Twenty-six years into the HIV epidemic, the development of a safe and effective preventive vaccine remains an urgent priority. Promising vaccine candidates must be tested in clinical trials, requiring study volunteers to consider the risks and benefits of participation.1 In addition to the risk of physical adverse events, trials of preventive HIV vaccines pose unique social risks to volunteers, attributable in large part to the persistent stigma surrounding HIV/AIDS.2-4 When trial volunteers disclose their study participation, friends, family members, or coworkers may judge them negatively by assuming that they are in the study because they are HIV-infected or at high risk for infection. This may occur more frequently in efficacy trials conducted in low-prevalence settings, where study sites seek subpopulations at risk for HIV infection, such as men who have sex with men (MSM), commercial sex workers, and injection drug users-groups that have been stigmatized historically. HIV testing can also be a source of negative social impact (NSI), because many experimental vaccines elicit antibody responses that can be misinterpreted as being diagnostic for HIV infection on commercially available tests. Volunteers are discouraged from seeking HIV testing outside of the study site to avoid confusion about their true HIV infection status; to prevent unblinding as to treatment assignment; and to minimize potential discrimination from insurance companies, employers, or immigration officials.2-5
The specter of experiencing a NSI could be a significant disincentive for the thousands of volunteers who are required to evaluate the efficacy of HIV vaccine candidates.6,7 To date, the largest published analysis of NSIs in North American HIV vaccine trials evaluated 1516 volunteers enrolled in US National Institutes of Health-sponsored safety and immunogenicity studies conducted by the AIDS Vaccine Evaluation Group (AVEG) from 1995 to 1998. That study found that 5% of volunteers reported trial-related negative events, that occurrences of a personal nature predominated, and that HIV testing-related harms were less frequent but often had significant impact.8 In those phase 1 and 2 studies, no particular demographic subgroup (eg, race, education) reported higher rates of NSI. However, that study primarily evaluated trials enrolling volunteers at lower risk for infection, which may not reflect the experience of efficacy trial participants. A more recent phase 2 HIV vaccine trial enrolled 420 participants, 360 of whom were at high risk for sexual exposure to HIV.9 Overall, 27% of the volunteers in that study reported at least 1 NSI. It is unknown whether this higher prevalence of events can be attributed to the inclusion of volunteers with higher levels of reported sexual risk or to the more frequent schedule of NSI assessments (every 3 months) compared with the annual assessments performed in the AVEG trials.
In the first HIV vaccine efficacy trial conducted in North America and Europe, VaxGen, Inc. (Brisbane, CA) monitored NSIs during trial participation. Here, we report the frequency and nature of these events and the predictors of reporting them in more than 5000 volunteers at risk for sexually acquired HIV infection.
Study Design and Population
The Vax004 study was a multicenter placebo-controlled trial that randomized participants in a 2:1 ratio to AIDSVAX B/B vaccine (300 μg each of MN and GNE8 recombinant gp120/HIV-1) adsorbed onto 600 μg of alum or placebo (alum only). Study recruitment began in 1998, and over the following 17 months, 57 trial sites in the United States, 3 sites in Canada, and 1 site in The Netherlands accrued 5108 HIV-negative MSM and 309 women at sexual risk for HIV infection. Participants were educated about the study objectives and potential risks, including NSIs; those who met medical and behavioral eligibility criteria and provided informed consent were followed prospectively for 36 months. Blinded study injections were given intramuscularly at 0, 1, 6, 12, 18, 24, and 30 months, and HIV testing with pre- and posttest counseling was performed by trained counselors at each visit. All participating trial sites obtained institutional review board approval before study initiation. Details of screening procedures and enrollment criteria are described elsewhere.10,11
Demographic data were collected at enrollment. An interviewer-administered questionnaire was completed at the 6-month visit and semiannually thereafter to assess the social impact of trial participation. Ten categories of NSI were routinely assessed, including events involving personal relationships, unintentional disclosure, employment, life and disability insurance, health insurance, housing, medical care, government agencies, travel and immigration, or other types not specified. At those visits where volunteers report an NSI, staff members were instructed to administer an additional questionnaire to gather contextual information about the nature of the event, the perceived impact on quality of life, and relation, if any, to HIV antibody testing. These supplemental questionnaires were completed for situations that clinicians and volunteers agreed were probably or definitely related to trial participation. HIV risk behavior was assessed at baseline and every 6 months for the duration of the trial. Individuals who became HIV-infected underwent CD4+ T-lymphocyte and HIV viral load testing every 4 months over a 2-year period; social impact assessments were also completed at these visits.
For those volunteers who provided additional contextual information, staff narratives briefly describing the NSIs were coded to identify common themes using AnSWR software, version 6.4.x.12 An iterative coding approach was undertaken, whereby the first 3 authors began by independently reviewing participant responses to develop an analysis code book. All text was then independently coded in AnSWR by 2 investigators (MD, EML), and intercoder agreement was assessed. This assessment resulted in modifications to code definitions and development of a hierarchic coding scheme that was reflective of the themes and patterns in the coded text. After this process was completed, a separate investigator (JF) provided input and made final modifications to the coding scheme.
Constructed Site-Level Predictors
The size of the metropolitan area and local cumulative AIDS case rates were used as proxies for overall community awareness of HIV and HIV-related issues, which we hypothesized might be associated with reduced risk of NSIs. Clinical trial sites were classified as situated within low-, medium-, or high-population cities according to 2000 US Census estimates13 and official Canadian and Dutch population estimates for the same time period.14,15 Similarly, cumulative AIDS case rates were classified in terciles, based on US Centers for Disease Control and Prevention, Canadian, and Dutch surveillance data.16-19 In addition, we used the number of accrued participants as a proxy for site experience with evaluating and mitigating NSIs. Sites were classified as low (1 to 50 participants), medium (51 to 100 participants), or high (≥101 participants) enrolling sites. To minimize bias, site-level variables were categorized a priori, without reference to NSI rates at the site.
Predictors of reporting a NSI at each follow-up visit were examined using logistic models for repeated binary outcomes, with robust standard errors obtained using generalized estimating equations (GEE) to account for the within-subject correlation of these outcomes.20 Visit was included as a categoric variable in single-predictor and multipredictor models. We used a χ2 test to assess trend in the frequency of reporting NSI events over the course of the trial. Sociodemographic variables, including age at enrollment, gender, race/ethnicity, and education; markers of HIV risk in the prior 6 months (eg, multiple partners, having an HIV-positive partner); and site variables were incorporated into a final multipredictor model if significant on univariate analyses. Because we hypothesized that correlates of NSIs might differ among MSM and high-risk women, separate models were estimated for each group; an analysis of the pooled data would primarily reflect results for MSM, who comprised 94% of the sample. All quantitative analyses were done using SAS version 6.12 software (SAS Institute, Cary, NC).
Among the 5417 participants enrolled, 5108 (94%) were MSM and 309 (6%) were women at sexual risk for infection. As seen in Table 1, male volunteers were predominantly white (86%), and most had completed college or graduate school (64%); female volunteers were predominantly African American (55%), and one third (32%) had not completed high school. Men and women reported substantial levels of sexual risk at baseline: 44% of men and 42% of women reported penetrative sex (anal or vaginal) with an HIV-positive male partner in the past 6 months. Women were more likely to be enrolled at sites located in cities with populations >750,000.
Social Risk of Efficacy Trial Participation
Overall, 973 volunteers (18%) reported 1496 NSI events during 36 months of follow-up, and 146 volunteers (3%) reported 2 or more events of different types (Table 2). A statistically significant decline was observed in the number of events reported after the first 6-month assessment (P for trend < 0.0001). Negative reactions from friends, family, and partners accounted for most trial-related social impacts, with 14% of the cohort reporting such events. These instances were primarily attributed to a misunderstanding of the volunteer's HIV status or risk of infection. Unintentional or undesired disclosure of vaccine participation was the second most commonly reported NSI, noted by 3.2% of the cohort. Fewer than 1% of the cohort reported each of the problems with disability or life insurance, employment, medical or dental care, health insurance, government agencies, or housing. No travel- or immigration-related events were reported, although some sites discouraged persons with potential immigration issues from enrolling in the study.
Negative Social Impacts Among HIV-Infected Volunteers
Among the 368 individuals who became HIV-infected after enrollment, 351 (95%) attended post-infection follow-up visits. Twelve (3.4%) of these volunteers reported a NSI during this period. Ten volunteers reported issues involving personal relationships, such as friends or family members asserting that the vaccine itself caused the volunteer to become HIV-infected or more susceptible to infection. One volunteer was denied life insurance because of a positive HIV test result obtained after infection but attributed this to his trial participation. Another volunteer noted that he was ineligible to participate in an unrelated clinical trial because of his involvement in the vaccine study.
Data were available for 169 (17%) of the 973 volunteers experiencing a NSI. Demographic and risk variables for this group did not differ from those in volunteers who reported a NSI but did not have contextual reports completed (data not shown). Most of these NSIs involved personal relationships and were considered of minimal or moderate intensity to these volunteers (Table 3). Systematic content analysis of interpersonal events revealed 2 major categories. The first involved others believing the volunteer was at risk of becoming HIV-infected by the vaccine or that there was some product-related safety issue. The second involved friends or family members mistakenly believing that they themselves were at risk of contracting HIV if they were exposed to volunteers who received virus-containing vaccine. These misperceptions led to avoidance of sexual activity with volunteers and resulted in the dissolution of relationships. In 3 cases, MSM volunteers reported their partners were verbally or physically abusive.
Unwanted disclosure of trial participation was reported by 46 individuals and involved different types of NSIs. Events were attributed to others seeing the volunteer at the clinic and family members finding volunteers' study materials at home. In a single case, a staff member inadvertently exposed a volunteer's participation in a public setting, and in another case, a close contact discovered a volunteer's vaccine study identification card.
Although NSIs related to employment and insurance seldom occurred, these events were often described by volunteers as moderately upsetting or were considered major disturbances in their lives. Two volunteers perceived that their employment was terminated because they had requested time off for study visits. Twenty volunteers reported that they believed their disclosure of study participation resulted in delay or denial of health or life insurance. Actual vaccine-induced antibody results did not prompt any of these insurance-related events. Volunteers also noted that they lost opportunities to take full time jobs because of the visit schedule or to join the military or apply for insurance because of the HIV testing requirement.
None of the HIV antibody-related NSI events reported by 29 volunteers were attributed to problems directly related to vaccine-induced antibody results. In addition to the insurance-related examples described previously, 1 volunteer stated that his HIV-positive partner became anxious as they both waited for his test results. Two volunteers pursued HIV testing outside of the study in response to high-risk sexual exposures and were told that they were HIV-positive. Both reported anxiety associated with the repeat testing required at the study sites. In both cases, additional testing confirmed they were HIV-infected.
Predictors of Negative Social Impact Reporting
Univariate and multipredictor models were run on the entire cohort of men (Table 4), and univariate models were run for the female volunteers. Among MSM, younger age and follow-up visit at an early time point after enrollment were associated with a greater likelihood of reporting a NSI, whereas level of education and race/ethnicity were not significant factors on multipredictor analysis. Reporting more than 1 partner in the prior 6 months was significantly associated with increased NSI risk, but there was no evidence for further increases in NSI reporting among participants with more than 2 partners. In addition, MSM enrolled at sites in cities with high cumulative AIDS case rates and in those enrolling fewer participants had a greater likelihood of reporting a NSI during follow-up.
In univariate analysis, only 2 factors were associated with NSI reporting among female volunteers: enrollment at a site with fewer than 50 participants (odds ratio [OR] = 1.65, 95% confidence interval [CI]: 1.31 to 2.07) and visit at an early time point (OR = 0.41, 95% CI: 0.20 to 0.86 for the 18-month vs. 6-month visit; OR = 0.15, 95% CI: 0.05 to 0.48 for the 30-month vs. 6-month visit; and OR = 0.37, 95% CI: 0.15 to 0.87 for the 36-month vs. 6-month visit). No other demographic, risk, or site variables were significant in univariate analysis (data not shown). Because of the relatively few women enrolled in this trial, we had limited power to identify relevant predictors in multipredictor models.
One in 5 volunteers noted a NSI during 36 months of follow-up in the first HIV vaccine efficacy trial, a prevalence consistent with earlier studies of mostly high-risk phase 2 trial volunteers but somewhat higher than trials enrolling lower risk volunteers.2,8,9 We found that volunteers having 2 or more sexual partners in the past 6 months were more likely to report NSIs than volunteers with 1 or no partner. We did not find an association between having a known HIV-positive partner and NSI reporting, however, suggesting that the number of partners rather than sexual risk per se may account for the increased NSI reporting. This is supported by our finding that most NSI events centered on problems with personal contacts who misunderstood basic trial requirements and the nature of the vaccines tested. As seen with previous studies, NSI events typically occurred early after enrollment and insurance- and employment-related issues were infrequent. Prior studies did not report on NSIs among HIV-infected vaccine trial volunteers; fortunately, this subgroup did not seem to experience a disproportionate number of NSI events.
The misperception that HIV vaccines can actually cause infection limits enrollment6 and is also a significant source of NSIs for volunteers.2,8 To address the public's limited knowledge of HIV vaccines,21,22 study sponsors should promote educational efforts for the community stressing that neither study participants nor their sexual partners can become infected from HIV vaccines. Study staff should highlight the most common areas of NSI in counseling study volunteers and create mechanisms to educate sexual partners, where needed.
Our trial and others2,8,9,23 have found that most NSIs arise from difficulty in interpersonal relationships and undesired disclosure of study participation. Trial staff should emphasize with volunteers the need for discretion in choosing with whom they discuss their participation and can role play different disclosure scenarios to reduce the risk of misunderstandings and negative consequences. Trial sites should reinforce with staff and volunteers how to preserve confidentiality, including proper handling of participant identification cards and materials. Given our findings and those of others24 of the association of young age with NSI reporting, particular care may be required in working with younger volunteers, including adolescents, a target population for future trials.25 This group may be more susceptible than older volunteers to peer and parental influence regarding trial participation.26
We hypothesized that community- and site-level factors could influence the likelihood of NSI reporting. In multipredictor analysis, we found that male volunteers were at greater risk for reporting NSIs if they enrolled at trial sites in cities with high cumulative AIDS case rates, independent of city size. This contradicted our original hypothesis that volunteers residing in such epicenters would be protected from NSIs, assuming their personal contacts would be more familiar with HIV-negative and AIDS-related issues. It is possible that higher disease burden in these communities may have paradoxically reassured volunteers that they could be forthcoming about their trial participation, yet they still faced negative reactions from misinformed friends and coworkers. Thus, staff from these settings should not presume that living in a city heavily affected by HIV/AIDS translates to greater overall community knowledge of HIV vaccines. We also found that volunteers enrolled at sites with fewer participants were more likely to report a NSI. Staff members from these sites may be less experienced in preventing NSIs among volunteers who may be at risk or more likely to ascribe impacts to the trial itself. Alternatively, volunteers at sites with fewer volunteers may more readily disclose NSIs to staff members who have more time to explore these issues with them. Although these proxy measures may not fully explain the relation between the site of enrollment and NSI reporting, staff training should focus on NSI prevention and mitigation, because future efficacy trials are likely to employ significantly fewer sites than Vax004 with the capacity to enroll large numbers of subjects from populations with sufficient seroincidence.27
Significantly fewer NSIs were related to employment or insurance; however, the perceived impact of these events on the volunteers' quality of life was substantial in many cases. To address job-related NSIs, investigators should design trials that limit the frequency and length of study visits to reduce volunteer time away from work or increase access to early morning, later evening, or weekend appointments. Most of the life and disability insurance-related NSIs were linked to the requirement for HIV testing common among most US carriers.28 Volunteers should be informed that investigators can intervene on their behalf by providing HIV status information as needed to prospective insurers. Nevertheless, it should be noted that companies are not legally compelled to use the testing facilities of the vaccine trial or to accept test results.29 As a result, broad constituencies of researchers, trial sponsors, and community groups must continue to educate the insurance industry and advocate for enhanced legal solutions, because federal disability discrimination law may not afford such protections.30
We found that only a limited number of NSIs occurred from HIV testing performed outside of the study; however, this may occur more frequently in the future as “opt-out” testing practices are incorporated into clinical practice.31 Unlike the gp120 vaccine tested in this study, vaccine candidates in current trials incorporate more HIV antigens, which can increase the likelihood of vaccine-induced antibody responses on screening and confirmatory tests.32 Targeted education to health care providers and other staff performing HIV testing should address the possibility of vaccine-induced antibody responses, and the possibility for “false-positive” HIV test results should be highlighted. New HIV testing strategies are also needed to differentiate vaccine-induced antibody responses from true infection.33
A number of study limitations should be acknowledged. First, the relatively few women enrolled in the Vax004 study limited power to identify predictors of NSI reporting, if they existed. Second, contextual data were collected for only a subset of NSI reporters based on trial staff members' perceived severity and/or relatedness of the events to trial participation. Therefore, these contextual data may not be generalizable to all NSI reports. Routinely collecting such information, including the methods used and time to resolve NSIs, should be helpful in future studies. Third, all NSI data were collected by interviewers; systems to collect NSI data anonymously, such as audio computer-assisted self-interviewing, may lead to increased reporting. Finally, our findings may not reflect the experiences of all types of trial volunteers or the wide range of international settings where efficacy trials are likely to take place.34
These data from the world's first completed HIV vaccine efficacy trial should inform collection of NSI data in future trials, particularly those enrolling new vulnerable populations. Staff members should be familiar with local norms that influence if and how volunteers confer with family and friends about their study participation and be knowledgeable about national legal requirements governing HIV testing. These and other steps should support culturally appropriate efforts to protect the well-being of future vaccine efficacy trial volunteers.
The authors thank Mary Allen for her input in the preparation of this manuscript and appreciate the contributions of Marta Ackers, Brad Bartholow, Alan Greenberg, Tim Mastro, and the Vax004/VISION study team. The authors also gratefully acknowledge the Vax004 study volunteers and participating community advisory boards.
2. Sheon AR, Wagner L, McElrath MJ, et al. Preventing discrimination against volunteers in prophylactic HIV vaccine trials: lessons from a phase II trial. J Acquir Immune Defic Syndr Hum Retrovirol
3. Chesney MA, Lurie P, Coates TJ. Strategies for addressing the social and behavioral challenges of prophylactic HIV vaccine trials. J Acquir Immune Defic Syndr Hum Retrovirol
4. Frey SE. Unique risks to volunteers in HIV vaccine trials. J Investig Med
. 2003;51(Suppl 1):S18-S20.
5. Koblin BA, Heagerty P, Sheon A, et al. Readiness of high-risk populations in the HIV Network for Prevention Trials to participate in HIV vaccine efficacy trials in the United States. AIDS
6. Buchbinder SP, Metch B, Holte SE, et al. Determinants of enrollment in a preventive HIV vaccine trial
: hypothetical versus actual willingness and barriers to participation. J Acquir Immune Defic Syndr
7. Mills E, Cooper C, Guyatt G, et al. Barriers to participating in an HIV vaccine trial
: a systematic review. AIDS
8. Allen M, Israel H, Rybczyk K, et al. Trial-related discrimination
in HIV vaccine clinical trials. AIDS Res Hum Retroviruses
9. Belshe RB, Stevens C, Gorse GJ, et al. Safety and immunogenicity of a canarypox-vectored human immunodeficiency virus Type 1 vaccine with or without gp120: a phase 2 study in higher- and lower-risk volunteers. J Infect Dis
10. Harro CD, Judson FN, Gorse GJ, et al. Recruitment and baseline epidemiologic profile of participants in the first phase 3 HIV vaccine efficacy trial. J Acquir Immune Defic Syndr
11. Flynn NM, Forthal DN, Harro CD, et al. Placebo-controlled phase 3 trial of a recombinant glycoprotein 120 vaccine to prevent HIV-1 infection. J Infect Dis
12. AnSWR: Analysis Software for World-based Records
[computer program]. Version 6.4. Atlanta, Georgia: Centers for Disease Control and Prevention; 2004. Available at: http://www.cdc.gov/hiv/software/answr.htm
. Accessed March 25, 2006.
16. Smit C, Geskus R, Uitenbroek D, et al. Declining AIDS mortality in Amsterdam: contributions of declining HIV incidence and effective therapy. Epidemiology
17. MacDougall RG, Rekart ML, Knowles L, et al. HIV/AIDS Update: Year End 2000
. Vancouver, BC, Canada: BC Centre for Disease Control; 2000.
18. Remis RS, Swantee C, Rottensten K, et al. Report on HIV/AIDS in Ontario
. Toronto, Ontario, Canada: Ontario HIV Epidemiologic Monitoring Unit, University of Toronto; 2003.
19. Santé et Services sociaux Quebec. Surveillance des cas de syndrome d'immunodeficience acquise (SIDA) cas cumulatifs 1979-2003
. Montréal, Québec, Canada: Programme de surveillance du SIDA du Quebec; 2003.
20. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics
21. Allen MA, Liang TS, La Salvia T, et al. Assessing the attitudes, knowledge, and awareness of HIV vaccine research among adults in the United States. J Acquir Immune Defic Syndr
22. Roberts KJ, Newman PA, Duan N, et al. HIV vaccine knowledge and beliefs among communities at elevated risk: conspiracies, questions and confusion. J Natl Med Assoc
23. Jenkins RA, Thapinta D, Morgan PA, et al. Behavioral and social issues among volunteers in a preventive HIV vaccine trial
in Thailand. J Acquir Immune Defic Syndr
24. Allen MA, Metch B, Lau C, et al. Negative social impacts in preventive HIV vaccine clinical trials [abstract OA06-02]. Presented at: AIDS Vaccine 2006; 2006; Amsterdam.
25. Gender, age, and ethnicity in HIV vaccine-related research and clinical trials: report from a WHO-UNAIDS consultation, 26-28 August 2004, Lausanne, Switzerland. AIDS
26. Swartz L, Kagee A, Kafaar Z, et al. Social and behavioral aspects of child and adolescent participation in HIV vaccine trials. J Int Assoc Physicians AIDS Care
27. Fuchs JD, Buchbinder SP. Lessons from the AIDSVAX B/B vaccine efficacy trial. In: Koff WC, Kahn P, Gust ID, eds. AIDS Vaccine Development. Challenges and Opportunities
. Norfolk, VA: Caister Academic Press; 2007:105-110.
28. Rothstein MA. Genetics and Life Insurance: Medical Underwriting and Social Policy
. 1st ed. Cambridge, MA: The MIT Press; 2004.
29. de Bruyn T. HIV/AIDS and Discrimination: A Discussion Paper
. Montreal, Quebec, Canada: Canadian HIV/AIDS Legal Network and Canadian AIDS Society; 1998.
30. Leider PA. Domestic AIDS vaccine trials: addressing the potential for social harm
to the subjects of human experiments. Calif Law Rev
31. Branson BM, Handsfield HH, Lampe MA, et al. Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings. MMWR Recomm Rep
. 2006;55(RR-14):1-17 [quiz: CE11-CE14].
32. Ackers ML, Parekh B, Evans TG, et al. Human immunodeficiency virus (HIV) seropositivity among uninfected HIV vaccine recipients. J Infect Dis
33. Khurana S, Needham J, Mathieson B, et al. Human immunodeficiency virus (HIV) vaccine trials: a novel assay for differential diagnosis of HIV infections in the face of vaccine-generated antibodies. J Virol
34. Pitisuttithum P, Gilbert P, Gurwith M, et al. Randomized, double-blind, placebo-controlled efficacy trial of a bivalent recombinant glycoprotein 120 HIV-1 vaccine among injection drug users in Bangkok, Thailand. J Infect Dis