Readiness of high-risk populations in the HIV Network for Prevention Trials to participate in HIV vaccine efficacy trials in the United States

Objective: To determine the willingness of populations at high risk of HIV-1 infection to participate in HIV vaccine efficacy trials, determine factors influencing decision-making, and evaluate knowledge levels of vaccine trial concepts.

Design: Cross-sectional study.

Methods: HIV-1-negative homosexual men, male and female injecting drug users and non-injecting women at heterosexual risk were recruited in eight cities in the United States (n = 4892).

Results: A substantial proportion of the study population (77%) would definitely (27%) or probably (50%) be willing to participate in a randomized vaccine efficacy trial. Increased willingness was associated with high-risk behaviors, lower education level, being uninsured or covered by public insurance, and not having been in a previous vaccine preparedness study. Altruism and a desire for protection from the vaccine were major motivators for participation. Major concerns included positive HIV-1 antibody test due to vaccine, safety of the vaccine, and possible problems with insurance or foreign travel. Baseline knowledge of vaccine trial concepts was low.

Conclusions: It is likely that high-risk volunteers will be willing to enroll in HIV vaccine efficacy trials. A variety of participant and community educational strategies are needed to address participant concerns, and to ensure understanding of key concepts prior to giving consent for participation.

Introduction

Certain subgroups of men who have sex with men (MSM), injecting drug users (IDU), and women remain at high risk of infection with HIV-1 [^1-3]. Continued research is needed to develop effective prevention strategies, including clinical trials of candidate HIV vaccines. Phase I and II studies on the safety and immunogenicity of various candidate vaccines are ongoing [^4,5], and eventually, promising vaccines may be tested for efficacy in preventing HIV-1 infection among populations who are at high risk.

In 1995, the Vaccine Preparedness Study (VPS) of the HIV Network for Prevention Trials (HIVNET) was initiated to gather baseline data to prepare for vaccine efficacy trials among several high-risk populations in the United States, and to establish the infrastructure to implement HIV vaccine and non-vaccine prevention trials rapidly [^6-9]. One of the goals of the VPS was to determine the readiness of several high-risk populations to participate in vaccine efficacy trials. To assess readiness, the answers to several important questions are required. First, are HIV-1-negative high-risk individuals, the target population for vaccine efficacy trials, willing to participate in HIV vaccine trials? Estimates of the proportion of high-risk populations willing to participate help determine the feasibility of conducting trials and the level of resources needed to conduct recruitment for a trial. Second, what is important to high-risk individuals in decision-making about participation? Analysis of factors influencing willingness can indicate which strategies need to be developed to address the concerns of potential participants. Finally, what is the level of knowledge regarding vaccines and vaccine trial concepts among high-risk individuals? This information can be used to develop educational strategies to improve understanding about vaccines and vaccine trials, which are critical components of the informed consent process. To address these questions, we present data from the VPS cohort of HIV-1- negative high-risk individuals.

Methods

From April to November 1995, individuals who were HIV-1-negative were recruited in eight cities in the United States: Boston (Massachusetts), Chicago (Illinois), Denver (Colorado), New York City (New York, two sites), Philadelphia (Pennsylvania), Pawtucket/Providence (Rhode Island), San Francisco (California), and Seattle (Washington). HIV-1-negative men were eligible if they reported anal sex with one or more men in the past year or drug injecting at least once in the previous 6 months. HIV-1-negative women were eligible if they reported drug injecting in the previous 6 months or were at high risk through heterosexual contact. Women were considered at high- risk through heterosexual contact if they reported any of the following in the past year: exchange of sex for money or drugs; crack use; five or more male partners; or diagnosis of syphilis, gonorrhea, chancroid, pelvic inflammatory disease, trichomonas or an initial episode of genital herpes. Women were also eligible if they had a current relationship with a man with any of the following characteristics: HIV-1-positive; had sex with other men; injected drugs in the past 5 years; or was diagnosed with syphilis, gonorrhea or chlamydia in the past year. Recruitment strategies varied by city but included advertising, referrals from other vaccine preparedness studies [^1,2,6], staff outreach, and referrals from current study participants, community agencies, HIV counseling sites and clinics for sexually transmitted diseases.

At the baseline visit, information on demographics and sexual and injecting risk behaviors was collected by trained interviewers using a standardized questionnaire. Participants were asked about sexual behaviors in the previous 6 months, including HIV-1 antibody status of the partners, and occurrence of specific sexual practices with each type of serostatus partner. Risk by injection was assessed by collecting data on occurrence of injection and sharing injection equipment.

Willingness to participate in future vaccine efficacy trials and knowledge of trial concepts were assessed at baseline by self-administered questionnaires. If necessary due to literacy problems, the questionnaires were administered by an interviewer, which occurred in 4.8% of visits. Questions on willingness and knowledge were asked without reference to a specific vaccine candidate or product.

Participants were first presented with 14 true/false statements about vaccine trial concepts and were asked whether they agreed or disagreed with the statements or were unsure of the correct answer. These data were collected to assess baseline knowledge levels before any educational efforts were initiated. The statements focused on concepts such as randomization, blinding, safety, adverse reactions, vaccine-induced seropositivity and potential social harms such as possible discrimination as a result of trial participation [¹⁰]. Next, a brief description of a randomized placebo-controlled vaccine efficacy trial was read to participants to provide information about the vaccine trial concepts included in the knowledge questionnaire (a copy of the description can be obtained from B.A.K.). After the trial description was read, participants had an opportunity to ask questions and were asked to rate their willingness to participate in a vaccine trial on a four-point scale (definitely, probably, probably not, definitely not willing). Finally, participants rated how important various factors were in making their decision about participating using a four-point scale (very, somewhat, slightly, not at all important).

Statistical analysis

Analyses of categorical variables were conducted using χ² tests of significance. Multivariate analysis of willingness by study population, demographics and risk behaviors was conducted using multiple logistic regression [¹¹]. Factors that may influence willingness to participate in future vaccine efficacy trials were grouped into four categories: social benefits, social risks, personal benefits, and personal risks. The association between each factor and willingness to participate in vaccine efficacy trials was determined using logistic regression, controlling for study population. Tests of hypotheses of equal odds ratios across study populations were conducted using the likelihood ratio test [¹¹]. Where significant differences were found across study populations, data are presented for each study population separately. Otherwise, data from study populations are pooled.

Results

Study population

A total of 4892 HIV-1-negative individuals were enrolled from April to November 1995: 3257 (66.6%) MSM, 770 (15.7%) male IDU, 511 (10.4%) women at heterosexual risk (WAHR), and 354 (7.2%) female IDU who also met eligibility criteria for heterosexual risk (WAHR-IDU). The demographic and behavioral characteristics of the four study populations are presented in Table 1. A larger proportion of MSM were white, college-educated, employed, covered by private health insurance and had participated in previous vaccine preparedness studies than the three other study populations. The populations recruited were at risk of HIV-1 infection based on recent behavior. One- quarter of MSM reported having a known HIV-1- positive sexual partner in the 6 months before enrolling in the study, as did 8% of women. Over one-third of MSM and 9% of women reported having unprotected receptive anal sex in the previous 6 months. Over three-quarters of women reported unprotected vaginal sex and over one-third of male and female IDU reported sharing injection equipment.

Table 1

Willingness to participate in HIV vaccine efficacy trials

Overall, 27% of participants indicated that they would be definitely willing to participate in an HIV vaccine efficacy trial, 50% were probably willing, 18% were probably not willing, and 5% were definitely not willing (Fig. 1). WAHR were significantly more likely to be definitely willing than the other three study populations (P < 0.001). No significant differences among study populations were observed when comparing those willing (definitely or probably willing) with those not willing (definitely or probably not willing; P = 0.12).

Fig. 1

Multivariate analysis of the association between willingness (definitely or probably willing versus definitely or probably not willing) and study population, demographics and risk behaviors are presented in Table 2. All odds ratios in Table 2 were adjusted by the other variables. Compared with participants with less than a high school education, those with a higher education level were less likely to be willing, with significant odds ratios for those with a high school education or at least a college degree. Those study subjects who had participated in previous vaccine preparedness studies were significantly less likely to be willing than new recruits. This relationship was even stronger among non-MSM populations as indicated by significant interaction terms between study population and previous study participation. Participants who were uninsured or received their health-care coverage through public insurance were significantly more likely to be willing than those covered by private insurance. Participants who reported having an HIV-1-positive partner and MSM who reported unprotected receptive anal sex were also significantly more likely to be willing than those who did not report these behaviors.

Table 2

Factors influencing willingness

The social benefit rated most frequently as very or somewhat important in decision-making about vaccine trial participation was 'helping to find a vaccine that works', followed closely by 'helping to stop the epidemic' (Table 3). For each social benefit, participants who considered the benefit to be important were significantly more likely to be willing to participate in future vaccine efficacy trials than those who did not rate the benefit as important (Table 3). The odds ratios for willingness in relation to each social benefit were not significantly different across study populations.

Table 3

The possibility of a positive HIV-1 antibody test result due to vaccine was the social risk most frequently rated as being very or somewhat important (Table 3). The next most frequently rated social risks were possible 'problems with health or life insurance' or 'problems with travel to foreign countries'. Participants who rated social risks as being important were significantly less likely to be willing to participate than those who did not rate these factors as important (Table 3). The odds ratios for willingness in relation to each social risk did not significantly differ by study population except for the possibility of a positive HIV-1 antibody test result. MSM who reported this social risk as important were significantly less likely to be willing than those not reporting this factor as important (odds ratio, 0.3; 95% confidence interval, 0.3-0.4). The odds ratios for the other study populations were not significantly different from 1.0. Overall, a larger percentage of participants indicated that social benefits were important factors in their willingness to participate than social risks.

The personal benefit rated most frequently as very or somewhat important was receiving 'current information about HIV research' (87% of participants), followed by feeling 'motivated to avoid risky behaviors' (75%), 'getting some protection against HIV infection from the vaccine' (74%), and getting 'free counseling and HIV tests' (70%; Table 4). Overall, 26% of participants indicated that a monetary incentive would be important in their decision. MSM were significantly less likely to rate each personal benefit as important than the other study populations, whereas WAHR were significantly more likely to rate each personal benefit as important, except for possible protection from the vaccine and monetary incentives. Although there were significant differences in the proportion of each study population rating the personal benefits as important, the rank order of the personal benefits was about the same across populations. All personal benefits, except monetary incentives, were significantly associated with willingness to participate in future vaccine trials, with the associations being strongest for male IDU, WAHR-IDU and WAHR.

Table 4

The personal risks most frequently rated as important were concerns that the 'vaccine could weaken the body's ability to fight off HIV infection' (72%) and the 'chance the vaccine could cause problems in the future' (71%; Table 4). Similar to personal benefits, MSM were significantly less likely to report personal risks as important than the other study populations. However, MSM who rated personal risks as important were significantly less likely to be willing than MSM who did not rate these items as important, except for the personal risk of thinking a lot about HIV/AIDS. Personal risks were not significantly associated with willingness among the other study populations.

Knowledge of vaccine trial concepts

Participants were asked to answer 'agree', 'disagree' or 'unsure' for the 14 knowledge items. Less than half of the items (mean, 6.5 across all participants) were identified correctly as true or false. A few items (mean, 2.9) were identified incorrectly, and for the remaining one-third of the statements (mean, 4.6), participants indicated that they were unsure. Only 8.4% of participants correctly answered at least 11 of the 14 questions.

Possible target areas for future educational efforts were suggested by those concepts frequently answered incorrectly or unsure. Table 5 shows 13 of the 14 items ordered according to the percentage incorrect. One item ('people may be discriminated against if they join an HIV vaccine study') was not included since there may have been some confusion about whether the question referred to the legality of discrimination or the potential for discrimination, whether legal or not. The concepts that the largest percentage of participants answered incorrectly concerned the possible adverse effect of the vaccine on the immune system (Table 5, item 1) and who would be eligible for a vaccine trial based on HIV-1 antibody status (item 2). Although the remaining items were incorrectly answered by 20% or less of the study population, at least one-third of the participants were unsure of the correct answer for five additional concepts relating to blinding (Table 5, items 4 and 11), vaccine-induced seropositivity (items 6 and 8), and vaccine efficacy (item 10).

Table 5

Discussion

The VPS of HIVNET provided an important opportunity to assess willingness to participate in future preventive HIV vaccine efficacy trials, determine factors influencing decision-making, and evaluate participant knowledge levels of vaccine trial concepts. Particular strengths of this study were the assessment of willingness across diverse high-risk populations with common instruments and the availability of data from non-IDU women at heterosexual risk, a population, to our knowledge, for which there are no data with regard to preparedness for future vaccine trials in the United States.

In this study, a large proportion of high-risk populations reported willingness to participate in future vaccine efficacy trials. Similar findings have been reported in previous studies of homosexual men, but a somewhat lower proportion who were willing to participate was reported in previous studies of IDU [^12-17]. In univariate analyses, women at heterosexual risk in VPS were somewhat more likely to be willing than homosexual men. However, once differences in demographic characteristics were considered in multivariate analyses, we found no significant differences by study population, suggesting that current recruitment strategies may be sufficient to recruit different risk populations who are willing. No differences in willingness were found by other demographic variables such as age, race/ethnicity or employment status.

Vaccine trials will require the participation of individuals at high risk of infection. In this study, participants reporting recent high-risk behaviors (an HIV-1-positive partner, or among MSM, unprotected receptive anal sex) were more likely to be willing to participate than those not reporting these behaviors, suggesting that it will be possible to recruit high-risk populations. Similar results have been reported from some cohorts of homosexual men [^13-15]. Analysis of HIV-1 incidence by willingness from longitudinal data from this study will provide more definitive evidence on the ability to recruit high-risk, willing populations.

Participants who were uninsured or insured by public medical coverage were more likely to be willing than those with private health insurance. It is possible that participants with public medical coverage or uninsured expect that vaccine trials will pay for any health problems that occur during the trial. Since monitoring safety of experimental vaccines will be an important objective of vaccine trials, it is likely that trials will cover expenses to evaluate possible adverse events. However, it is unclear at this time to what extent medical care associated with adverse events, whether or not vaccine-related, will be paid for during or after vaccine trials. The issue of medical care coverage will need to be resolved before the initiation of large-scale trials.

An alternative explanation for these findings is that participants with private insurance may be more reluctant to participate if they fear potential loss of insurance coverage due to association with a high-risk group or misinterpretation of vaccine-induced HIV-1 antibody-positive test results. The possibility of insurance problems will need to be discussed with study participants, and study sites will need to be prepared to help participants verify their HIV-1 infection status, if needed. Insurance-related issues will need to be clearly detailed in the informed consent documents.

Participants who had been in previous vaccine pre-paredness studies were less willing than new recruits. This reluctance to participate could be due to greater knowledge of candidate vaccines and possible risks associated with candidate vaccines acquired through previous study participation. However, with these cross-sectional baseline data, we cannot conclude that willingness decreases with increasing knowledge. Longitudinal data from the VPS will be available to assess whether changes in knowledge are related to changes in willingness.

Altruism was a main motivator for participating in efficacy trials across all study populations, and monetary incentives were important to only a small proportion of participants. However, of concern is the observation that three-quarters of participants were motivated by desire for possible protection from vaccine and this factor was highly associated with an increase in willingness. This problem could be compounded if participants were able to unblind themselves by using standard HIV-1 antibody testing and assume that if positive, they had received vaccine and were protected. Accordingly, careful counseling will be needed with frequent reinforcement before and during a trial to help participants understand that they should not assume they will be protected since they may receive placebo or a vaccine of unknown efficacy. Such counseling and education should also discourage testing outside of the study.

A major concern across all study populations was vaccine-induced seropositivity. Study sites will need to emphasize to participants that testing to discriminate actual HIV-1 infection from vaccine-induced sero-positivity will be available at the study site, as needed. However, participants need to understand that if they test outside of the study site, HIV-1 antibody-positive results found by conventional serologic tests may not be able to distinguish infection from vaccine-induced seropositivity, and these results may lead to possible discrimination or other social harms. Study sites must also be prepared to assist participants with verifying their HIV-1 infection status in situations of unavoidable, mandatory testing.

Safety of the vaccine was also an important concern for participants, particularly for IDU and women compared with homosexual men, although safety issues were more likely to affect willingness among homosexual men. Absolute safety cannot be guaranteed for any vaccine trial. However, information from prior HIV vaccine trials will contribute safety data that, together with information on uncertain risks, will need to be discussed with participants and the communities from which they are recruited. Informed consent documents will need to outline the possible safety risks.

Knowledge about vaccine trial concepts among study participants was assessed in this study before any formal education process was initiated. Hence, the low levels of knowledge found across all study populations should not necessarily be interpreted to mean that subjects who enroll in trials will not be well-informed. Attention to certain concepts will be especially important, including safety, blinding, vaccine-induced seropositivity and lack of knowledge about vaccine efficacy until an efficacy trial is completed. As important vaccine concepts, these findings underscore the need for a variety of approaches to educate about vaccine trials concepts in diverse populations. HIVNET is conducting further studies of educational approaches to ensure participant understanding of key concepts, an important prerequisite for proper informed consent. In addition, efforts are needed to educate the larger communities, since it is likely that participants will seek external advice about their decision to participate in vaccine trials [¹²].

The results of this study may be limited in terms of predicting actual patterns of willingness once recruitment for a vaccine efficacy trial begins. The hypothetical vaccine trial presented did not refer to any specific vaccine product or strategy and we do not know the actual proportion of these study populations who might be willing once presented with a specific vaccine product and trial. In addition, about one-third of participants were recruited from previous vaccine preparedness studies and their willingness may not reflect that of potential high-risk participants recruited through other methods that might be utilized for vaccine efficacy trials.

The VPS of HIVNET found that a large proportion of several high-risk populations were willing to participate in future HIV vaccine efficacy trials. Effective HIV prevention counseling to avoid increases in risk behaviors associated with trial participation will be critical. Participant counseling and community education need to address vaccine-induced seropositivity, safety, protection from the vaccine and other key vaccine trial concepts.

Acknowledgement

The authors acknowledge the contributions of the site coordinators and other staff at each site and the help and dedication of the study participants.