• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Will Sexual Risk Behaviour Increase After Being Vaccinated for AIDS?

Posted on: Wednesday, 29 March 2006, 09:01 CST

By Crosby, Richard A; Holtgrave, David R

Summary: The question of whether people will engage in greater levels of HIV-associated risk behaviour after receiving an AIDS vaccine has not been sufficiently investigated. Three objectives were: (1) assess the likelihood that people will engage in greater levels of HIV-risk behaviour after receiving an AIDS vaccine; (2) determine the association of increases with vaccination intent; and (3) identify differences between people who would and would not increase their HIV-risk behaviour after vaccination.

A cross-sectional survey was conducted of 278 adults from three populations: gay men, African-American women, and persons who used illicit drugs.

Nearly one-quarter of the sample indicated a likelihood that their HIV-risk behaviour would increase after vaccination. This increase was positively associated (r = 0.24) with increased intent to be vaccinated. Previous worry about having HIV (adjusted odds ratio [AOR] = 3.4, P = 0.004), being 32 years of age or older (AOR = 2.9, P = 0.0007), and having less than a high school education (AOR = 2.3, P = 0.027) were each associated with a post-vaccination increase in HIV risk.

With the seemingly real potential for increased HIV-risk behaviours after being vaccinated against AIDS, intervention studies are warranted to identify strategies with potential to minimize this phenomenon.

Keywords: AIDS vaccine, sexual risk, behaviour, HIV

Introduction

Although the advent of an AIDS vaccine may represent a substantial turning point in the epidemic, the question of whether people will engage in greater levels of HIV-associated risk behaviour after receiving an AIDS vaccine has not been sufficiently investigated. In clinical trials of vaccine candidates, increased risk behaviour among vaccinated volunteers might be considered an adverse event by an internal review board, thereby precluding ongoing study of this phenomenon. Indeed, the same dynamic could undermine the success of an AIDS vaccine. Clearly, even modest increases in risk behaviours that follow from a perception of 'safety' could substantially detract from the overall value of the widespread use of a vaccine that is not completely protective.1'2

Given the ethical improprieties of passively observing increases in HIV-associated risk behaviour among persons enrolled in clinical trials designed to test efficacy, an alternative strategy is to ask people to respond to a hypothetical scenario. While admittedly exploratory in nature, such a study could be quite informative as progress is made toward the development of effective and PDA- approved AIDS vaccine. Accordingly, the purpose of this study was to (1) assess the likelihood that people would increase their level of HIV-risk behaviour if they received an AIDS vaccine; (2) determine whether people, who would increase their level of HIV risk taking, are also more likely to state an intent to get vaccinated; and (3) identify basic demographic and psychosocial differences between people who would and would not increase their HIV-risk behaviour if they received an AIDS vaccine.

Methods

Study sample

From October 2002 through April 2003, project staff members approached members of three populations: men attending gay- identified venues, AfricanAmerican women attending a wide variety of venues (e.g., small neighbourhood grocery stores, fast food restaurants in low-income downtown areas, homeless shelters, and a university campus), and persons who abuse substances and attend venues that provide recovery services. Recruitment occurred at arbitrary times and days of each week. The recruitment area was limited to a large metropolitan area of the southeastern US. Persons were eligible for study participation if they were at least 18 years of age, could read and speak English, and indicated that (to the best of their knowledge) they were not HIV-positive.

Three hundred and fifty-one people were approached and asked to participate in the study. Of these, 278 (79%) answered questions designed to assess eligibility. All of those who agreed to answer questions about eligibility were eligible and subsequently provided written informed consent. A $25 incentive for participation was provided. The Emory University Institutional Review Board approved the study protocol prior to study implementation.

Data collection

Participants completed a 40-item self-administered questionnaire in a private or semi-private area. Research assistants made themselves available to participants asking for clarification in wording.

Measurement

Assessment of the outcome measure Participants were asked, Imagine that you get a vaccine that would protect your immune system if you ever were infected by HIV. How likely is it that you would engage in greater-levels of HIV-risk behaviour?' Response alternatives were provided on a five-point scale ranging from (1) Very likely' to (5) Very unlikely'.

Assessment of intent to be vaccinated Intent to accept a hypothetical AIDS vaccine was assessed by a single-item: 'If an AIDS vaccine (approved by the Food and Drug Administration) was made available in the next 12 months, how likely is it that you would get this vaccine?' Response alternatives were provided on the same five- point scale previously described. Of note, we have previously published findings, from this study, that specifically addresses correlates of intent to be vaccinated for AIDS if an AIDS vaccine became available.3

Assessment of demographic and psychosocial correlates Initial questions assessed basic demographic measures (i.e., age, sex, race, education, employment status). Next, substance use was assessed (using a 12-month recall period) by asking participants if they had injected illicit drugs or used crack, cocaine, heroin, ecstasy, crystal meth, LSD, 'Special K', or 'Roofies'. The same recall period was also used to assess the number of same-sex and opposite-sex sex partners of each participant. Questions also assessed whether participants had ever been diagnosed as having a sexually transmitted disease and how many times they had been tested for HIV infection. Finally, participants were asked how often they worry about having HIV (responses were provided on a four-point scale ranging from (1) 'never' to (4) 'frequently'.

Data analysis

Because the outcome measure was non-normally distributed, we dichotomized this distribution by grouping responses of 'very likely' and 'likely' and comparing these to the remainder of responses. Of the three correlates assessed on a continuous scale, none were normally distributed. Subsequently, age was dichotomized by a median split and the number of HIV tests was dichotomized by comparing those who had repeated tests (two or more tests were reported by 72%) to those who have never been tested or those testing only one time (28%). Those who indicated that they 'frequently' worry about having HIV (10.1%) were compared to the remainder who indicated less worry. Finally, it should be noted that we collapsed racial categories to compare African-Americans (a population in the US that is disproportionately at-risk of HIV infection4'5) to the remainder of the sample.

The third purpose of the study (i.e., to identify basic demographic and psychosocial differences...) necessitated extensive analysis. First, continuous variables were assessed for normality by calculating their degree of skewness and kurtosis. Skewness or kurtosis ratios exceeding an absolute value of 2.0 were considered an indication of nonnormality. Continuous variables with substantial deviations from normality were dichotomized. Next, bivariate significance was assessed by contingency table analyses. Strength of the bivariate associations was measured by calculating prevalence ratios, their 95% confidence intervals, and corresponding P values. Finally, correlates achieving a screening level of bivariate significance (P < 0.10) with the outcome (i.e., increased sexual risk behaviour) were entered into a forward stepwise logistic regression model to calculate adjusted odds ratios (AOR), their 95% confidence intervals, and corresponding P-values. The Hosmer^emshow Goodness-of-Fit Test was done to ensure that the model achieved a satisfactory fit with the data. Multivariate significance was defined by an alpha of 0.05.

Results

Characteristics of the sample

Average age of the sample was 33.1 years (standard deviation [SD] = ILl). About three-fifths of the sample was female. The majority (81.8%) identified as African-American/Black, 12.4% identified as White, and the remainder identified as members of other races. About one-sixth (16.0%) reported they had not completed high school. More than one-third (37.2%) reported they were not employed. Substance abuse was reported by 22.7% of the sample. About one-third (34.4%) reported ever being diagnosed with an STD. Nearly one-quarter (24.1%) of the sample was comprised of men having sex only with men.

Responses to the outcome measure

Participants were asked whether their HIV-risk behaviour would increase if they were vaccinated against AIDS. Fourteen percent stated it is 'very likely' that their HIV-risk behaviour would increase and another 9.4% stated it was likely'. Another 14% said they were 'unsure', 28.4% stated they were 'unlikely' to increase their risk behaviour, and the remainder (31.7%) stated that were 'very unlikely' to increase this risk behaviour. \After coding the response of 'very likely' as 1, 'likely' as 2, 'unsure' as 3, and so on, the mean score on this measure was 3.56 (SD = 1.4), suggesting the possibility that an increase in HIV risk would be relatively modest when averaged across a diverse population of people.

Relation of the outcome measure to intent to be vaccinated

Intent to be vaccinated if an PDA-approved vaccine became available was mixed. Thirty percent indicated they would be 'extremely likely' to accept such a vaccine, followed by 25.3% who said they would be 'likely', 24.9% who were unsure, 12.8% who said they would be 'unlikely' and 7.0% indicating they would be 'extremely unlikely'. This measure was significantly (P < 0.01) and positively correlated with the measure assessing whether HIV-risk behaviours would increase (Spearman's Rho = 0.23). Thus, as likelihood of greater HIV-risk behaviour increased, the likelihood of reporting an affirmative intent to be vaccinated also increased.

Bivariate associations

Table 1 displays the obtained bivariate associations between the outcome measure and each of the assessed correlates. As shown, six of the 11 assessed correlates achieved at least a screening level of significance (five of these six were less than 0.05).

Multivariate associations

Table 2 displays the AOR, their 95% confidence intervals, and respective P-values for variables that achieved significance (P < 0.05). in the multivariate model. The model was significant (χ^sup 2^ with 3 df = 29.5, P < 0.0001), and achieved a satisfactory fit with the data (Goodness-of-Fit χ^sup 2^ with 3 df = 2.15, P = 0.54). The model classified 77% of cases correctly.

The strongest multivariate correlate of increased HIV-risk after being vaccinated was reporting frequent past worry about having HIV. In contrast to those reporting less past worry, participants reporting frequent worry were 3.4 times more likely to report they would increase their level of HIV-risk behaviour after being vaccinated. Also, compared to their younger counterparts, older participants were nearly three times as likely to report a likely increase. Finally, those with less than a high school education were 2.3 times more likely than their counterparts earning a high school diploma to report a likely increase.

Discussion

Although exploratory, findings from this diverse sample suggest that a substantial percent of people being vaccinated against AIDS may increase their level of sexual risk. The extent of this increase and whether post-vaccination counselling can effectively avert this problematic response warrants further investigation. Also noteworthy is the finding that people most likely to report a post-vaccination response of increasing HIV-risk behaviour are also more likely to seek and receive an AIDS vaccine. Given that this observation may be supported by more definitive studies, one implication is that an important objective of prevaccination counselling protocols may be to help people understand that the vaccine should be viewed as merely one of several protective measures that could prevent AIDS. Further, people being vaccinated should understand that HIV infection is not averted by vaccines designed to prevent immune decline (i.e., to prevent AIDS).

The diverse nature of the sample-enabled comparisons between variously defined groups of people with respect to their reported intent to increase HIV-risk behaviour after receiving an AIDS vaccine. Three of these comparisons achieved multivariate significance. First, people who reported they had frequently worried about having HIV were clearly more likely to report a likely increase in HIV risk after being vaccinated. This may reflect a 'relief experience' conferred by the vaccine. To the extent this relief from worry acts as precursor of increased HIV-risk behaviour, the finding warrants further research. The finding suggests that an AIDS vaccine could conceivably produce a 'backlash effect' that - on a population-wide basis - dilutes the impact of the vaccine on AIDS incidence, unless the eventual vaccine is highly effective.

Multivariate findings also suggest that people older than 32 and those with less than a high school diploma may be more likely to increase their level of HIV risk after being vaccinated against AIDS. In all likelihood, these two demographic measures are probably only proxies of one or more behavioural predispositions that were not assessed in our exploratory study.

Table 1 Bivariate associations between the outcome measure and the assessed correlates

Table 2 Significant multivariate associations between correlates achieving screening significance and indications that HIV-risk behavior would increase after being vaccinated (n=266)

Many of the bivariate findings can also be quite informative. For example, gay men did not appreciably differ from the remainder of the sample with respect to the outcome of interest (the fact that the effect size was extremely small suggests the validity of this observed null association). The same observation applies to a comparison of men and women having only opposite sex partners to the remainder of the sample as well as to African-Americans compared to the remainder and those having repeated HIV tests compared to the remainder. Thus, attention to a potential backlash effect in pre- and post-vaccination counselling protocols should not vary based on sexual orientation, African-American race, or history of being tested for HIV. Although the effect size was somewhat larger for sex of the participants, our findings nonetheless suggest that a potential backlash effect may apply equally to men and women.

Limitations

Findings are limited by several factors, including the inherent limitations of a cross-sectional study design and the use of a convenience sample. The use of a convenience sample implies that the findings may not be generalizable. An important limitation is that the study was based, by necessity, on asking people 'to imagine' a hypothetical scenario and respond accordingly. Whether people would actually respond to a real vaccine as they indicated in this artificial scenario is problematic. It should also be noted that participation bias in a study about an AIDS vaccine is particularly likely (i.e., it is conceivable that people having extremely negative intent to receive an AIDS vaccine may be the least likely to volunteer for such a study).

Conclusions

These results indicated that substantial further research in this arena is needed. With the seemingly real potential for increased risk behaviours with the advent of an HIV vaccine, and given that increases in risk behaviour may offset vaccine effectiveness in a significant way, understanding the patterns and predictors of such behavioural offsets is critical for the public health. This work should proceed with further, larger-scale descriptive and predictive research; thereafter, intervention studies appear necessary to identify ways to decrease risk behaviours in the presence of HIV vaccines. Many studies have shown that behavioural HIV prevention interventions can work to reduce risk behaviours, but these interventions have not been specifically tested in the context of HIV vaccine availability. This is an area in need of substantial further attention.

Acknowledgements: We are grateful to the Emory Center for AIDS Research for their support and contribution to this study. This study was supported by a grant from NIAID to the second author (P30 DA12121 0351).

References

1 Blower S, Schwartz EJ, Mills J. Forecasting the future of HIV epidemics: the impact of antiretroviral therapies & imperfect vaccines. AIDS Rev 2003;5:113-25

2 Blower SM, McLean AR. Prophylactic vaccines, risk behavior change, and the probability of eradicating HIV in San Francisco. Science 1994;265:1451-4

3 Crosby RA, Holtgrave DR, Bryant L, Frew PM. Correlates of negative intent to receive an AIDS vaccine: an exploratory study. Int J STD AIDS 2004;15: 552-7

4 Hader SL, Smith DK, Moore JS, Holmberg SD. HIV infection in women in the United Sates: status at the millennium. IAMA 2001;285:1186-92

5 Rosenberg PS. Scope of the AIDS epidemic in the United States. Science 1995;270:1372-5

(Accepted 6 April 2005)

Richard A Crosby PhD1 and David R Holtgrave PhD2,3

1 College of Public Health, University of Kentucky, Lexington, KY; 2 Rollins School of Public Health, Department of Behavioral Sciences & Health Education, Lexington, KY; 3 Emory Center for AIDS Research, Lexington, KY, USA

Correspondence to: Dr Richard Crosby, College of Public Health, University of Kentucky, 121 Washington Ave., Room 111C, Lexington, KY 40506-0003, USA

Email: crosby@uky.edu

Copyright Royal Society of Medicine Press Ltd. Mar 2006

Source: International Journal of STD & AIDS

More News in this Category