a high school education, have private health insurance, rent a house or apartment, visit gay venues at least weekly, visit gay Internet chat rooms at least weekly, and to use noninjection illicit drugs (Table 2). In addition, men enrolled at sex venues were less likely to engage in unprotected anal sex than men enrolled in bars and clubs (Table 2). Results from a multivariable model were similar to the bivariate analysis, indicating that enrollment venue category was independently associated with older age, race/ethnicity, sexual identity, health insurance status, frequency of gay venue attendance, and larger numbers of male sex partners, after adjusting for NHBS metropolitan area (Table 3).
Compared with MSM enrolled at bars and clubs, MSM enrolled at other venues were more likely to be ≤20 years or ≥40 years of age, be Asian/Pacific Islander, identify as heterosexual, not to have told anyone that they were attracted to or had sex with men, have public health insurance, report being HIV positive, to live in a hotel or rooming house, be homeless, frequently visit gay Internet chat sites, or have 10 or more male sex partners (Table 2). MSM enrolled at other venues were less likely than MSM enrolled at bars and clubs to visit gay venues, be tested for HIV in the past 12 months, have a main sex partner, or use noninjection illicit drugs (Table 2). Results from a multivariable model were similar to the bivariate analysis, indicating that enrollment venue category was independently associated with age, race/ethnicity, sexual identity, disclosure of sexual identity, living situation, frequency of gay venue attendance, frequency of gay Internet chat, and noninjecting drug use, after adjusting for NHBS metropolitan area (Table 3).
The purpose of this analysis was to describe differences in demographic and behavioral characteristics of MSM participating in NHBS by the type of venue from which they were enrolled. In general, the prevalence estimates of demographic and behavioral characteristics were similar among MSM enrolled in bars and clubs compared with the prevalence estimates overall. However, MSM enrolled at sex venues and other venues were different on many important demographic and behavioral characteristics than MSM enrolled at bars and clubs. Although the enrollment efficiency rate was higher among men enrolled in bars and dance clubs, the overall participation rate was the same across the venue type. Therefore, findings from our analyses do not support restricting NHBS sampling frames to bars and clubs.
There were several demographic differences between NHBS respondents by venue category. Respondents who were enrolled at sex venues or at other venues were more likely to be older than those enrolled at bars and clubs. Black and Hispanic MSM were more likely to be enrolled at sex venues than at bars and clubs. Given that the incidence of HIV is highest among black and Hispanic MSM,1 continuing to include NHBS enrollment at venues other than bars and clubs will be important to ensure that these groups are adequately represented.
Monitoring HIV testing among MSM is an important national indicator for HIV prevention. MSM enrolled at other venues were less likely to report having been tested for HIV in the past 12 months, although the absolute differences were small. Therefore, limiting the NHBS sampling frame to bars and clubs would have resulted in the same proportion of MSM who have recently been tested for HIV infection.
However, several other characteristics were different by venue category. Although there were no differences in reported unprotected anal sex with either main or casual partners, MSM who reported 10 or more sex partners were more likely to be enrolled at sex venues than at bars and clubs. This suggests that enrollment at sex venues may sample a subset of MSM that is more likely to engage in sexual behaviors that might not be represented at other sampling venues.
There were differences in both sexual identity and the frequency of gay venue attendance by venue category. MSM who identified as bisexual were more likely to be enrolled in sex venues and less likely to be enrolled at other venues than at bars and clubs. MSM who had not told anyone of their attraction to men were more likely to be enrolled at other venues than at bars and clubs. This is similar to the results of the Young Men's Survey, which showed that MSM who did not disclose their attraction to men were less likely to attend gay bars or dance clubs.10 Studies have shown that nondisclosing MSM report less high-risk activity10,11 and are less likely to be HIV positive12 than MSM who do disclose. The impact of restricting NHBS to bars and clubs would likely reduce enrollment of MSM who do not identify as homosexual or bisexual, and might overestimate the national prevalence estimates of high-risk behavior among MSM.
There are several limitations to this analysis. Although NHBS did not collect information on the overlap in venue attendance across the 3 categories, several studies have shown that MSM who visit gay bars and clubs also visit other gay venues.13,14 However, our results showed that respondents enrolled from either sex venues or other venues were less likely to attend gay venues, such as gay bars and clubs. Behaviors may be either over- or underestimated depending on the extent to which they are associated with the frequency of venue attendance, and therefore the likelihood of selection for NHBS. Five of the NHBS metropolitan areas did not include sex venues in their sampling frame, which limits the comparability among NHBS metropolitan areas and the external validity of the NHBS method.
Monitoring HIV risk behaviors among MSM is a crucial component of HIV prevention.4 Pollack et al. have suggested that NHBS venue-based sampling should be validated periodically using population-based probability sampling.6,15 A comparison of respondent characteristics from the CDC's HIV Testing Survey, which used convenience sampling in gay bars and clubs, with a random digit dial probability sampling survey of MSM in San Francisco showed that there were significant differences in demographic characteristics, although sexual risk behaviors were similar.15 One important limitation of using telephone surveys for surveillance is that sampling MSM would be inefficient outside of metropolitan areas with a high concentration of MSM. The Urban Men's Health Study screened more than 95,000 households in San Francisco, New York, Los Angeles, and Chicago, identified 3700 eligible households with at least 1 MSM, and enrolled 2881 respondents.12 Although the participation rate of the Urban Men's Health Study was comparable with NHBS (78% vs. 81%), the efficiency rate was lower (3% vs. 60%). In addition, response rates to telephone surveys in general have declined in recent years.16 Significant demographic and behavioral differences, including recent HIV testing, have been described between persons in households with landlines, who are eligible for telephone surveys, and persons in cell phone-only households, who generally are not eligible for or effectively reached by telephone surveys.17 These differences may be particularly important for younger and low-income populations, who are at high risk for HIV infection and, therefore, especially important to accurately monitor in HIV behavioral surveillance.18 The changing dynamics of the gay community mean that NHBS must remain flexible in future cycles to ensure inclusion of a broad sample of MSM, particularly young MSM and MSM of color, and highlight the importance of the formative research to identify NHBS sampling venues before data collection.19 Periodic evaluation of NHBS using comparisons with other MSM sampling methods, including Internet-based surveys, is critical to our understanding of possible biases and limitations to our method.
NHBS is the largest, most geographically diverse, ongoing and systematic surveillance system to monitor HIV risk among MSM in the United States. The success of the surveillance system depends on the support of stakeholders, including Community Advisory Boards and the MSM community, to achieve high participation rates and a diverse sample of venues and participants.5 The perception and participation of stakeholders is important to the success of NHBS and should be carefully considered before the sampling methods are modified. NHBS depends on community support and venue management approval. Restricting sampling frames to gay bars and clubs would place a greater burden on fewer venues to accommodate interview teams, particularly in metropolitan areas with fewer gay bars and clubs. The restricted sampling frame may also unnecessarily limit the size of the population that is eligible to participate. Although NHBS may not change the types of venues included in the sampling frame for this reason, our analysis suggests that limiting venue-based enrollment of MSM to bars and clubs may be a viable option for other studies or behavioral surveillance activities. Because there is no gold standard method for sampling MSM, those conducting surveys must be aware of the potential biases of each sampling strategy.
1. Prejean J, Song R, Hernandez A, et al.. Estimated HIV incidence in the United States, 2006–2009. PLoS One 2011; 6:e17502.
2. Purcell DW, Johnson C, Lansky A, et al.. Calculating HIV and syphilis rates for risk groups: Estimating the national population size of men who have sex with men. Paper presented at: National STD Prevention Conference; 2010; Atlanta, GA.
3. Lansky A, Sullivan PS, Gallagher KM, et al.. HIV behavioral surveillance in the US: A conceptual framework. Public Health Rep 2007; 122(suppl 1):16–23.
4. Gallagher KM, Sullivan PS, Lansky A, et al.. Behavioral surveillance among people at risk for HIV infection in the US: The National HIV Behavioral Surveillance System. Public Health Rep 2007; 122(suppl 1):32–38.
5. MacKellar DA, Gallagher KM, Finlayson T, et al.. Surveillance of HIV risk and prevention behaviors of men who have sex with men—a national application of venue-based, time-space sampling. Public Health Rep 2007; 122(suppl 1):39–47.
6. Pollack LM, Osmond DH, Paul JP, et al.. Evaluation of the Center for Disease Control and Prevention's HIV behavioral surveillance of men who have sex with men: Sampling issues. Sex Transm Dis 2005; 32:581–589.
7. Lansky A, MacKellar D, Gallagher KM, et al.. Evaluation of the new National HIV Behavioral Surveillance System (NHBS) is imperative. Sex Transm Dis 2006; 33:272–273; author reply 274–275.
8. Allen DR, Finlayson T, Abdul-Quader A, et al.. The role of formative research in the National HIV Behavioral Surveillance System. Public Health Rep 2009; 124:26–33.
9. Sanchez T, Finlayson T, Drake A, et al.. Human immunodeficiency virus (HIV) risk, prevention, and testing behaviors—United States, National HIV Behavioral Surveillance System: Men who have sex with men, November 2003–April 2005. MMWR Surveill Summ 2006; 55:1–16.
10. CDC. HlV/STD risks in young men who have sex with men who do not disclose their sexual orientation—six US cities, 1994–2000. Morb Mortal Wkly Rep 2003; 52:81–86.
11. Pathela P, Hajat A, Schillinger J, et al.. Discordance between sexual behavior and self-reported sexual identity: A population-based survey of New York City men. Ann Intern Med 2006; 145:416–425.
12. Catania JA, Osmond D, Stall RD, et al.. The continuing HIV epidemic among men who have sex with men. Am J Public Health 2001; 91:907–914.
13. Xia Q, Tholandi M, Osmond DH, et al.. The effect of venue sampling on estimates of HIV prevalence and sexual risk behaviors in men who have sex with men. Sex Transm Dis 2006; 33:545–550.
14. Fisher HH, Purcell DW, Hoff CC, et al.. Recruitment source and behavioral risk patterns of HIV-positive men who have sex with men. AIDS Behav 2006; 10:553–561.
15. Schwarcz S, Spindler H, Scheer S, et al.. Assessing representativeness of sampling methods for reaching men who have sex with men: A direct comparison of results obtained from convenience and probability samples. AIDS Behav 2007; 11:596–602.
16. Mokdad AH, Stroup DF, Giles WH. Public health surveillance for behavioral risk factors in a changing environment. Recommendations from the Behavioral Risk Factor Surveillance Team. MMWR Recomm Rep 2003; 52(RR-9):1–12.
17. Blumberg SJ, Luke JV, Cynamon ML. Telephone coverage and health survey estimates: Evaluating the need for concern about wireless substitution. Am J Public Health 2006; 96:926–931.
18. Blumberg SJ, Luke JV, Davidson G, et al.. Wireless substitution: State-level estimates from the National Health Interview Survey, January-December 2007. Natl Health Stat Report 2009:1–13, 16.
19. Simon Rosser BR, West W, Weinmeyer R. Are gay communities dying or just in transition? Results from an international consultation examining possible structural change in gay communities. AIDS Care 2008;20:588–595.