Introduction

Several clinical trials have demonstrated daily oral HIV pre-exposure prophylaxis (PrEP) with the antiretroviral combination of either tenofovir disoproxil fumarate and emtricitabine (TDF/FTC) or tenofovir alafenamide and emtricitabine (TAF/FTC) to be well tolerated and effective in decreasing the risk of HIV acquisition in high-risk populations, such as men who have sex with men (MSM) and transgender women (TGW) [1,2,3,4,5]. The 2021 guidelines from the Centers for Disease Control and Prevention (CDC) now include intermittent dosing strategies (i.e., 2-1-1 oral PrEP) and long-acting injectable options such as cabotegravir in addition to the mainstay of consistent, daily dosing of either TDF/FTC or TAF/FTC. Even with these new additions, PrEP adherence remains paramount to successful HIV prevention [6].

The Southeastern United States experiences the highest rates of new HIV diagnoses nationwide. In 2018, of 37,968 new HIV diagnoses in the U.S., 51% were in the Southeastern U.S. [7]. Despite the large burden of disease faced by this region and PrEP’s efficacy and safety, the Southeastern U.S. has the lowest levels of PrEP use of any region in the nation, especially relative to its HIV burden [8]. Research on the psychosocial predictors of non-adherence among individuals taking PrEP is limited. Studies on perceived barriers to PrEP by potential PrEP users suggest that social stigma may be an important barrier to adherence [9].

Stigma is defined as a personal characteristic or attribute that confers a negative value or judgment onto an individual, thereby socially discrediting the individual [10]. Several factors associated with PrEP make it stigmatizing for those who take it, including the fact that it is an HIV medication and, thus, puts users at risk of being perceived by partners as HIV-positive [11, 12]. Another factor is the perception that being on PrEP insinuates participation in promiscuous sexual behavior that may put one at risk of HIV [13, 14]. PrEP-related stigma is further magnified when it intersects with stigmatized identities such as being part of a racial, sexual, or gender minority [15]. Understanding the nuances and drivers of stigma experienced by potential and current PrEP users could provide insights into how providers can optimize PrEP uptake and adherence. There has been great interest in developing instruments to measure PrEP stigma in recent years [16,17,18], but fully validated instruments for U.S. populations are few [19]. These existing instruments also do not measure for multiple dimensions of stigma.

This work is informed by stigma frameworks in the HIV stigma literature, which define stigma as societal beliefs that affect individuals with perceived stigmatized traits (i.e., living with HIV) via multiple dimensions including community stigma, internalized stigma, anticipated stigma, and experienced stigma [20,21,22]. These stigma mechanisms have downstream effects on health behavior and outcomes such that people living with HIV (PLWH) who experience higher levels of HIV-related stigma have lower adherence to antiretroviral therapy (ART) and less engagement in HIV care [23,24,25,26]. One potential pathway for these effects of stigma on adherence is self-efficacy—the confidence in oneself to adhere to a treatment plan [27, 28]. Previous research suggests that the association between internalized HIV stigma and health outcomes may be mediated by adherence self-efficacy [26, 29]. There are emerging data as well related to the impact of PrEP stigma on PrEP adherence. Multiple studies suggest that PrEP adherence may be negatively influenced by PrEP stigma in some populations, specifically MSM [12, 30]. It is also likely that PrEP stigma impacts other points on the PrEP care continuum, such as initiation and retention in care, leading to low numbers of patients receiving and remaining on PrEP [31].

In this study, we developed measures of dimensions of stigma related to PrEP and evaluated associations with PrEP adherence, using both self-reported adherence measures and plasma TDF levels. To date, most studies investigating the role of PrEP-related stigma on PrEP adherence have only utilized self-reported adherence measures, so the addition of biological measures in this study is novel. We also sought to explore self-efficacy as a potential mediator in PrEP adherence. Informed by qualitative work with people currently using or eligible for PrEP in the community of Birmingham, Alabama [32], we developed a multi-part survey instrument and administered it to patients currently on PrEP in either of the only two clinics providing PrEP services in Birmingham, Alabama at the time of the study.

Methods

Measure Development

Measures for the study were developed in two phases. In the first phase, 44 current and potential PrEP users participated in qualitative in-depth interviews (emerging themes were reported elsewhere [32]). Recruitment flyers were posted in community-based health clinics providing sexual health services in Birmingham. Participants met with experienced interviewers and were asked to share their perspectives on attitudes about PrEP in their community. The research team then developed measures relevant to understanding PrEP-related stigma as well as PrEP adherence. Questionnaire items were adapted from in-depth interview content as well as validated HIV stigma measures such as the HIV Stigma Scale [21, 33]. The item inventory was then tested using cognitive interviews with 17 current PrEP users. Based on the feedback obtained in the cognitive interviews, the research team met again and narrowed down the items by scales (internalized, perceived, experienced, anticipated PrEP-related stigma and PrEP-related disclosure concerns). We also developed measures of PrEP stigma in healthcare settings and community HIV stigma, which were not included in the current study. The study measures are detailed in Table S1.

Participants, Settings, and Procedure

Participants in this study were current PrEP users recruited from two clinical sites between May 2018 and November 2018. Individuals were eligible for the study if they were over 18 years old, English-literate, and they attended a return visit for PrEP services. Potential participants were approached by research staff during their visit and asked if they were willing to participate in a research survey on attitudes about PrEP. Participants who expressed interest were invited to a private room for a verbal and written explanation of the study procedures by the research staff. They were informed that their participation in the study would have no effect on their healthcare and their providers would not have access to study-related data. All participants underwent an informed consent process.

The adapted questionnaire and socio-demographic questions were given to 100 participants who agreed to participate in the study. The surveys were administered electronically via Qualtrics (Provo, UT, USA) on an in-room computer, laptop, or iPad in the exam room prior to their provider visit. Participants were compensated $30 for completing the survey and received an additional $10 for a blood draw to measure their plasma tenofovir level. This study protocol was approved by the Institutional Review Board of the University of Alabama at Birmingham (Protocol # IRB=170414005).

Measures

PrEP-Related Stigma Dimensions

All PrEP-related stigma items were adapted from existing, validated HIV stigma measures [20, 21, 33, 34] and the qualitative interview data collected in our study [32]. Unless otherwise specified, participants responded to items using a 4-point Likert scale. For each scale, composite scores were calculated using the mean of the items.

Internalized PrEP Stigma Participants responded to 8 items, including: “I feel ashamed of being on PrEP,” and “Being on PrEP makes me think I’m a slut.”

Perceived PrEP Stigma in the Community Participants responded to 14 items (e.g., “Most people look down on people who are on PrEP” and “Most people think PrEP is a green light to have as much sex as you want.”).

Experienced PrEP Stigma Participants responded to 19 items assessing their experiences of being treated differently because they were taking PrEP for HIV prevention. Example items included, “Some family members have looked down on me because I am on PrEP,” and “Someone who learned I was on PrEP assumed I was sleeping around.”

Anticipated PrEP Stigma Seventeen items assessed participants’ perceived likeliness they would be treated differently if others found out they were taking PrEP. Example items included, “Sexual or romantic partners will look down on me if they learn I am taking PrEP,” and “Some people will assume I am sleeping around if they find out I am on PrEP.” Participants responded using a 5-point scale from 1 = Very unlikely to 5 = Very likely.

PrEP-related Disclosure Concerns Participants responded to 16 items regarding concerns about others’ learning they were taking PrEP for HIV prevention. Items included, “I would not tell others I am taking PrEP because they would make assumptions about my sex life,” and “I regret having told someone I am taking PrEP.”

PrEP Adherence Outcomes

PrEP Adherence Self-efficacy Adherence self-efficacy was assessed using 11 items assessing perceived ability to start and maintain a PrEP treatment regimen, adapted for this study from a measure of ART treatment adherence self-efficacy [28]. Participants received a stem question, “How confident are you that you can do these things related to PrEP?” Items included, “Integrate PrEP into your daily routine,” and “take PrEP daily as prescribed.” For these measures, participants were asked to rate the responses on a scale of 1–10, with 1 being “cannot do at all” and 10 being “completely certain can do.”

Self-reported PrEP Adherence We adapted the one-item validated measure of ART adherence to assess PrEP adherence [35,36,37]. Participants were asked to rate how good a job they did taking PrEP as prescribed in the past 30 days using a 6-point scale from Very poor to Excellent. This single item can be used as a tool to establish thresholds for guiding patient-provider discussions about ART adherence in clinical settings [38]. Self-reported PrEP adherence was high in our sample, having a negative skew. Following the dichotomization strategy when assessing ART adherence with this measure, responses were dichotomized for analysis to reflect 0 = suboptimal adherence (any response less than “excellent” job at taking PrEP as prescribed) and 1 = optimal adherence (“excellent” job at taking PrEP as prescribed) [35].

Plasma TDF Levels Plasma TDF levels have been used to measure PrEP adherence in several previous trials [39,40,41]. Participants who volunteered for the optional blood draw had an additional 1 mL of blood drawn during their regular clinic lab work. Specimens were immediately centrifuged for 10 min to separate the plasma. The plasma was transferred to a polyprophylene vial with screw top cap labeled with the participant’s ID number and date/time of blood draw. Cryovials were frozen at – 20 °C until transferred to a lab for measurement of plasma TDF levels. Continuous TDF levels were measured for plasma samples with quantifiable levels. For analysis, TDF concentrations were dichotomized as > 40 ng/mL suggestive of recent consistent dosing and < 10 ng/mL suggestive of recent inconsistent PrEP dosing [40]. It is important to note that measurement of TDF levels are not a perfect measure of adherence given their half-life of 17 h [42].

Data Analysis

First, descriptive statistics for the sample were calculated. Second, exploratory factor analyses were performed to determine the factor structure of each PrEP-related stigma dimension. Next, separate regression analyses (binary logistic and linear regression) were used to examine the association of each PrEP-related stigma dimension with PrEP adherence self-efficacy, self-reported PrEP adherence, or plasma TDF levels. Lastly, mediation analyses were conducted to examine whether PrEP adherence self-efficacy mediates the association between PrEP-related stigma dimensions and self-reported PrEP adherence (or plasma TDF levels) using the PROCESS macro for SPSS. We utilized 95% percentile confidence intervals (CIs) and 2000 bootstrapping resamples in mediation analyses. In this procedure, if confidence interval does not include zero, the indirect effect is significant, suggesting mediation [43]. In all analyses, covariates included education, race, and time on PrEP.

Results

Descriptive statistics regarding demographics are presented in Table S2. The average age of the sample was 35.8 ± 11.4 years. The average duration on PrEP was 15.7 ± 13.1 months. Ninety-one participants identified as cisgender men (91%), with the remainder identifying either as cisgender women, transgender women, or another gender identity (n = 9). Sixty-six participants identified as white, 23 as black, 6 as Hispanic or Latinx, 1 as Asian, and 4 as multi-race. Regarding relationship status, 47 participants reported being sexually active, but not being in a dedicated relationship, 27 were in an exclusive sexual relationship with one partner, 13 were in a relationship and having sex with people outside of that relationship, and 8 were not sexually active. Seventeen participants reported having a sexual partner living with HIV. Regarding adherence, self-reported adherence was available for 99 participants and plasma TDF levels for 84 participants.

Factor Structure and Reliability of PrEP-Related Stigma Dimensions

A series of factor analyses were conducted to identify the factor structure of PrEP-related stigma dimensions (i.e., internalized, anticipated, experienced, perceived community stigma, and disclosure concerns). First, data for each stigma dimension were checked for suitability for factor analysis. Bartlett’s test of sphericity was significant (p < 0.001) for internalized [χ2(28) = 888.64], anticipated [χ2(136) = 1341.25], experienced [χ2(171) = 1470.13], perceived community stigma [χ2(91) = 831.54], and disclosure concerns [χ2(120) = 1119.02]. Additionally, Kaiser–Myer–Olkin (KMO) value was 0.86 for internalized stigma, 0.88 for anticipated stigma, 0.88 for experienced stigma, 0.89 for perceived community stigma, and 0.91 for disclosure concerns, suggesting suitability for factor analysis [44]. Next, we ran separate factor analyses to determine the number of factors to retain. The method of scree plots suggested single-factor solutions for all stigma dimensions. Results yielded single factors for all stigma scales that explained 73.90% of the variance for internalized stigma, 53.92% for anticipated stigma, 51.06% for experienced stigma, 51.15% for community stigma, and 55.54% for disclosure concerns. Factor loadings for items ranged between 0.77 and 0.96 for internalized stigma, 0.64 and 0.80 for anticipated stigma, 0.55 and 0.81 for experienced stigma, 0.50 and 0.80 for perceived community stigma, and 0.54 and 0.88 for disclosure concerns.

Next, internal consistencies of each dimension were calculated. Cronbach’s alpha coefficients were 0.94 for internalized stigma, 0.94 for anticipated stigma, 0.94 for experienced stigma, 0.92 for perceived community stigma, and 0.94 for disclosure concerns. Correlations between PrEP-related stigma dimensions and other study variables as well as descriptive statistics are shown in Table S3.

Associations of PrEP Stigma Dimensions with PrEP Adherence

To examine the association between PrEP-related stigma and variables related to adherence to PrEP (i.e., self-reported PrEP adherence, plasma TDF levels, and PrEP adherence self-efficacy), separate regression analyses (logistic or linear, controlling for covariates) were conducted for each stigma dimension. In the logistic regression analyses, only internalized PrEP stigma was significantly associated with self-reported PrEP adherence (B = − 1.10, AOR = 0.33, p = 0.045, CI [0.11, 0.97]). There may be other psychological mechanisms that help explain the relationships between other dimensions of stigma and PrEP adherence, in addition to PrEP adherence self-efficacy. Results of all regression analyses are presented in Table S4.

Exploratory Mediation Analyses: Indirect Associations Between PrEP Stigma Dimensions and PrEP Adherence Through Self-efficacy

We conducted a series of simple mediation models where the association between PrEP stigma dimensions (i.e., internalized, perceived, experienced, anticipated PrEP stigma, and disclosure concerns) and self-reported PrEP adherence is mediated by PrEP adherence self-efficacy. The indirect effect of internalized PrEP stigma on PrEP adherence through PrEP adherence self-efficacy was significant (B = − 0.48, SE = 0.41, CI [− 1.55, − 0.02]; see Fig. 1A). Next, we tested the mediating effect of PrEP adherence self-efficacy in the association between perceived PrEP stigma and PrEP adherence. This analysis yielded a significant indirect effect on PrEP adherence (B = − 0.30, SE = 0.27, CI [− 1.04, − 0.03]; see Fig. 1B). The indirect effect of experienced stigma on PrEP adherence through PrEP adherence self-efficacy was also significant (B = − 0.30, SE = 0.27, CI [− 1.04, − 0.03]; see Fig. 1C).

Fig. 1
figure 1

A–E The indirect effects of PrEP stigma dimensions on self-reported PrEP adherence through PrEP adherence self-efficacy. Note Associations are presented as path coefficients (unstandardized). aIt indicates the coefficient (direct effect) when the mediator is entered in the model. *p < 0.05, **p < 0.01, ***p < 0.001

Full size image

We also tested the mediating effect of PrEP adherence self-efficacy in the association between anticipated PrEP stigma and PrEP adherence. This analysis also yielded a significant indirect effect on PrEP adherence (B = − 0.22, SE = 0.21, CI [− 0.79, − 0.02]; see Fig. 1D). Lastly, in a simple mediation model depicted in Fig. 1E, the indirect effect of disclosure concerns on PrEP adherence through PrEP adherence self-efficacy was also significant (B = − 0.27, SE = 0.25, CI [− 0.95, − 0.03]). These results provide evidence for the mediating effect of adherence self-efficacy in the association between all PrEP stigma dimensions and self-reported PrEP adherence.

Discussion

In this study, we developed and administered measures of PrEP stigma dimensions to a cohort of 100 current PrEP users in Birmingham, Alabama. Factor analyses confirmed the unidimensional structure of each stigma dimension and reliability analyses suggested good internal consistency for each. We also aimed to understand the association between PrEP-related stigma and adherence to PrEP using these newly developed measures and found only internalized PrEP stigma to be associated with low self-reported PrEP adherence. No measures of PrEP stigma were associated with low PrEP adherence measured with TDF levels. We used exploratory mediation analyses to assess whether the relationships between each measure of PrEP stigma and self-reported PrEP adherence could be explained by PrEP adherence self-efficacy and found significant indirect effects between each PrEP stigma measure (internalized, perceived, experienced, anticipated, and disclosure concerns) and self-reported PrEP adherence.

Our data aligns with findings in the HIV stigma literature that internalized stigma has the greatest impact on adherence outcomes [20, 23]. Applying negative societal beliefs to oneself about a health prevention tool like PrEP may undermine daily decisions to take one’s medication as prescribed. However, there may be psychological mechanisms that help explain the relationships between other dimensions of stigma and PrEP adherence, like PrEP adherence self-efficacy. There are limited instruments available to measure the nuanced dynamics of PrEP stigma as it is experienced by current users. Further validation in larger and more diverse samples with this instrument could provide a key tool for investigating the impact of PrEP stigma on PrEP uptake and adherence.

PrEP adherence self-efficacy as a potentially important mechanism in the association between each dimension of PrEP stigma and self-reported PrEP adherence is crucial to consider. Self-efficacy is also known to be a strong predictor of antiretroviral adherence in PLHW [26]. Two of the currently FDA-approved PrEP drugs that are available are daily oral pills [4, 5]. However, many people have difficulty taking medications daily [45]. Broader psychosocial and structural factors such as poverty, substance use, and unstable housing may also be involved in this relationship and were not investigated in this study. Innovations in intermittent PrEP dosing and long-acting injectable PrEP options will be important tools for some patients as less frequent dosing intervals may empower patients to feel that they can adhere to these medications.

Our data emphasize the importance of not only considering PrEP stigma in HIV prevention efforts but finding methodologically sound ways to measure PrEP-related stigma. Existing literature suggests that PrEP stigma is a major driver of disparities in PrEP access [12], thus preventing optimal implementation of PrEP in many settings [46]. A growing body of qualitative work in this area points to a need to develop interventions to reduce PrEP stigma, but in order to execute such interventions, we must be able to accurately measure PrEP stigma [47]. Existing studies suggest that PrEP stigma is a problematic experience for people accessing PrEP, but few examine different mechanisms of PrEP stigma on adherence among people using PrEP. Our data suggest that PrEP users in Birmingham do experience stigma, but it is internalized PrEP stigma that potentially has the greatest impact on adherence to PrEP. This is in keeping with the fact that internalized stigma has also been identified in the HIV literature to be a major predictor of poor adherence among PLWH compared to other stigma dimensions [20, 48].

One limitation of our study was the lack of racial and gender diversity represented. Most of our sample were white cisgender men who had been using PrEP for a year or more. In the Southeastern U.S., HIV most dramatically impacts racial and ethnic minorities; 74% of diagnoses occurred in Black/African-American or Hispanic/Latinx individuals in 2018 [49]. Nationwide, sexual and gender minorities (SGM) are also disproportionately impacted by HIV, with MSM and TGW experiencing higher rates than their counterparts [50,51,52]. Future studies are needed administering this instrument to a larger and more representative cohort of eligible PrEP users, including Black and transgender individuals. Of note, our sample seems to be quite representative of current PrEP users, emphasizing the fact that PrEP is underutilized among these marginalized populations. Measuring adherence was another challenge in our study. Self-reported adherence by participants can be subject to reporting bias [53]. Measurement of plasma TDF levels is more objective, but is limited as a measure of adherence by the fact that it has a half-life of 17 h and, therefore, may not be reliably detected even if someone is adherent if the sample is not collected in a timely manner [42]. In addition, measurement of plasma TDF levels do not capture the true dynamics of adherence among patients. For example, many may take their PrEP right before a clinic visit (which coincided with our research visit that included blood samples to assess TDF levels) because of a desire to present themselves to clinicians as adherent and actually demonstrate less adherence between visits, leading to other ways the accuracy of plasma levels could be diminished. All analyses reported are cross-sectional and should not be interpreted to imply causality. Additional research is needed to validate these measures as potential predictors of retention in PrEP care and longitudinal adherence in the context of HIV risk.

Conclusions

In conclusion, among PrEP users in the Southeastern U.S., internalized stigma is associated with lower self-reported PrEP adherence. PrEP adherence self-efficacy may play a role in this relationship such that the various dimensions of PrEP stigma may diminish self-efficacy and thereby PrEP adherence. Additional longitudinal research is needed on these relationships. No dimensions of PrEP stigma were found in this study to influence PrEP adherence as measured by plasma TDF levels. This sample lacked racial and gender identity diversity and while it represents the demographics of current PrEP users in the Southeastern U.S., more research is needed to investigate how PrEP-related stigma dimensions impact adherence among PrEP users, especially those from racial and gender minorities with less experience with PrEP.