HIV prevention options expanded to include antiretroviral-based products after several clinical trials demonstrated that once-daily emtricitabine/tenofovir disoproxil fumarate (FTC/TDF) as pre-exposure prophylaxis (PrEP) reduced the risk of HIV acquisition.1–3 However, 2 clinical trials assessing the role of FTC/TDF as PrEP for HIV prevention in women—FEM-PrEP and MTN-003 [Vaginal and Oral Interventions to Control the Epidemic (VOICE)]—did not demonstrate effectiveness of FTC/TDF in their study populations because of low adherence.4,5 Low risk perceptions among study participants may have been one of the reasons for the overall poor adherence observed in FEM-PrEP.4
Perceived risk is one of several key constructs in many theories of health behavior change, such as the health belief model6 and the protection motivation theory.7 Within such frameworks, individuals may be motivated to take action to reduce risk behaviors if they perceive some risk; individuals may also be discouraged to modify their risk behaviors if they perceive no risk. Perception of risk is frequently assessed by asking individuals the question “What is your chance of getting X [ie, the illness] in the next X [ie, timeframe]?”8 However, defining and measuring perception of risk is complex. Several terms are used interchangeably in public health to describe perceived risk, such as “perceived likelihood,” “perceived susceptibility,” and “perceived vulnerability.”8,9 Adding to the complexity, perceived risk may be the outcome of past sexual behavior or it may predict future sexual behavior.10–12
In the field of HIV prevention, a substantial body of literature exists on HIV risk perceptions, including from early in the epidemic.10,11 Research has primarily focused on how perceived risk should be defined and measured, on identifying factors associated with perceived risk, and on assessing the associations between perceived risk and engagement in risk behaviors.10,11,13–20 Findings have varied on the associations between perceived risk and risk-reduction behaviors; however, these variations may be due in part to the interpretation of results from cross-sectional studies and differences in measurement.10–12
In FEM-PrEP, we explored perceptions of HIV risk within a different context than which they are explored in HIV prevention programs that encourage adherence to efficacious risk-reduction methods. We assessed risk perceptions within the context of a placebo-controlled clinical trial, in which counseling focused on encouraging adherence to proven HIV risk-reduction methods (eg, condoms) and to a daily, investigational pill as PrEP for HIV prevention (ie, FTC/TDF). We hypothesized that one reason for the poor adherence observed in FEM-PrEP was low perceived risk of HIV among study participants. Here we describe participants' HIV risk perceptions, factors associated with perceived HIV risk, and the association between HIV risk perceptions and adherence to once-daily, investigational FTC/TDF for HIV prevention.
Overview of the FEM-PrEP Clinical Trial
FEM-PrEP was a phase 3, placebo-controlled clinical trial to assess the safety and effectiveness of oral FTC/TDF for the prevention of HIV in women in Bondo, Kenya; in Bloemfontein and Pretoria, South Africa; and in Arusha, Tanzania. The trial has been described elsewhere.4 In brief, women who were considered to be at higher risk of HIV acquisition were randomized to once-daily FTC/TDF or placebo and asked to take their assigned study pill for 52 weeks. Women who reported at screening that they had had sex at least once within the past 2 weeks or had had sex with more than 1 sexual partner in the past month were considered to be at higher HIV risk. These criteria were chosen given the overall context of a generalized epidemic, high HIV prevalence at the sites, and information on incidence and sexual risk eligibility criteria from previous HIV prevention trials with women in sub-Saharan Africa. At each 4-week clinic visit (for up to 60 weeks), participants received risk-reduction counseling and a free supply of condoms, separate adherence counseling and a resupply of study pills (until week 52), and treatment of any curable sexually transmitted infections. The trial began in July 2009 and closed early in April 2011 because of lack of effectiveness.21 A total of 2120 participants were enrolled. Among the subcohort of participants described in this manuscript, 12% achieved good adherence (defined below) to FTC/TDF, 23% rarely or never took FTC/TDF, and 60% had fluctuating adherence throughout their participation in the trial.22
We assessed perceptions of HIV risk among all participants through a quantitative, face-to-face interview at enrollment and at quarterly follow-up visits. We defined perceived HIV risk as a woman's assessment of the likelihood that she would acquire HIV over a specific period (ie, perceived absolute risk23). We used the standard single item to measure this construct: participants were asked to categorize their belief in the likelihood of becoming infected with HIV in the next 4 weeks. Response options, which were read aloud, were (1) no chance: no possibility of becoming infected with HIV, (2) a small chance: could happen but not likely, (3) a moderate chance: some possibility of becoming infected with HIV, and (4) a high chance: likely to become infected with HIV. The short time frame for the assessment—perceived risk in the next 4 weeks—allowed participants to consider their immediate HIV risk in the context of their current sexual partner(s), risk reduction practices, and surrounding environment rather than their risk over a longer period in which their partners and risk situations could change. Participants were also asked to describe (1) whether they agreed or disagreed (on a 4-point Likert scale—strongly agreed, agreed, disagreed, and strongly disagreed) that their own behavior or their primary partner's behavior puts them at risk for HIV, to better understand perceptions of personal versus partners' behaviors and perceived risk and (2) whether they agreed or disagreed (on the same 4-point Likert scale) that they can protect themselves from acquiring HIV, as a single measure to represent self-efficacy.
We also assessed the following variables through a quantitative, face-to-face interview with all participants at enrollment and at each quarterly follow-up visit: their sexual partnerships (primary partner, other partners); the number of sexual partners (primary and other) they had vaginal sex within the past 7 days; the number of vaginal sex acts with primary and other partners in the past 7 days; if they had anal sex within the past 4 weeks; condom use during vaginal and anal sex with primary and other partners in the past 4 weeks; knowledge of primary and others sexual partners' HIV status; their belief that their primary partner had vaginal or anal sex with another partner in the past 4 weeks; and whether they exchanged sex for money or goods in the past 4 weeks. The quantitative interview was conducted toward the beginning of the study visit before counseling on risk reduction and adherence and by different staff from those providing counseling. Answers were not used to inform subsequent risk reduction counseling. Demographic information was collected at screening.
Samples were collected to test for sexually transmitted infections and other reproductive tract infections (ie, syphilis, chlamydia, gonorrhea, trichomoniasis, bacterial vaginosis, and candidiasis) at screening, as clinically indicated during follow-up visits, and at the final visit, if the participant agreed to a pelvic examination. We also collected blood specimens from all participants at each study visit between week 4 and week 52 for analysis of drug concentrations after trial closure.
Adherence Study Sample
We assessed the association between perceived risk and adherence among a subcohort of 150 participants assigned to FTC/TDF who completed at least 1 follow-up visit. The sample consisted of 50 participants selected completely at random from each site where HIV infections occurred (Bloemfontein, Bondo, and Pretoria); no participants from the Arusha site were included in the subcohort because the site had only recently been initiated at the time of study closure and no participants had seroconverted. Based on empirical estimates of within-subject correlations of responses over time, the size of this subcohort provided 80% power to detect an odds ratio (OR) of 2.25 for the association between reporting some perceived risk of HIV and good pill adherence, given that 40%–60% of participants reported some perceived risk and given an overall prevalence of good adherence greater than 20%.
Concentrations of tenofovir (TFV) in plasma and of tenofovir diphosphate in upper layer packed cells24,25 were assessed for samples collected at weeks 4, 16, 28, and 40 when drug was available to use. Visits were excluded when drug was not available for use because of a protocol-defined pill withdrawal (eg, an adverse event) or a missed pill supply visit. We defined good adherence as having both a TFV concentration in plasma exceeding 10 ng/mL and an intracellular tenofovir diphosphate concentration in upper layer packed cells exceeding 100,000 fmol/mL. With these drug concentrations, the estimated doses per interval are between approximately 4 doses per week and daily dosing. We used the drug concentration measurement that was closest to 4 weeks after the risk assessment [median days = 28 (range: 19–37; interquartile range = 0)] and excluded all other drug concentration measurements. Additional details of the FEM-PrEP analysis of drug concentrations and adherence are described elsewhere.22
We used logistic regression to assess the association between some perceived HIV risk (ie, small, moderate, or high) reported at enrollment and individual- and dyad-level factors among the user population sample of participants from Bondo, Bloemfontein, and Pretoria. This analysis population included all participants who were randomized and excluded participants who never received the study pills, returned all pills unused, never returned for a follow-up visit, or were HIV infected at enrollment. We dichotomized participants' risk perceptions by none versus some (small, moderate, or high) rather than by the ordinal responses (none, small, moderate, or high) because of the few responses in some of the small, moderate, and high categories in the sub-cohort across sites. We also chose to combine the participants who perceived a small, moderate, or high risk into 1 group because they acknowledged a potential likelihood of becoming infected with HIV and therefore may have differed from participants who believed they were not at risk at all. We first conducted a bivariate analysis by study site and pooled across sites, followed by a multivariate analysis using factors initially identified to be significantly associated (P < 0.05) with some perceived risk at enrollment in the pooled bivariate model. The final multivariate model was simplified using stepwise variable selection (P < 0.05). Factors considered were age, education, marital status, occupation, partner type, number of primary and other partners in the past 7 days, number of vaginal sex acts with primary and other partners in the past 7 days, engaged in transactional sex in the past 4 weeks, had vaginal or anal sex with primary or other partner without using a condom in the past 4 weeks, knowledge of whether primary or other partner has HIV, belief that primary partner had vaginal or anal sex with another partner in the past 4 weeks, and diagnosis of a sexually transmitted infection. Variables assessed for both primary and other partners were combined in the analysis to maximize the analysis sample size.
We also used logistic regression with robust variance estimation to examine the association between risk perception [none versus some (small, moderate, or high)] reported among participants in the adherence sub-cohort at enrollment and at weeks 12, 24, and 36 and good adherence from the corresponding measurement of drug concentration 4 weeks later at weeks 4, 16, 28, and 40, respectively. We initially adjusted for all 3 factors previously found to be significantly associated with adherence in FEM-PrEP22: use of oral contraceptive pills at enrollment, study site, and liking the color of the pill. Liking the pill color was removed from the final model during stepwise variable selection (P > 0.05). Using oral contraceptive pills at enrollment and study site remained significant and were adjusted in the final model. SAS 9.326 was used for the analyses and statistical significance was defined as P ≤ 0.05.
The protocol for the FEM-PrEP clinical trial was approved by the University of The Free State Ethics Committee in South Africa (Bloemfontein), the Kenyatta National Hospital/University of Nairobi Research and Ethics Committee in Kenya (Bondo), the Medunsa Campus Research Ethics Committee in South Africa (Pretoria), and the Protection of Human Subjects Committee at FHI 360 in the United States. All participants provided their written informed consent.
Demographic characteristics, sexual behaviors, and beliefs about sexual partners collected at baseline among all participants (ie, the user population) generally differed between participants in Bondo and participants in the 2 South African sites (Table 1). In Bondo, the majority of participants were 25 years or older (61%) and were married (74%). Many participants in Bondo (43%) reported having more than 1 sexual partner, and a high percentage (82%) reported having had sex with a primary partner without using a condom in the past 4 weeks. Numerous participants (51%) were unsure whether their primary partner had HIV, and 33% believed their primary partner had or might have had other sexual partners in the past 4 weeks. In contrast, most women from the 2 South African sites—Bloemfontein and Pretoria—were 24 years or younger (75% and 69%, respectively) and not married (94% and 93%, respectively). Few participants at the Bloemfontein and Pretoria sites (14% and 13%, respectively) reported having more than 1 partner. Although condom use was higher among women in Bloemfontein and Pretoria than among women in Bondo, a sizable percentage of participants in both Bloemfontein and Pretoria (52% and 65%, respectively) reported having had sex with their primary partner without using a condom in the past 4 weeks. Also in comparison with participants in Bondo, fewer participants in Bloemfontein and Pretoria (18% and 31%, respectively) were unsure whether their primary partner had HIV, and fewer (14% and 8%, respectively) believed their primary partner had or might have had other sexual partners. One noticeable difference among all 3 sites was the proportion of participants who were students: 45% in Bloemfontein, 22% in Pretoria, and 2% in Bondo. These data were similar among participants randomly assigned to the adherence sub-cohort.
Perceived HIV Risk
Among the user population, HIV risk perceptions varied significantly at each time point across study sites (all P < 0.001, Fisher's exact test) and were much lower among participants in Bloemfontein and Pretoria than among participants in Bondo (Fig. 1). At enrollment, a higher percentage of participants from Bloemfontein and Pretoria (66% and 61%, respectively) than from Bondo (20%) perceived they had no chance of acquiring HIV in the next 4 weeks. Similarly, more participants in Bondo (20%) than in Bloemfontein and Pretoria (4% and 3%, respectively) believed they had a high chance of becoming HIV infected in the 4 weeks following enrollment. The frequency of each category of risk perception fluctuated slightly throughout the trial among participants, but the overall trend remained: participants in Bondo continued to have a wider distribution and higher perceived risk than participants in the 2 South African sites and most participants in the 2 South African sites continued to report no perceived risk or a small perceived risk. We found no difference in the distribution of none versus some perception of HIV risk at each time point among participants in the adherence sub-cohort and the larger user population (P > 0.1, Fisher's exact test).
When asked to all participants at baseline, most agreed or strongly agreed that they can protect themselves from getting HIV (Bloemfontein: 96%; Pretoria: 89%; Bondo: 93%). Perceptions varied, however, among the sites on whether the participants' own behavior and the behavior of their primary partner put the participant at risk. In Bondo, 65% of participants agreed or strongly agreed that their own behavior puts them at risk for HIV and 85% said their primary partner's behavior puts them at risk compared with 14% and 44% of participants from Pretoria and 25% and 45% of participants from Bloemfontein, respectively.
Factors Associated With Perceived HIV Risk
Bivariate analyses found several individual- and dyad-level factors associated with having some perceived risk at enrollment, at the site level and across all sites (Table 2). In multivariate analyses, the following factors remained statistically significant when all sites were combined: having more than 1 sexual partner [OR: 1.9; 95% confidence interval (CI): 1.4 to 2.5], having more vaginal sex acts with a primary or other partners in the past 7 days (OR: 1.1 per vaginal sex act; 95% CI: 1.0 to 1.1), having vaginal or anal sex without a condom with a primary or other partner within the past 4 weeks (OR: 1.8; 95% CI: 1.4 to 2.2), not knowing if a primary or other partner had HIV (OR: 1.8; 95% CI: 1.4 to 2.2), believing that a primary partner had or might have had other sexual partners in the past 4 weeks (OR: 4.0; 95% CI: 2.8 to 5.6), and not knowing if a primary partner has had other partners in the past 4 weeks (OR: 2.4; 95% CI: 1.9 to 3.0) (Table 2). Most associations remained statistically significant at the site level (data not shown).
Perceived HIV Risk and Adherence
We observed a significant positive association between some perceived HIV risk reported at enrollment and at each quarterly visit and good adherence based on drug concentrations measured 4 weeks later. The association was significant when all 3 sites were combined (OR: 2.0; 95% CI: 1.1 to 3.5; P = 0.016) (Table 3) and at the site level among participants in Bondo (OR: 4.7; 95% CI: 1.3 to 17.0; P = 0.018) and Bloemfontein (OR: 4.5; 95% CI: 1.8 to 11.5; P = 0.001). We did not, however, find a significant association between risk perception and adherence among participants in Pretoria (OR: 0.9; 95% CI: 0.3 to 2.6; P = 0.878).
However, among participants from all 3 sites, the percentage who reported some perceived risk and who also achieved good adherence was moderate to low: 41% (29/70) at week 4, 31% (17/54) at week 16, 34% (19/56) at week 28, and 25% (8/32) at week 40.
Many FEM-PrEP participants, particularly women from South Africa, perceived that they had little or no risk of acquiring HIV within the next 4 weeks. Perception of HIV risk is believed to motivate adherence to PrEP among women27; however, to our knowledge, our study is the first to provide empirical evidence on the association between risk perceptions and adherence within the context of a placebo-controlled PrEP clinical trial. In the weeks following the risk assessment, adherence to the study pill was higher among participants who perceived some risk than among those who perceived no risk. These data suggest that risk perceptions may have played a role in some women's decisions to adhere to the study pill during FEM-PrEP. Yet, even among participants with some perceived risk, overall adherence was moderate to low, suggesting that other factors also played a role. In a previous analysis,22 we identified that being from the Bloemfontein site and liking the color of the pill was positively associated with good adherence, whereas using oral contraceptive pills at enrollment was negatively associated with good adherence. We will also describe elsewhere findings from a follow-up study conducted after trial closure on participants' explanations for adherence, and non-adherence, during the trial.
Our findings on the factors associated with having some perceived risk suggest that some participants likely understood the sexual behaviors that increase women's risk of HIV and may have reflected on such risks in their lives when describing their own risk perceptions. We also observed statistically significant differences among sites in perceived risk and in almost all sexual behaviors and beliefs. Elsewhere we have commented on such differences among participants in the Bondo and Pretoria sites, focusing on risk behaviors and the broader context28 and HIV risk perceptions and risk rationalizations among participants who seroconverted.29
Perceived HIV risk may have also played a role in adherence in other PrEP clinical trials. In Partners PrEP, for which high adherence to the study pill was demonstrated,1,30 sexual behavior (having no sex or having sex with a non-study partner) was positively associated with nonadherence to the study pill among the HIV-negative partners in serodiscordant relationships.30 The authors postulate that participants' perceptions of risk surrounding their sexual behaviors may have influenced their adherence decisions; participants may have perceived no HIV risk during times when they were not having sex and perceived low risk when they were having sex with a non-study partner, particularly if they knew that partner did not have HIV. PrEP was also perceived among some HIV-negative participants in Partners PrEP to be a solution for resolving the “discordance dilemma”—the conflict that emerges among HIV-negative partners who want to remain HIV negative but also wish to maintain their relationship with the partner who has HIV.31
A strength of our analysis was the prospective collection of data.12 Risk perceptions were assessed before participants performed the subsequent behavior of pill adherence, which allowed us to compare drug concentrations representing the overlapping time in which participants were asked to reflect upon the likelihood of becoming infected with HIV. Furthermore, risk perceptions may change over time, and our analysis captured participants' adherence at any measured time in the trial when they perceived some HIV risk. A limitation of our assessment is that we measured risk perception using a single measure. Bradley et al15 reported a stronger correlation between perceived risk and self-reported risk behavior among women seeking HIV voluntary counseling and testing in Ethiopia when using a risk perception scale than when using a single measure similar to that used in FEM-PrEP. Additionally, individuals may have different interpretations of the question that assesses risk perception8,15,32; therefore, some researchers9,15,32 recommend measuring the different aspects of the construct separately. For example, Brewer et al9 argue that “perceived likelihood” (ie, an individual's perception of the probability of experiencing a harm because of a behavioral risk) and “perceived susceptibility” (ie, an individual's feelings on how vulnerable he or she is to experiencing the harm) are distinguishable and measurable dimensions of perceived risk, together with “perceived severity.” In FEM-PrEP, it is unknown what specific factors participants considered when they reported their perceived risk. We will describe elsewhere the findings from follow-up interviews and focus group discussions on how perceptions of HIV are constructed, which were conducted with FEM-PrEP participants and other women in the FEM-PrEP communities after trial closure. Finally, our decision to choose a limited time frame (ie, the subsequent 4 weeks) as the anchor point for perceived risk may have helped participants reflect on their current sexual risk context rather than hypothesize about their risk over a longer period. Alternately, it may have been emotionally difficult for participants to understand or admit their immediate risk.
It is important to note that our findings describe the association between perceived risk and adherence to an investigational drug within the context of a placebo-controlled clinical trial. As such, participants were told at each study visit that they were assigned either the study pill or a placebo and that it was unknown whether the study pill could prevent HIV. This overall context may have—or may not have—affected participants' decisions about adherence, even among those with higher perceptions of risk. Now that FTC/TDF has been shown to be effective at reducing HIV acquisition,1–3 future research can explore the role of risk perception in the uptake of and adherence to a daily HIV prevention pill.
The authors are grateful to the women who participated in the FEM-PrEP trial, to the study staff, to the communities who partnered with them to conduct the trial, and to all of the collaborators in Africa, Belgium, and the United States. The views expressed in this publication do not necessarily reflect those of FHI 360, the funding agencies, or Gilead Sciences, Inc.
FEM-PrEP Study Group: Walter Agingu, Julie Ambia, Jesse Asewe, Irith De Baetselier, Savi Chetty-Tulsee, Sarah Chiduo, Tania Crucitti, Jennifer Deese, Katrien Fransen, John Gardner Gaddiel, Saidi Kapiga, Haddie Kiernan, Stella Kirkendale, Tharnija Lalbahadur, Michele Lanham, Temu Lucky, Paul Mak’Oketch, Shumani Makatu, Mookho Malahleha, Justin Mandala, Rachel Manongi, Makanda Mankalimeng, Martha Masaki, Timothy Mastro, Kevin McKenna, Modie Constance Monedi, Sarah Mullins, Jacob Odhiambo, Kavita Nanda, Immaculate Olango, Paul Omullo, Zablon Omungo, Fred Owino, Caleb Parker, Brian Perry, Malebo Ratlhagana, Ilse Reblin, Lisa Saylor, Phumzile Siguntu, Joseph Short Skhosana, Doug Taylor, Amanda Troxler, Valentine Veena, Lalitha Venkatasubramanian, Gustav Venter, and Christina Wong.
1. Baeten JM, Donnell D, Ndase P, et al.. Antiretroviral prophylaxis for HIV prevention in heterosexual men and women. N Engl J Med. 2012;367:399–410.
2. Thigpen MC, Kebaabetswe PM, Paxton LA, et al.. Antiretroviral preexposure prophylaxis for heterosexual HIV transmission in Botswana. N Engl J Med. 2012;367:423–434.
3. Grant RM, Lama JR, Anderson PL, et al.. Preexposure chemoprophylaxis for HIV prevention in men who have sex with men. N Engl J Med. 2010;363:2587–2599.
4. Van Damme L, Corneli A, Ahmed K, et al.. Preexposure prophylaxis for HIV infection among African women. N Engl J Med. 2012;367:411–422.
5. Marrazzo JM, Ramjee G, Nair G, et al.. Pre-exposure prophylaxis for HIV in women: daily oral tenofovir, oral tenofovir-emtricitabine, or vaginal tenofovir gel in the VOICE Study (MTN 003) [26LB]. Presented at: The 20th Conference on Retroviruses and Opportunistic Infections; March 3–6, 2013; Atlanta, GA.
6. Rosenstock IM. The health belief model and preventive health behavior. Health Educ Behav. 1974;4:354–386.
7. Rogers RW. Cognitive and physiological processes in fear appeals and attitude change: a revised theory of protection motivation. In: Cacioppo JR, Perry RE, eds. Social Psychology: A Sourcebook. New York, NY: Guilford Press; 1983:153–176.
8. Brewer NT, Weinstein ND, Cuite CL, et al.. Risk perceptions and their relation to risk behavior. Ann Behav Med. 2004;27:125–130.
9. Brewer NT, Chapman GB, Gibbons FX, et al.. Meta-analysis of the relationship between risk perception and health behavior: the example of vaccination. Health Psychol. 2007;26:136–145.
10. Kowalewski MR, Henson KD, Longshore D. Rethinking perceived risk and health behavior: a critical review of HIV prevention research. Health Educ Behav. 1997;24:313–325.
11. Gerrard M, Gibbons FX, Bushman BJ. Relation between perceived vulnerability to HIV and precautionary sexual behavior. Psychol Bull. 1996;119:390–409.
12. Weinstein ND, Nicolich M. Correct and incorrect interpretations of correlations between risk perception and risk behaviors. Health Psychol. 1993;12:235–245.
13. Tenkorang EY, Rajulton F, Maticka-Tynadle E. Perceived risks of HIV/AIDS and first sexual intercourse among youth in Cape Town, South Africa. AIDS Behav. 2009;13:234–245.
14. Do M, Meekers D. Multiple sex partners and perceived risk of HIV infection in Zambia: attitudinal determinants and gender differences. AIDS Care. 2009;21:1211–1221.
15. Bradley H, Tsui A, Hindin M, et al.. Developing scales to measure perceived HIV risk and vulnerability among Ethiopian women testing for HIV. AIDS Care. 2011;23:1043–1052.
16. Macintyre K, Rutenberg N, Brown L, et al.. Understanding perceptions of HIV risk among adolescents in KwaZulu-Natal. AIDS Behav. 2004;8:237–250.
17. Majaraj P, Cleland J. Risk perception and condom use among married or cohabiting couples in KwaZulu-Natal, South Africa. Int Fam Plan Perspect. 2005;31:24–29.
18. Stringer EM, Sinkala M, Kumwenda R, et al.. Personal risk perception, HIV knowledge and risk avoidance behavior, and their relationship to actual HIV serostatus in an urban African obstetric population. J Acquir Immune Defic Syndr. 2004;35:60–66.
19. Akwara PA, Madise NJ, Hinde A. Perception of risk of HIV/AIDS and sexual behaviour in Kenya. J Biosoc Sci. 2003;35:385–411.
20. Mbizvo EM, Msuya SE, Hussain A, et al.. HIV prevalence in Zimbabwean women: 54-67% knowledge and perceived risk. Int J STD AIDS. 2003;14:202–207.
22. Corneli AL, Deese J, Wang M, et al.. FEM-PrEP: adherence patterns and factors associated with adherence to a daily, oral study product for pre-exposure prophylaxis. J Acquir Immune Defic Syndr. 2014;66:324–331.
23. Waters EA, McQueen A, Cameron LD. Perceived risk and its relationship to health-related decisions and behavior. In:Martin LR, DiMatteo MR, eds. The Oxford Handbook of Health Communication, Behavior Change, and Treatment Adherence. New York, NY: Oxford University Press; 2014:193–213.
24. Anderson PL, Glidden DV, Liu A, et al.. Emtricitabine-tenofovir concentrations and pre-exposure prophylaxis efficacy in men who have sex with men. Sci Transl Med. 2012;4:151ra125.
25. Adams JL, Sykes C, Menezes P, et al.. Tenofovir diphosphate and emtricitabine triphosphate concentrations in blood cells compared with isolated peripheral blood mononuclear cells: a new measure of antiretroviral adherence? J Acquir Immune Defic Syndr. 2013;62:260–266.
26. SAS. Version 9.2. Cary, NC: SAS Institute, Inc; 2009.
27. Aaron E, Cohan D. Preexposure prophylaxis for the prevention of HIV transmission to women. AIDS. 2013;27:F1–F5.
28. Headley J, Lemons A, Corneli A, et al.. The sexual risk context among the FEM-PrEP study population in Bondo, Kenya and Pretoria, South Africa. PLoS One. 2014;9:e10641.
29. Corneli A, McKenna K, Headley J, et al.. A descriptive analysis of perceptions of HIV risk and worry about acquiring HIV among FEM-PrEP participants who seroconverted in Bondo, Kenya, and Pretoria, South Africa. J Int AIDS Soc. 2014;17:19152.
30. Haberer JE, Baeten JM, Campbell J, et al.. Adherence to antiretroviral prophylaxis for HIV prevention: a substudy cohort within a clinical trial of serodiscordant couples in East Africa. PLoS Med. 2013;10:e1001511.
31. Ware NC, Wyatt MA, Haberer JE, et al.. What's love got to do with it? Explaining adherence to oral antiretroviral pre-exposure prophylaxis for HIV-serodiscordant couples. J Acquir Immune Defic Syndr. 2012;59:463–468.
32. Windschitl PD. Measuring and conceptualizing: perceptions of vulnerability/likelihood. Presented at: The Conceptualizing and Measuring Risk Perceptions Workshop; February 13-14, 2003; Washington, DC.