NHBS is a CDC-funded behavioral surveillance system conducted through annual rotating cycles of surveys and HIV testing targeting one of three populations at high risk for acquiring HIV: MSM, persons who inject drugs, and heterosexuals at increased risk for HIV infection. NHBS data are used to monitor prevalence and trends in HIV-related risk and testing behaviors and access to HIV prevention services [23]. During the MSM cycles, local surveillance staff conduct venue-based, time-space sampling following the national NHBS protocol, which organizes sampling activities into three components. First, staff conduct formative research to identify the venues, times, and methods to recruit MSM. Next, staff construct sampling frames of eligible venues and venue-specific, daytime periods that meet MSM attendance, logistical, and safety eligibility criteria. A calendar of sampling events is constructed by randomly selecting venues and day-time periods from the sampling frame. CDC is in the process of developing weights to account for variations in venue attendance and likelihood of being selected for NHBS participation. The final component of venue-based time-space sampling involves recruiting and interviewing men during sampling events.

In Denver, sampling events were conducted between August and November 2008 for the second NHBS cycle among MSM (NHBS-MSM2) and between August and November 2011 for the third cycle of MSM (NHBSMSM3). During each 4-h sampling event, NHBS field staff approached men at selected venues to assess eligibility. Selected venues included bars, dance clubs, bathhouses, parks, and local grocery stores. The venues at which sampling events took place remained largely unchanged between the 2008 and 2011 MSM cycles. For both cycles, all potential participants had to be: (1) born male (2) approached by study staff at the sampled venues, (3) 18 years or older, (4) residents of the Denver metropolitan statistical area, (5) not previously completed an interview for the current NHBS-MSM cycle, (6) able to complete the survey in English or Spanish, and (7) able to provide informed consent. Neither sexual orientation nor same-sex behaviors were eligibility criteria. All NHBS activities were voluntary and anonymous. Verbal informed consent was obtained from eligible participants. Participants were also offerred HIV testing.

Participants completed an interviewer-administered behavioral risk survey using handheld computers. The behavioral risk survey included questions about sexual behaviors, substance use, STI history, and HIV testing. In addition to the behavioral risk survey, which is a standardized questionnaire used across all NHBS sites, participants completed a shorter local questionnaire. In Denver, a PrEP-specific supplemental questionnaire was included during the 2008 (NHBS-MSM2) and 2011 (NHBS-MSM3) cycles to assess knowledge of PrEP; willingness to use PrEP if the clinical trials showed few or no side effects; willingness to use PrEP if the clinical trials showed PrEP prevents HIV infection in 75 and 50 % of people who take it daily; and potential risk compensation, including intentions to have sex with more partners or use condoms less frequently if PrEP were found to be effective.

To introduce the concept of PrEP, the following statement was read to participants: “Scientists are currently doing studies to find new ways of preventing people from becoming infected with HIV. In these studies, people take a pill every day that contains the same medicine that is used to treat people who are infected with HIV. Scientists want to know if taking this medicine will prevent people exposed to HIV from becoming infected with it. They call this method pre-exposure prophylaxis or PrEP.” Participants were then asked if they had ever heard of PrEP. This was followed by the question: “imagine that PrEP was available today. If studies showed that PrEP has few or no side effects, would you be willing to take PrEP pills every day to try to protect yourself from becoming infected with HIV?” If participants responded in the affirmative to this question, they were asked “If these studies also showed that PrEP prevents HIV infection in three-quarters or 75 % of the people who take it daily, would you be willing to take PrEP pills every day to try to protect yourself from becoming infected with HIV?” If their response was in the affirmative, they were asked the same question but specifying 50 % rather than 75 % effectiveness. The wording of the questions remained the same during both cycles.

The current analysis was restricted to participants who reported having had oral or anal sex with a man in the preceding 12 months and who did not report being HIV-positive. Descriptive frequencies are presented for each sample. χ² statistics were calculated to compare the two samples from 2008 and 2011. We also generated a multivariable model using a generalized linear model with log link and binomial error distribution to assess the primary association between knowledge of PrEP and year of data collection, controlling for characteristics that differed significantly between the two samples. To account for potential correlation of observations within venues, a robust variance estimator was calculated [24]. Unadjusted and adjusted prevalence ratios (PR) and associated 95 % confidence intervals (CI) are presented. As a final step, comparisons between participants who reported knowledge of PrEP and those who did not were made using Chi square (χ²) statistics separately for each sample. Fisher's exact test was calculated when minimum expected frequency requirements were not met. All analyses were conducted using Stata Version 12.0 (StataCorp, College Station, TX).

Between August and November 2008, 735 men were screened for the NHBS-MSM2 survey in Denver, Colorado. Of these men, 503 were eligible for NHBS and completed the PrEP supplemental questionnaire. After excluding men who did not have sex with a man in the previous 12 months and those who reported being HIV positive, the resulting sample size was 425. Of the 425 participants included in the analysis, almost two-thirds (63 %) described their race/ethnicity as non-Hispanic white, 3 % identified as non-Hispanic black, 26 % as Hispanic, and the remaining 7 % described themselves as Asian, Pacific Islander, American Indian/Alaskan Native or other race/ethnicity (Table 1, NHBS-MSM2). Almost two-thirds (66 %) of participants were over the age of 30 years, 73 % had more than a high school education, and 52 % reported an annual income of $40,000 or more. A little less than one-fourth (22 %) of participants reported living with a man they considered to be a boyfriend, spouse, significant other, or life partner. More than three-quarters of participants (77 %) reported two or more male sexual partners during the past 12 months and 37 % reported at least one act of unprotected anal intercourse during the past 12 months.

A total of 91 (21 %) participants reported that they had ever heard of PrEP before their interview (Table 2, NHBSMSM2). Two-thirds of participants (66 %) reported that they would take PrEP every day if studies showed no side effects. The proportion of men in Denver reporting their willingness to take PrEP daily if shown to be 75 and 50 % effective was 60 and 45 %, respectively. Prior to the release of the iPrEX trial results, the majority of participants did not anticipate changing either their condom use (85 %) or the number of sex partners (92 %) if they were taking PrEP every day.

Between August and November 2011, 678 men were screened for NHBS-MSM3. Of these men, 527 were eligible for NHBS and completed the PrEP supplemental questionnaire. After excluding men who did not have sex with a man in the previous 12 months and those who reported being HIV positive, the resulting sample size was 461. Slightly more than half (54 %) of the 461 participants described their race/ethnicity as non-Hispanic white, 11 % identified as non-Hispanic black, 27 % as Hispanic, and the remaining 8 % described themselves as Asian, Pacific Islander, American Indian/Alaskan Native or other race/ethnicity (Table 1, NHBS-MSM3). Almost two-thirds (62 %) of participants were over the age of 30 years, 76 % had more than a high school education, and 43 % reported an annual income of $40,000 or more. In the 2011 sample, one-fifth (20 %) of participants reported living with a man they considered to be a boyfriend, spouse, significant other, or life partner. A large majority (80 %) of participants reported two or more male sexual partners during the past 12 months and 36 % reported at least one act of unprotected anal intercourse during the past 12 months.

In the 2011 sample, a total of 131 (28 %) participants reported that they had heard of PrEP before their interview (Table 2, NHBS-MSM3). Compared to 66 % of the sample in 2008, 62 % of participants in the 2011 sample reported that they would take PrEP every day if studies showed no side effects. In 2011, the proportions of men reporting their willingness to take PrEP at different levels of effectiveness was also slightly lower with 56 % of men reporting willingness to take PrEP every day if shown to be 75 % effective and 44 % willing to take daily PrEP if shown to be 50 % effective. Post iPrEx trial results being released, a slightly higher proportion of men in the 2011 sample reported some anticipated changes in behavior with 11 % reporting less frequent condom use and 10 % reporting more condom use. Similarly, a higher proportion of the 2011 sample reported a change in anticipated number of partners with 8 % anticipating fewer partners and 6 % anticipating more partners. This increase is compared to the 2008 sample where 3 % anticipated having fewer partners and 4 % anticipated having more partners (χ² = 12.59, P = 0.002).

Among participants in the 2011 sample, 131 (28 %) had heard of PrEP compared with 91 (21 %) participants in the 2008 sample (crude PR: 1.32, 95 % CI: 1.07, 1.65; Table 3). Accounting for potential clustering within each venue and the potentially confounding effects of race/ethnicity, age, income, anticipated changes in condom use and number of partners as these factors differed between the two samples, participants in the 2011 sample were 43 % more likely to have heard of PrEP (adjusted PR: 1.43, 95 % CI: 1.18, 1.72).

In both samples, knowledge of PrEP differed significantly by income level (Table 4). In the 2008 sample, among participants who had heard of PrEP 65 % reported an annual income of $40,000 or more, 24 % reported an annual income between $20,000 and $39,000, and 11 % reported an annual income of $19,999 or less (χ² = 9.23, P = 0.01). In the 2011 sample, among participants who had heard of PrEP 55 % reported an annual income of $40,000 or more, 21 % reported an annual income between $20,000 and $39,000, and 24 % reported an annual income less of $19,999 or less (χ² = 10.66, P = 0.005).