Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Aug 1.
Published in final edited form as: AIDS Care. 2015 Mar 2;27(8):1000–1004. doi: 10.1080/09540121.2015.1017443

Relationship-level analysis of drug users’ anticipated changes in risk behavior following HIV vaccination

April M Young 1,2, Daniel S Halgin 3, Jennifer R Havens 2
PMCID: PMC4425595  NIHMSID: NIHMS666312  PMID: 25730519

Abstract

Formative research into the behavioral factors surrounding HIV vaccine uptake is becoming increasingly important as progress is made in HIV vaccine development. Given that the first vaccines on the market are likely to be partially effective, risk compensation (i.e. increased risk behavior following vaccination) may present a concern. This study characterized the relationships in which HIV vaccine-related risk compensation is most likely to occur using dyadic data collected from people who use drugs, a high-risk group markedly underrepresented in extant literature. Data were collected from 433 drug users enrolled in a longitudinal study in the US. Respondents were asked to provide the first name and last initial of individuals with whom they had injected drugs and/or had sex during the past 6 months. For each partner, respondents reported their likelihood of increasing risk behavior if they and/or their partner received an HIV vaccine. Using generalized linear mixed models, relationship-level correlates to risk compensation were examined. In bivariate analysis, risk compensation was more likely to occur between partners who have known each other for a shorter time (OR=0.95, 95% CI: 0.90-0.99, p=0.028) and between those who inject drugs and have sex together (OR=2.52, CI: 1.05-6.04, p=0.039). In relationships involving risk compensation, 37% involved partners who had known each other for a year or less compared to only 13% of relationships not involving risk compensation. Adjusting for other variables, duration (OR: 0.95, CI: 0.90-1.00 p=0.033) was associated with risk compensation intent. These analyses suggests that risk compensation may be more likely to occur in less established relationships and between partners engaging in more than one type of risk behavior. These data provide further support for the need to expand measures of risk compensation in HIV vaccine preparedness studies to assess not only if people will change their behavior, but also with whom.

Keywords: HIV vaccine, risk compensation, social network, injection drug use, condom use

Introduction

As advancements are made in HIV vaccine development, the likelihood of reversing the HIV epidemic increases. However, early vaccines on the market are likely to be only partially effective, and post-vaccination increases in risk behavior (i.e., risk compensation (Hogben & Liddon, 2008)) could reduce the public health benefit (Andersson et al., 2007; Blower, Schwartz, & Mills, 2003; Eaton & Kalichman, 2007; Fonseca et al., 2010; Gray et al., 2003). Findings from previous studies examining individuals’ likelihood of risk compensation have been mixed; when queried about others’ behaviors, study participants reported a likely increase in sexual risk behavior if vaccinated (Koniak-Griffin, Nyamathi, Tallen, González-Figueroa, & Dominick, 2007; Newman et al., 2009; Newman, Roungprakhon, Tepjan, Yim, & Walisser, 2012; Olin et al., 2006; Sayles, MacPhail, Newman, & Cunningham, 2010; Webb, Zimet, Mays, & Fortenberry, 1999). However, when asked about personal behaviors, relatively few anticipated engaging in such risk compensation (Barrington, Moreno, & Kerrigan, 2008; Crosby & Holtgrave, 2006; MacPhail, Sayles, Cunningham, & Newman, 2012; O'Connell et al., 2002). These seemingly incongruous findings indicate a need for a more nuanced understanding of risk compensation related to HIV vaccination. Risk compensation is inherently a relationship-level, or dyadic, phenomenon, yet most existing studies employed only global measures (i.e., inquiring about individuals’ likelihood of changing their risk behavior in general). These global measures determine who is most likely to increase risk behavior but not the types of relationships in which behavioral change is likely to occur. This presents a significant gap in knowledge about the potential impact of risk compensation on HIV vaccine initiatives.

Recent network-level research using dyadic data on anticipated HIV vaccine-related risk compensation revealed that increases in risk behavior in certain relationships could impact the connectivity of the overall risk network in which individuals are embedded (Young, Halgin, DiClemente, Sterk, & Havens, 2014). While the study focused on the larger network structure, local dyadic patterns remained largely unexplored. Therefore, the purpose of the present study was to characterize the dyadic relationships in which HIV vaccine-related risk compensation is most likely to occur.

Methods

Sample

The sample consisted of 433 participants enrolled in a longitudinal study investigating HIV, HCV, and herpes-simplex 2 incidence among illicit drug users [described in detail elsewhere (Havens et al., 2013; Young, Jonas, Mullins, Halgin, & Havens, 2013)]. Eligibility criteria included being at least 18 years of age, residing in Appalachian Kentucky, and having used prescription opioids, heroin, crack/cocaine or methamphetamine “to get high” in the prior 30 days. Participants completed interviewer-administered questionnaires and HIV testing at baseline and in 6-month intervals thereafter. A questionnaire to examine attitudes toward HIV vaccination and risk compensation intent was administered during participants’ 24-month assessment; the data from this questionnaire were used for the analyses presented below. Details on the HIV vaccine survey methods have been described previously (Young, DiClemente, Halgin, Sterk, & Havens, 2014a, 2014b; Young, Halgin, et al., 2014). All participants tested HIV negative at the 24-month visit. The protocol was approved by the University Institutional Review Board and a Certificate of Confidentiality was obtained.

Dyadic data

Respondents were asked to provide the first name and last initial of individuals with whom they had injected drugs and/or engaged in sex during the past 6 months. Each partnership constituted a ‘risk relationship’. As described elsewhere (Young, Halgin, et al., 2014), for each partner, respondents were asked about their likelihood of increasing risk behavior if they, the partner, or they and their partner received an HIV vaccine. Specifically, respondents were asked three sex-related items, “If [you/partner/you and partner] got an HIV vaccine that was 90% effective, would you use a condom with them... [‘Much less often’, ‘Less often’, ‘More often’, ‘Much more often’, ‘We wouldn't change how often we used a condom’]”; and three injection-related items, “If [you/partner/you and partner] got an HIV vaccine that was 90% effective, would you use share injection equipment...” [‘Much less often’, ‘Less often’, ‘More often’, ‘Much more often’, ‘We wouldn't change how often we shared equipment’]. Using these data, a binary risk compensation variable was created in which the value of 1 was assigned to risk relationships with anticipated increases in risk behavior in any one of the above six items and 0 if otherwise.

Additional variables included gender similarity (binary), absolute difference in age (years), duration of relationship (years), frequency of communication (6-point Likert scale, with increasing values representing more frequent communication), geographic distance between residences (9-point Likert scale with increasing values indicating farther distances), trust (10-point scale), and whether the respondent received social and/or financial support from each partner (both binary). Participant and partner gender, age, and recent (past 6 month) injection drug use (binary) as reported by the respondent, were also analyzed.

Behavioral dyadic measures included engagement in the following behaviors with each partner: used drugs, injected drugs, injected drugs and had sex, and distributive and/or receptive sharing of injection equipment. Frequency of risk behavior was examined using a variable representing the sum of three Likert scales on which participants rated the frequency of unprotected sex (4-point scale) and needle and cooker sharing (5-point scales) with the partner. Perceived risks and benefits are key to the theoretical foundation of risk compensation (Eaton & Kalichman, 2007; Hogben & Liddon, 2008; Wilde, 1982). Risk perceptions were assessed using the following: “How likely do you think it is that [partner] would ever get infected with HIV?” and “How likely do you think it is that [partner] would ever infect you with HIV?” (4-point Likert scales ranging from “very unlikely” to “very likely”). Participants’ perceived benefit of vaccination was assessed using the following: “In your opinion, how much would an HIV vaccine benefit you?” [1=not at all, 2=a little, 3=some, 4=a lot].

Statistical analyses

Given potential autocorrelation among dyads involving the same respondent, generalized linear mixed models with a random effect for respondent were used. Models were estimated using the PROC GLIMMIX (SAS Institute, 2011) procedure (SAS v9.3). Anticipated risk compensation was regressed on each of the key measures described above. Covariates reaching p<0.10 in bivariate analyses were entered into multivariate analyses. Unadjusted and adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were reported.

Results

Sample demographic and behavioral characteristics have been reported elsewhere (Young, DiClemente, et al., 2014a, 2014b; Young, Halgin, et al., 2014). Participants were predominantly White (94%), male (55%), and unmarried (75%). The median age was 34 years (range: 21–68). Most (82%) reported at least one sexual partner in the past 6 months, 24% reported having multiple partners, 71% reported unprotected sex with at least one partner, and 20% reported unprotected sex with a person who injects drugs. Characteristics of risk partnerships are presented in Table 1. The risk network contained 582 relationships; 78% were sexual only, 12% involved injecting together, and 10% involved injection and sex. There were 30 relationships in which a respondent reported anticipated risk compensation; three anticipated increases in equipment sharing and 27 in unprotected sex.

Table 1.

Bivariate correlates to risk compensation intent in 582 risk relationships

Risk Compensation (n=30) No risk compensation (n=552) OR (95% CI) p-value
Demographic similarities
    Same gender 1 (3.3) 76 (13.8) 0.23 (0.02 – 2.26) 0.209
    Age difference1 (years) 5.4 (5.0) 7.0 (6.7) 0.97 (0.90 – 1.04) 0.344
Relationship characteristics
    Duration (years) 6.3 (7.3) 10.4 (8.8) 0.95 (0.90 – 0.99) 0.028*
    Frequency of contact 5.0 (1.2) 5.4 (1.0) 0.86 (0.60 – 1.23) 0.404
    Distance2 4.4 (2.9) 4.0 (2.7) 1.00 (0.86 – 1.17) 0.973
    Trust 6.1 (3.4) 6.5 (3.2) 1.03 (0.92 – 1.15) 0.625
    Social support 13 (43.3) 265 (48.0) 0.98 (0.38 – 2.56) 0.968
    Receives financial support 11 (36.7) 218 (39.5) 1.34 (0.57 – 3.15) 0.507
    Partner's risk for HIV3 2.3 (0.7) 1.9 (0.9) 1.40 (0.99 – 1.99) 0.055
    HIV risk posed by partner3 1.9 (0.7) 1.6 (0.8) 1.38 (0.83 – 2.28) 0.215
Behavior
    Used drugs together 17 (56.7) 348 (63.0) 0.88 (0.40 – 1.96) 0.757
    Sexual relationship (Ref: inject together only) 29 (96.7) 482 (87.3) 4.15 (0.41-41.76) 0.227
    Inject together and sexual partners 6 (20.0) 54 (9.8) 2.52 (1.05 – 6.04) 0.039*
    Distributive needle sharing 2 (6.7) 40 (7.3) 0.86 (0.13 – 5.66) 0.871
    Receptive needle sharing 2 (6.7) 37 (6.7) 1.05 (0.19 – 5.78) 0.955
    Frequency of risk behavior 2.1 (2.6) 2.6 (1.9) 0.93 (0.69 – 1.26) 0.652
Respondent characteristics
    Male 21 (70.0) 316 (57.3) 1.20 (0.47 – 3.07) 0.698
    Age 33.3 (5.2) 33.1 (8.0) 1.02 (0.98 – 1.06) 0.446
    Recent injection drug use 16 (53.3) 257 (46.6) 1.55 (0.61 – 3.97) 0.359
    Perceived benefit of HIV vaccination 1.47 (0.88 – 2.45) 0.138
Partner's characteristics
    Male 10 (33.3) 248 (44.9) 0.80 (0.31 – 2.09) 0.649
    Age2 31.4 (9.6) 34.0 (9.6) 0.99 (0.94 – 1.03) 0.593
    Recent injection drug use 11 (36.7) 194 (35.1) 1.14 (0.50 – 2.57) 0.762
Open in a new tab
*

p < 0.05; OR: odds ratio; CI: confidence interval

1

Due to two missing values among those not intending risk compensation, 580 relationships were included in analysis.

2

Due to three missing values among those not intending risk compensation and one among those intending risk compensation, 578 relationships were included in analysis.

3

Due to one missing value among those not intending risk compensation, 581 relationships were included in analysis.

Dyadic correlates to likelihood of risk compensation

Table 1 describes bivariate analysis of dyadic correlates of intended risk compensation. Risk compensation was more likely to occur between partners who inject drugs and have sex together (p=0.039) and in partnerships of shorter duration (p=0.028). Among relationships involving risk compensation, 37% were in existence for a year or less compared to only 13% of relationships not involving risk compensation. In multivariate analysis (Table 2), only relationship duration (p=0.046) was significantly associated with risk compensation intent. For every one-year increase in relationship duration, the odds of risk compensation decreased by 5% (OR: 0.95; CI: 0.90–1.00; p=0.033).

Table 2.

Multivariate correlates to risk compensation intent in 581 risk relationships
Adjusted OR (95% CI) p-value
Relationship characteristics
    Duration (years) 0.95 (0.90 – 1.00) 0.033*
    Partner's risk for HIV1 1.34 (0.95 – 1.90) 0.096
Behavior
    Inject together and sexual partners 2.08 (0.82 – 5.23) 0.119
Open in a new tab
*

p < 0.05; OR: odds ratio; CI: confidence interval

1

Due to one missing value among those not intending risk compensation, 581 relationships were included in analysis.

Discussion

While some studies have examined individual-level correlates to risk compensation, few have characterized the types of relationships in which risk compensation is most likely to occur. This gap is significant given that different types of risk relationships pose different levels of HIV transmission risk (Baggaley, Boily, White, & Alary, 2006; Boily et al., 2009). Although risk compensation in our study was relatively uncommon, results suggest that risk compensation was more likely to occur in what could be considered as less established relationships (i.e., relationships of shorter duration). In bivariate analyses, participants were also more likely to engage in risk compensation with partners with whom they had a multiplex risk relationship and to a lesser extent, with partners who they perceived to be at a higher risk for HIV. While seemingly counterintuitive, the increased likelihood of risk compensation in more ‘risky’ partnerships is consistent with the cognitive mechanism underlying risk compensation (Hogben & Liddon, 2008) and relevant to risk calibration (Eaton & Kalichman, 2007). Inhibition is a necessary prerequisite for disinhibition; that is, increased risk behavior would only be expected to occur in those relationships originally perceived as posing a risk.

If risk compensation is most likely to occur in the relationships that present the greatest HIV transmission risk, then risk compensation could pose a greater threat for HIV acquisition than that projected from global measures of anticipated behavior change. More advanced mathematical modeling and simulation is needed to fully explore this possibility. To our knowledge, this is the first study of its kind to identify relationship-level correlates to risk compensation using dyadic analyses. Thus, similar research in different settings and at-risk groups is needed to determine if trends observed in this sample of drug users is consistent with those that can be anticipated in other settings. Overall, these findings emphasize the importance of acknowledging dyadic factors that might influence future risk behaviors and have important implications for vaccine program planning.

Limitations of the present study are described in detail elsewhere (Young, Halgin, et al., 2014), and include factors such as reliance on self-report, measures based on intention rather than actual behavior, and cross-sectional study design. Nevertheless, our findings provide further support for the need to expand measures of risk compensation in HIV vaccine preparedness studies to assess not only if people will change their behavior, but also with whom.

Acknowledgement

The authors would also like to acknowledge Hannah Cooper, Ralph DiClemente, and Claire Sterk for their input during the conceptualization of the study.

Funding

This work was supported by the National Institute on Drug Abuse under Grants R01DA024598 and R01DA033862, National Center for Research Resources and the National Center for Advancing Translational Sciences at the National Institutes of Health under Grant UL1TR000117.

Footnotes

The authors report no conflicts of interest.

References

  1. Andersson KM, Owens DK, Vardas E, Gray GE, McIntyre JA, Paltiel AD. Predicting the impact of a partially effective HIV vaccine and subsequent risk behavior change on the heterosexual HIV epidemic in low- and middle-income countries: A South African example. Journal of Acquired Immune Deficiency Syndromes. 2007;46(1):78–90. doi: 10.1097/QAI.0b013e31812506fd. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Baggaley RF, Boily M-C, White RG, Alary M. Risk of HIV-1 transmission for parenteral exposure and blood transfusion: a systematic review and meta-analysis. AIDS. 2006;20(6):805–812. doi: 10.1097/01.aids.0000218543.46963.6d. [DOI] [PubMed] [Google Scholar]
  3. Barrington C, Moreno L, Kerrigan D. Perceived influence of an HIV vaccine on sexual-risk behaviour in the Dominican Republic. Culture, Health & Sexuality. 2008;10(4):391–401. doi: 10.1080/13691050801898877. [DOI] [PubMed] [Google Scholar]
  4. Blower SM, Schwartz EJ, Mills J. Forecasting the future of HIV epidemics: the impact of antiretroviral therapies & imperfect vaccines. AIDS Reviews. 2003;5(2):113–125. [PubMed] [Google Scholar]
  5. Boily M-C, Baggaley RF, Wang L, Masse B, White RG, Hayes RJ, Alary M. Heterosexual risk of HIV-1 infection per sexual act: systematic review and meta-analysis of observational studies. The Lancet infectious diseases. 2009;9(2):118–129. doi: 10.1016/S1473-3099(09)70021-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Crosby R, Holtgrave D. Will sexual risk behaviour increase after being vaccinated for AIDS? International Journal of STD & AIDS. 2006;17(3):180–184. doi: 10.1258/095646206775809204. [DOI] [PubMed] [Google Scholar]
  7. Eaton LA, Kalichman SC. Risk compensation in HIV prevention: implications for vaccines, microbicides, and other biomedical HIV prevention technologies. Current HIV/AIDS Reports. 2007;4(4):165–172. doi: 10.1007/s11904-007-0024-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Fonseca MGP, Forsythe S, Menezes A, Vuthoori S, Possas C, Veloso V, Stover J. Modeling HIV vaccines in Brazil: assessing the impact of a future HIV vaccine on reducing new infections, mortality and number of people receiving ARV. PLoS ONE. 2010;5(7):e11736–e11736. doi: 10.1371/journal.pone.0011736. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Gray RH, Li X, Wawer MJ, Gange SJ, Serwadda D, Sewankambo NK, Quinn TC. Stochastic simulation of the impact of antiretroviral therapy and HIV vaccines on HIV transmission; Rakai, Uganda. AIDS. 2003;17(13):1941–1951. doi: 10.1097/00002030-200309050-00013. [DOI] [PubMed] [Google Scholar]
  10. Havens J, Lofwall MR, Frost SD, Oser CB, Leukefeld CG, Crosby RA. Individual and network factors associated with prevalent hepatitis C infection among rural Appalachian injection drug users. American Journal of Public Health. 2013;103(1):e44–e52. doi: 10.2105/AJPH.2012.300874. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Hogben M, Liddon N. Disinhibition and risk compensation: scope, definitions, and perspective. Sexually Transmitted Diseases. 2008;35(12):1009–1010. doi: 10.1097/OLQ.0b013e31818eb752. [DOI] [PubMed] [Google Scholar]
  12. Koniak-Griffin D, Nyamathi A, Tallen L, González-Figueroa E, Dominick E. Breaking the silence: what homeless 18- to 24-year-olds say about HIV vaccine trials. Journal of Health Care for the Poor and Underserved. 2007;18(3):687–698. doi: 10.1353/hpu.2007.0061. [DOI] [PubMed] [Google Scholar]
  13. MacPhail CL, Sayles JN, Cunningham W, Newman P. Perceptions of sexual risk compensation following posttrial HIV vaccine uptake among young South Africans. Qualitative Health Research. 2012;22(5):668–678. doi: 10.1177/1049732311431944. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Newman P, Lee SJ, Duan N, Rudy E, Nakazono TK, Boscardin J, Cunningham WE. Preventive HIV vaccine acceptability and behavioral risk compensation among a random sample of high-risk adults in Los Angeles (LA VOICES). Health Services Research. 2009;44(6):2167–2179. doi: 10.1111/j.1475-6773.2009.01039.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Newman P, Roungprakhon S, Tepjan S, Yim S, Walisser R. A social vaccine? Social and structural contexts of HIV vaccine acceptability among most-at-risk populations in Thailand. Global Public Health. 2012;7(9):1009–1024. doi: 10.1080/17441692.2012.692388. [DOI] [PubMed] [Google Scholar]
  16. O'Connell JM, Hogg RS, Chan K, Strathdee SA, McLean N, Martindale SL, Remis R. Willingness to participate and enroll in a phase 3 preventive HIV-1 vaccine trial. Journal of Acquired Immune Deficiency Syndromes. 2002;31(5):521–528. doi: 10.1097/00126334-200212150-00010. [DOI] [PubMed] [Google Scholar]
  17. Olin J, Kokolamami J, Lepira FB, Mwandagalirwa K, Mupenda B, Ndongala ML, Mokili J. Community preparedness for HIV vaccine trials in the Democratic Republic of Congo. Culture, Health & Sexuality. 2006;8(6):529–544. doi: 10.1080/13691050600888434. [DOI] [PubMed] [Google Scholar]
  18. SAS Institute . The GLIMMIX Procedure SAS/STAT 9.3 User's Guide. SAS Institute; Cary, NC: 2011. [Google Scholar]
  19. Sayles JN, MacPhail CL, Newman P, Cunningham WE. Future HIV vaccine acceptability among young adults in South Africa. Health Education & Behavior. 2010;37(2):193–210. doi: 10.1177/1090198109335654. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Webb PM, Zimet GD, Mays R, Fortenberry JD. HIV immunization: Acceptability and anticipated effects on sexual behavior among adolescents. Journal of Adolescent Health. 1999;25(5):320–322. doi: 10.1016/s1054-139x(99)00066-x. [DOI] [PubMed] [Google Scholar]
  21. Wilde GJ. The theory of risk homeostasis: implications for safety and health. Risk Analysis. 1982;2(4):209–225. [Google Scholar]
  22. Young AM, DiClemente RJ, Halgin DS, Sterk CE, Havens JR. Drug Users’ Willingness to Encourage Social, Sexual, and Drug Network Members to Receive an HIV Vaccine: A Social Network Analysis. AIDS and Behavior. 2014a;18:1753–1763. doi: 10.1007/s10461-014-0797-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Young AM, DiClemente RJ, Halgin DS, Sterk CE, Havens JR. HIV vaccine acceptability among high-risk drug users in Appalachia: a cross-sectional study. BMC Public Health. 2014b;14(1):537. doi: 10.1186/1471-2458-14-537. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Young AM, Halgin DS, DiClemente RJ, Sterk CE, Havens JR. Will HIV vaccination reshape HIV risk behavior networks? A social network analysis of drug users' anticipated risk compensation. PLoS One. 2014;9(7):e101047. doi: 10.1371/journal.pone.0101047. doi: 10.1371/journal.pone.0101047. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Young AM, Jonas A, Mullins U, Halgin D, Havens J. Network structure and the risk for HIV transmission among rural drug users. AIDS and Behavior. 2013;17(7):2341–2351. doi: 10.1007/s10461-012-0371-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES