Despite intensive HIV prevention efforts and strong care and treatment programs in the United States, the number of annual new infections has not decreased over the past 10 years.1 Over 40,000 new HIV infections are estimated to occur in the United States every year, with 50.0% of those in persons <25 years of age and the majority in racial and ethnic minorities, women, and men who have sex with men (MSM).1-3 Scientists and public health officials agree that a preventive HIV vaccine remains our best hope for ending the AIDS pandemic. Despite significant efforts in HIV research over the past decade, there is still no vaccine to prevent HIV infection. However, scientists continue to make progress toward the goal of developing a preventive HIV vaccine that is safe, effective, affordable, and appropriate for use around the world. Development of an HIV vaccine requires not only laboratory research and strong vaccine candidates but also a social environment that is supportive of clinical testing of the candidate vaccines in human clinical trials.
In 2001, the Division of AIDS at the National Institutes of Allergy and Infectious Diseases (NIAID), part of the National Institutes of Health, an agency of the Department of Health and Human Services, launched an HIV vaccine communications campaign (HVCC) aimed at raising awareness and acceptance of HIV vaccine research and creating and sustaining a supportive environment suitable for clinical HIV vaccine research in the general public and in specific populations disproportionately affected by HIV, such as African Americans, Hispanics, and MSM. Given that >30 phase I, II, and III HIV preventive vaccine clinical trials are currently underway or planned, which will require tens of thousands of volunteers, preparing communities to consider participating in the development and testing of an HIV vaccine is essential.4
Although successful treatments for HIV infection have taken AIDS off the “crisis” radar for most Americans, new HIV infections and AIDS cases remain at alarmingly high rates, particularly among women of color and young MSM of color.1 These 2 groups are vulnerable populations that must be educated and encouraged to participate in HIV vaccine clinical trials. Participation of these populations and other groups at risk for HIV is vital to ensure that the results of the study represent everyone and are applicable to all populations. Unfortunately, current data suggest that there are many barriers for trial participation, including perceived vaccine safety concerns, time commitment, social stigma associated with HIV/AIDS, perceived social harms associated with vaccine-induced antibody positivity, mistrust of government, homophobia, and misunderstanding of research.4-7 Although African Americans and Hispanics represent approximately 25.0% of the general population and >50.0% of all new HIV infections in the United States, they accounted for only 13.0% and 4.0%, respectively, of participants in US trials conducted by the NIAID-sponsored HIV Vaccine Trials Network (HVTN).8 In the privately funded VaxGen phase III HIV preventive vaccine clinical trial conducted in North America and Europe, African Americans and Hispanics represented only 6.4% and 6.8%, respectively, of 5009 trial volunteers.9 Strategies to enhance the representation of racial and ethnic minorities in HIV preventive vaccine clinical trials are sorely needed.9 To develop a campaign and education program to increase the awareness of and support for vaccine research, it was necessary to determine a baseline for these parameters. To accomplish this, a telephone survey was developed to assess the attitudes, awareness, and knowledge of HIV vaccine research among various populations in the United States. To date, there has not been a comparable quantitative study of this scope conducted. Limited quantitative data currently exist on how Americans and communities disproportionately affected by the epidemic in general feel about an HIV vaccine. Previous research efforts have focused on specific issues in limited populations and included vaccine trial volunteers or a cohort representing potential trial participants. To our knowledge, the study presented here is the first of its kind.
One survey, administered to 4 different populations, was fielded using the following sampling methods. The general population was nationally sampled using the Mitofsky-Waksberg Random Digit Dial sampling design (Survey Sampling, Inc., Fairfield, CT). Up to 10 call attempts were made at various times during the day and week to each household in an effort to reach a respondent. The highly impacted populations were sampled from geographically targeted areas in the nation. The MSM group was drawn from San Francisco, New York, Chicago, and Los Angeles using high-density zip codes identified by researchers at the Center for AIDS Prevention Studies at the University of California at San Francisco (Urban Disease Prevention Survey, supported by National Institutes of Mental Health). The African American and Hispanic groups were sampled by including high-density areas identified from an analysis of Claritas' PRIZM, the geodemographic targeting database used to identify areas of key target US markets in which to concentrate research efforts (Claritas, Inc., San Diego, CA). PRIZM defines every neighborhood in the United States in terms of demographically and behaviorally distinct types or “clusters.” Sampling for the Hispanic population was from PRIZM's Hispanic mix, Latino America, and “starter family” clusters, and for African Americans, sampling was obtained from Inner City and Southside City Clusters. Several of these clusters were identified by the Centers for Disease Control and Prevention as racial, ethnic, and demographic populations living in areas disproportionately affected by HIV/AIDS, populations that could potentially have participants for future HIV vaccine clinical trials.
For each group, the cooperation rate was defined as the proportion of eligible individuals completing the interview after they were identified. The adjusted response rate was defined as the proportion of eligible households in the Random Digit Dial sample frame in which an interview was completed, adjusting for the number of unanswered and unscreened telephone numbers who would not have been eligible if reached (according to the American Association for Public Opinion Research).
From December 2002 through February 2003, Field Research, Inc., San Francisco, CA, conducted 3509 telephone interviews among adults in the United States, which included a sample of randomly selected American adults and adults from 3 highly impacted populations (those who are disproportionately affected by HIV/AIDS): African Americans, Hispanics, and MSM. All participants were at least 18 years old. Participants were initially screened in a presurvey questionnaire to ensure they met the selection criteria. The selection criteria for the African American sample was men and women living in US cities that bear a disproportionate burden of the HIV/AIDS epidemic due to socioeconomic factors, including poverty, unemployment, low educational attainment, and lack of access to health care. The Hispanic group was selected under the same conditions with the exception that participants could be from any race but self-identified as being of Hispanic or Latino origin. African American and Hispanic participants were not asked about their sexual orientation. MSM were identified and selected as men of any race who identify themselves as gay or bisexual, or who have had sex of any kind with a man since age 18, and lived in high-density neighborhoods in San Francisco, Los Angeles, Chicago, and New York. All participants were offered a choice of completing the interview in English or Spanish and the average interview time ranged from 17 to 20 minutes.
Development of Key Themes and Messages
To develop key themes and messages for the survey questionnaire, qualitative one-on-one interviews were conducted with 45 community leaders and members of the HVCC's steering group, which is composed of approximately 25 individuals who represented health care providers, HIV prevention educators, public health professionals, AIDS treatment providers, and representatives of AIDS service organizations and AIDS-related community-based organizations. In addition, 28 focus groups were conducted with 251 members of the general public in 9 US cities recruited by professional focus group research vendors. Finally, between July and August 2001, a media content review was performed using electronic and traditional sources of data, including newspapers, journals, and other publications. Both national and regional publications were analyzed, as well as publications in both English and Spanish. They included a total of 7 national, 25 regional, 29 alternative, and 61 audience-specific publications targeting black/African Americans, Hispanics/Latinos, or gay and lesbian audiences.
The focus groups' results and media content analysis were used to identify several key themes or messages surrounding HIV vaccine research. For the focus groups, 3 characteristics of respondents' comments were considered following coding of recorded transcripts: frequency, extensiveness, and intensity. Frequency refers to the number of times a comment was made; extensiveness refers to the number of people who said it; and intensity refers to how strongly the opinion or comment was expressed. Key findings were drawn from central tendency data, where a majority of focus group participants expressed similar opinions. In addition to questions included in the discussion guide, respondents reacted to concept statements about HIV vaccine research presented on paper. Three sets of message concepts were used as discussion stimuli across the 28 focus groups. Following an initial round of 6 groups in Columbus, Chicago, and Houston, revisions were made to the 1st set of messages and new ones were created. This set of messages was then revised based on additional findings from 11 focus groups in Los Angeles, San Francisco, and Fort Lauderdale. The 3rd and final set of messages was used for the last 11 focus groups in Nashville, Baltimore, and New York. With each new set of concept statements, respondents' comments were incorporated and improvements had been made to the messages. From this exercise, 5 key concepts surrounding HIV vaccine research were identified, as follows: there is currently no vaccine to prevent HIV infection; only HIV-negative individuals can volunteer for a preventive HIV vaccine trial; you cannot be infected with HIV from the vaccines being tested; all populations must be involved in HIV vaccine research; and an HIV preventive vaccine, complemented by strong behavioral prevention programs and AIDS care and treatment, is the best way to end the epidemic.
To ascertain and validate our preliminary key themes and messages, a 27-question survey was developed. The survey questions were organized into categories including salience of HIV/AIDS as a concern, attitudes toward fighting AIDS, awareness of HIV vaccine research, attitudes concerning HIV vaccine research, knowledge of HIV vaccines research, interest in and support of HIV vaccine research, and standard demographics. A variety of ordinal variables were used for many of the questions, most of which included standard, balanced 5-point Likert-type scales with labels attached to each point on the scale (“Strongly agree,” “Somewhat agree,” “Neither agree or disagree,” “Somewhat disagree,” “Strongly disagree”) and some unbalanced 5-point scales, also with labels attached to each point on the scale (“Extremely interested,” “Very interested,” “Somewhat interested,” “Not that interested,” “Not at all interested”). In addition, 4 open-ended questions were included that asked respondents to name or identify their answer; for these questions, precoded lists of answer categories were prepared and additional verbatim responses were later coded during data analysis.
The survey was reviewed by the US Office of Human Subjects Protection and classified as exempt from institutional review board review/approval. After the survey was approved by the US Office of Management and Budget, interviews and preliminary analyses were conducted under contract by Field Research, Inc., through Ogilvy Public Relations Worldwide.
Data were reviewed at the end of the survey and discrepancies were resolved and recoded. All analyses were conducted using SPSS 11.0.1 (SPSS, Inc., Chicago, IL) and Stata 7.0 (Statacorp., College Station, TX). To test for differences between sample populations, χ2 goodness-of-fit tests were performed. To identify predictors of survey responses among the populations, bivariate and multivariate logistic regression analyses were conducted.
Our interviewers attempted to contact a total of 12,605 individuals for the general population sample and 31,701 for the African American, Hispanic, and the MSM groups. Of these, we completed 2008 interviews from the general population, 501 from the African American and Hispanic, and 500 from the MSM group. The cooperation rates were 77.1% for the general population, 73.8% for African Americans, 78.7% for Hispanics, and 91.1% for MSM. After accounting for unanswered and unscreened telephone calls, the adjusted response rate for each group was as follows: general population (29.1%), African American (33.1%), Hispanic (40.6%), and MSM (34.6%).
Table 1 lists sociodemographics for each sample of participants. Among the general population sample, the plurality were aged 35 to 54 years of age, female, had an income less than $50,000 per year, and had some college education or beyond. The racial breakdown of the respondents was 77.7% white, 8.4% Hispanic, and 8.1% African American. Compared with the 2000 US Census, the general population sample in our survey had an identical percentage of men and women, a slightly different racial distribution (higher percentages of whites, Native Americans, and other; lower percentages of blacks, Hispanics, and Asians), and attained somewhat higher levels of education. Over 17.0% of those in our general population survey did not report household income, resulting in lower percentages at all income levels, and making comparison with this characteristic and US census data difficult. In the African American sample, 38% were aged 18 to 34, 62% were female, 22% had income <$25,000 per year, and 47% had some high school or less or were a high school graduate. Compared with the African American population in the National census, our sample had fewer men and younger individuals and was less educated (data not shown). For the Hispanic sample, nearly 49% were 18 to 34 years of age, >56% were female, >54% had income <$25,000 per year, and >39.0% had some high school or less education. Compared with the general Hispanic population, our sample had fewer men and lower income (data not shown). For those responding to the MSM survey, 36.0% were between the ages of 35 and 44 years, >38.0% had an income of ≥$75,000 or more per year, and slightly more than 6.0% had either a high school education or less.
Table 2 shows responses and statistical comparisons of survey questions related to HIV/AIDS awareness, knowledge, and attitudes in the general population sample. Notably, 18.2% of the general public believed that a vaccine secretly exists, 23.6% correctly believed that vaccines being tested today cannot cause infection, and 28.8% would be supportive of someone they knew who was considering participating in an HIV vaccine study. Moreover, 15.4% cited HIV/AIDS as the most urgent current health problem, compared with 23.1% for cancer and 21.9% for health care access, costs, and insurance; 43.5% read or heard about HIV vaccine research in the past 12 months; and 29.8% were interested to learn more about HIV vaccine research. Despite these findings, 62.5% agreed that an HIV vaccine is the best hope to stop the global AIDS epidemic.
A subanalysis of the general population sample showed that, compared with men, women had less confidence in the development and effectiveness of an HIV vaccine, were more likely to incorrectly believe that the vaccines being used in clinical trials could cause HIV infection, did not trust the government to protect volunteers in HIV vaccine clinical trials, and overall had less hope that an HIV vaccine would control the global AIDS epidemic. Despite these findings, a statistically higher percentage of women were interested in learning more about HIV research efforts and cited HIV/AIDS as the most urgent current health problem. Interestingly, statistically fewer women had read or heard about HIV research efforts in the past 12 months.
Highly Impacted Populations
Within the highly impacted sample groups, 47.1% of African Americans, 26.5% of Hispanics, and only 13.4% of MSM believed that an HIV vaccine already exists and is being kept secret. These results were statistically significant in cross-comparisons between the highly impacted groups. Multivariate logistic regression analyses also showed the following among those who agreed or did not know whether an HIV vaccine is being kept secret: African Americans were less likely to be supportive of vaccine volunteers; Hispanics were less likely to believe it is important to personally support vaccine research; MSM were less likely to believe that vaccines tested do not cause infection but were more likely to personally want to learn more about HIV vaccine research.
However, only 4.6% and 1.0% of African Americans and Hispanics, respectively, listed distrust of the government among their top 3 concerns for not supporting someone who is considering volunteering for an HIV vaccine trial (data not shown). Moreover, an overwhelming 78.0% of Hispanics believed the US government can be trusted to protect trial volunteers, which was comparatively higher than African Americans and MSM.
Seventy-eight percent of African Americans, 57.7% of Hispanics, and 68.0% of MSM did not know or incorrectly believed that HIV vaccines being tested can cause HIV infection in volunteers. These responses were all statistically significant in a cross-comparison analysis. For each of these samples, the majority correctly believed (although to varying degrees) that HIV vaccines require extensive testing for approval, although to varying degrees. Multivariate logistic regression analyses also showed the following among those who disagreed or did not know that HIV vaccines being tested do not cause infection: African Americans were less likely to be supportive of vaccine trial volunteers, believe the government will protect trial volunteers, and believe a vaccine will be 100% efficacious; Hispanics were less likely to be supportive of vaccine trial volunteers, believe it is important to personally support vaccine research, agree that vaccines are our best hope, believe the government will protect volunteers, and believe a vaccine will be 100% efficacious; MSM were less likely to believe that it is possible to develop a vaccine, vaccines tested require extensive testing, believe the government will protect volunteers, and heard or read of vaccine research in the past 12 months.
With regard to support for HIV vaccine trials, statistically fewer African Americans (34.9%) would be supportive of an HIV vaccine trial volunteer, compared with Hispanics (45.9%) and MSM (67.8%). In contrast, the majority (86.0% or higher) within each of the highly impacted groups believed that it is important to support HIV vaccine research.
Moreover, a larger percentage of MSM (77.2%) cited HIV/AIDS as the most urgent current health problem, compared with Hispanics (10.6%) and African Americans (22.4%). Despite these findings, our survey found that statistically fewer African Americans (61.9%) than Hispanics (74.3%) or MSM (70.6%) believe that an HIV preventive vaccine is the best way. Yet, all of these groups overwhelmingly (>86.0%) believe that HIV vaccine research is just as important as HIV prevention education.
Of particular concern to those conducting clinical HIV vaccine research is the potential impact that negative attitudes and misinformation can have on trial participation, as either a deterrent to participation or supporting the participation of others. Whereas attitudes, beliefs, knowledge, and cultural characteristics have been shown to affect participation in clinical trials for cancer and Alzheimer disease research, there are limited data on factors influencing participation in studies of investigational vaccines or medicines.10,11 With many thousands of volunteers from diverse backgrounds needed to participate in HIV vaccine research in the United States alone, this study sought to examine American attitudes and beliefs about HIV vaccines.
We found that there is significant variation among the populations with respect to their awareness and knowledge of, as well as their support for, HIV vaccine research. Our results underscore the need for and uses of exploratory research to build effective community engagement programs that aim to educate and address the attitudes and perceptions of key communities affected by HIV/AIDS.
Additionally, we used the audience segmentation theory to further understand the results of this survey. In the audience segmentation theory, fundamentally distinct populations, grouped by common identifiable social characteristics, reside along a continuum of knowledge/awareness/supportiveness of a concept such as support of HIV vaccine trials. Along the continuum, populations are categorized as (A) opposed; (B) having negative attitudes; (C) being aware but likely unsupportive; (D) being unaware/uninformed; (E) aware/likely supportive; (F) having positive attitudes; or (G) being strongly supportive of volunteering (Fig. 1).12-15 Knowing and understanding the target audience(s) along this continuum is important to determine which key messages will aid in shifting the audience toward being more supportive of HIV vaccine research and trial participation. Overall, the results of the media analysis, focus groups, and the telephone survey confirmed that most Americans were in the unaware/uninformed group (category D) of the audience segmentation continuum, and thus the challenge is to implement a national education/awareness program to move individuals toward the supportive end of the continuum (E, F, and G) while maintaining the high level of support and knowledge in the MSM population (Fig. 1).
Across all groups, the majority of those surveyed believe HIV vaccine research is as important as education on HIV prevention and that it is important to personally help support HIV vaccine research. The majority were also optimistic that a preventive HIV vaccine can and will be developed. Most groups surveyed indicated some awareness of news items regarding HIV vaccine research and were knowledgeable that vaccines must be tested in thousands of persons before they can be approved for use. In addition, most of the groups were at least somewhat trustful of the government to protect volunteers (Hispanics were particularly trustful). However, in contrast to these favorable findings, only 21.6% to 42.3% of each group surveyed was aware that HIV vaccines cannot cause infection.
In comparison with the general population, African Americans, Hispanics, and MSM surveyed were more likely to be interested in learning about HIV vaccine research, were more supportive of others volunteering for trials, and were more likely to feel that it was important for them to personally help support HIV vaccine research.
Compared with all other groups surveyed, MSM were most likely to view HIV/AIDS as an urgent problem (77.2%) and were overwhelmingly more likely to be supportive of others volunteering for HIV vaccine trials (67.8%). MSM were also most likely to have seen news items on HIV vaccine research (73.8%), were most knowledgeable in regard to vaccines needing to be tested in thousands of persons (72.6%), and were most likely (66.8%) to express interest in learning more about the trials. MSM were least likely to have misinformation regarding a vaccine already existing (13.4%) but were somewhat less trusting than other groups that the government would protect volunteers. In summary, MSM were highly knowledgeable about the epidemic and HIV/AIDS research and had the highest interest and support for preventive HIV vaccines. Hence, these results verify the need to work with this population to maintain their knowledge/awareness and improve their trust so that they will continue to be supportive of HIV vaccine research.
In a similar comparison, Hispanics were more likely than all others surveyed to be trustful of the government to protect volunteers (78%), were most knowledgeable that vaccines cannot cause infection (42.3%), and were second only to MSM in supportiveness of someone volunteering (45.9%). However, Hispanics were also least likely to cite HIV/AIDS as the most urgent health problem (10.6%). In addition, despite the high level of trust, a significant portion of Hispanics believe a vaccine already exists. These findings emphasize the need for more education of Hispanics to remove misconceptions and increase the awareness of Hispanics about the HIV epidemic and how it affects their community.
African Americans were second only to MSM in viewing HIV/AIDS as an urgent problem (22.4%) but also were the group most likely to be misinformed that a vaccine already exists (47.1%), were least likely to be aware of news items on HIV vaccines (36.5%), and were unlikely to know that vaccines cannot cause infection (22.0%). There is a great need for more education to remove misconceptions about HIV vaccine research to increase the awareness and supportiveness of the African American community.
The need for outreach and education efforts is also strengthened by the fact that, whereas the general public and highly impacted populations overwhelmingly believe that an HIV preventive vaccine is the best hope to control global AIDS epidemic, they differed in their supportiveness of HIV vaccine trial volunteerism for someone they knew. These issues may make trial recruitment and retention more difficult.
The results of this survey provide interesting insights in the various populations surveyed. Limitations and potential biases should be considered, however. The major limitation is that the results of this survey may not be generalizable to the US population(s) as a whole. This is due to the fact that the highly impacted groups were sampled from preselected urban locations and each of the population samples had statistically significant differences compared with the data from the 2000 US Census. Despite this limitation, these study results are extremely relevant to the activities of the HVCC, which focuses on outreach to communities disproportionately affected by HIV/AIDS, with a special emphasis in urban areas where HIV vaccine clinical trials are ongoing or planned. Using the results of this survey will help to formulate the necessary educational efforts to help potential volunteers make a more informed decision about participation in HIV vaccine clinical trials. Also, although efforts were undertaken to reduce potential bias and misclassification, the results from the MSM survey may be influenced by nonresponder bias. Compared with the other study groups, the MSM sample had a lower response rate but did have a higher level of cooperation. To reduce potential bias and misclassification, the study design involved performing random telephone surveys and screening individuals through a presurvey questionnaire.
Taken together, our results outline the areas to address in a number of American communities. We believe these survey results validate the key messages of the HVCC (Fig. 2) and underscore the need to provide accurate information, correct misinformation, and promote favorable attitudes about HIV preventive vaccine research. In consultation with the HVCC Steering Group, NIAID will use the results of this study to continue to engage African Americans, Hispanics, MSM, and the general population in a national awareness and education program.
The provision of clear, accurate information to communities that have historically had lower levels of participation in clinical trials is essential to addressing community-specific barriers to support for HIV preventive vaccine trials. Participation in and support for HIV vaccine trials from target populations in which new HIV infections are rising is critical to finding an HIV vaccine that works for everyone. Building broad-based support for HIV vaccine research in the context of other research efforts in AIDS care and treatment will strengthen a comprehensive HIV/AIDS program. This research has allowed NIAID to build a stronger research program that can provide accurate and timely information to community members so they can make informed, educated decisions about the role they hold in the development of an HIV preventive vaccine.
The authors thank the following for their contributions to the manuscript, including Rona Siskind (NIH/NIAID), Benjamin Perkins (Henry M. Jackson Foundation), Martin Gutierrez (NIH/NIAID), Joanna Katzman (Henry M. Jackson Foundation), George Counts, MD (HVTN), Sarah Alexander (HVTN), and Steve Wakefield (HVTN).