Longitudinal predictors of post-exposure prophylaxis awareness among latino sexual minority men in South Florida

Abstract

Biomedical tools for HIV prevention such as post-exposure prophylaxis (PEP) continue to be underutilized by subgroups experiencing significant HIV inequities. Specifically, factors associated with both PEP awareness and uptake both cross-sectionally and longitudinally are under-researched, despite PEP being a part of the United States’ Plan for Ending the HIV Epidemic. The current study examined longitudinal predictors of PEP awareness among Latino sexual minority men (LSMM) living in South Florida. This current study (N = 290) employed hierarchal linear modeling across three timepoints (baseline, 4-months, 8-months) to assess within-person and between-person effects over time for several psychosocial and structural factors. Most participants (67.5%) reported little to no awareness of PEP at baseline with general PEP awareness growing slightly across the study (60.5% reporting little to no awareness of PEP at 8 months). Results of the final conditional model suggest significant within-person effects of PrEP knowledge (p = 0.02) and PrEP self-efficacy (p < 0.001), as well as a significant positive between-person effect of PrEP knowledge (p < 0.01) on PEP awareness. Between-person HIV knowledge was also a significant predictor in this model (p = 0.01). This longitudinal analysis of LSMM’s PEP awareness indicates that more must be done to increase PEP awareness among this subgroup. Future studies should explore how to build on existing interventions focused on HIV and PrEP knowledge and PrEP self-efficacy to incorporate information about PEP to increase the reach of this effective biomedical HIV prevention tool.

Introduction

HIV continues to be a major public health concern both domestically and internationally. More than 1.1 million people in the United States (U.S.) are living with HIV, with HIV incidence averaging around 35,000 cases over the past few years (CDC, 2020). Although improvements to antiretroviral therapy (ART) have been the primary drivers for managing the HIV epidemic over the past two decades, the development of biomedical HIV prevention tools like post-exposure prophylaxis (PEP) and pre-exposure prophylaxis (PrEP), have also contributed to decreasing HIV incidence more recently (CDC, 2020). Despite their efficacy, PEP and PrEP continue to inadequately reach key populations affected by HIV disparities (e.g., racial, ethnic, and sexual minorities) (Walters et al., 2017). Underscoring their importance, biomedical prevention interventions like PEP and PrEP are key parts of the national Ending the HIV Epidemic (EHE) plan (Fauci et al., 2019). Therefore, it is essential to increase awareness, and in turn, reach of biomedical HIV prevention tools (e.g., PEP) among groups currently experiencing disparities.

HIV incidence among certain groups who have historically experienced HIV-related inequities have begun to level out since the implementation of the EHE plan (CDC, 2022). Although incidence has decreased among sexual minority men (SMM; i.e., gay, bisexual, and men who have sex with men) overall in recent years, SMM continue to bear the brunt of new U.S. HIV diagnoses. For instance, SMM experienced more than 70% of all new HIV diagnoses in 2020, more than double that of their heterosexual counterparts (CDC, 2022). Furthermore, for SMM living at the intersection of other marginalized identities (e.g., race and ethnicity) HIV related disparities are exacerbated (Harkness et al., 2021; Singer et al., 2017; CDC, 2022). For example, Latino sexual minority men (LSMM) experience higher HIV incidence − 23% of new HIV diagnoses in 2019 were among LSMM (HIV Surveillance | Reports| Resource Library | HIV/AIDS | CDC, 2021), less engagement with pre-exposure prophylaxis (PrEP) for HIV prevention (Smith et al., 2015), and less frequent HIV testing (Painter et al., 2019) than their White SMM peers.

Structural and psychosocial barriers rooted in minority stress worsen HIV-related health outcomes for SMM. According to the minority stress theory, SMM experience sexual orientation-based stressors related to discrimination, prejudice, hostility, and rejection that lead to mental and physical health disparities (Brooks, 1981; Meyer, 1995). For instance, prior research has found that SMM who experienced sexual orientation-based prejudice are more likely to report poorer mental and physical health (Frost et al., 2015; Zheng et al., 2020). For racially/ethnically marginalized SMM living at the intersection of multiple systems of oppression, these unique stressors may be magnified which, in turn, may amplify health disparities (Crenshaw, 1991). This intersectional minority stress is especially apparent for LSMM living in Miami, Florida (Harkness et al., 2022a) and may be a contributing factor of the uncontrolled HIV epidemic in this region (AIDSVU Report for Miami-Dade County, 2022). Therefore, given the cumulative impact of intersectional minority stress, it is imperative to employ more active and engaged participatory data science approaches that center the lived experiences of LSMM as researchers develop, implement, and maintain culturally grounded HIV prevention interventions.

PEP is an effective tool to reduce HIV incidence and could mitigate health disparities if scaled up alongside other forms of HIV prevention. PEP consists of a 28-day course of ART for people who have already been exposed to HIV and, when taken within 72 h of exposure, can reduce HIV acquisition by almost 80% (Cardo et al., 1997; DeHaan et al., 2021; Dominguez et al., 2018). As an evidence-based tool for HIV prevention identified in EHE national priorities, PEP may be particularly valuable for subgroups at increased risk for HIV acquisition who are not currently engaged in other forms of HIV prevention (e.g., PrEP, consistent condom use). Although less common than PrEP given its indication for urgent situations in which an exposure to HIV has already occurred, PEP may be particularly useful for people who are not currently taking PrEP, especially in situations related to sexual violence (e.g., rape), condom failures, or condomless sex in the context of concurrent substance use. For example, although individuals may not take PrEP due to low perceived risk of HIV or feeling as if they are not a good candidate for daily PrEP due to engaging in only isolated instances of high HIV-risk activity (e.g., condomless sex with someone with unknown HIV status), they still may become a PEP candidate in the context of an exposure. Despite its relevance and effectiveness, PEP awareness is low among MSM in general (25–59%; Jin et al., 2022; Walters et al., 2017) as well as among subgroups experiencing HIV disparities, including 33% among LSMM (Weinstein et al., 2022), 63% among transwomen (Koblin et al., 2018), and 34% among cisgender women of color (Koblin et al., 2018). PEP utilization is even lower than PEP awareness with data indicating that between 5 and 13% of people with increased susceptibility to HIV had ever used PEP (Jin et al., 2022; Koblin et al., 2018). Similarly, PEP awareness and prescribing among both HIV and non-HIV related providers is low with one study indicating that 12.5% of providers were unaware of PEP and less than 50% of prescribers have a history of recommending PEP to patients with heighted HIV risk exposure (John et al., 2020). Even though MSM have expressed specific interest in utilizing PEP for HIV prevention (Dolezal et al., 2015), research on determinants of PEP awareness and use is lacking, particularly among LSMM. Thus, without scientific knowledge regarding awareness, it is difficult to know how to intervene to improve PEP’s awareness and ultimately, reach.

Even though PEP has been approved as a biomedical intervention for HIV prevention by the Federal Drug Administration (FDA) for over 15 years, there is minimal literature focused on factors associated with PEP utilization and awareness, particularly among groups with more HIV-related inequities. Theories of behavior change, such as the Social Cognitive Theory, suggest that cognitive factors, like knowledge and self-efficacy, are essential components of engaging in a health behavior (Bandura, 1986). Therefore, in the realm of HIV prevention, increasing an individual’s awareness of PEP is a necessary first step in improving engagement (Cohen et al., 2015). Some of the most salient factors associated with PEP awareness are related to overall HIV awareness and more engagement with other forms of HIV prevention including increased HIV testing (Prati et al., 2016) and a history of current or past PrEP use (Weinstein et al., 2022). Additionally, greater PrEP knowledge, HIV knowledge, and PrEP self-efficacy have all been cited as important predictors in overall PEP awareness, particularly among LSMM (Weinstein et al., 2022). Furthermore, factors that either perpetuate (e.g., HIV stigma) (Jin et al., 2022; Prati et al., 2016) or alleviate intersectional minority stress (e.g., including involvement with HIV/AIDS organizations, being a part of an interconnected community) (Koblin et al., 2018; Prati et al., 2016) are also significantly related to PEP awareness among groups affected by HIV.

The HIV epidemic among LSMM living in South Florida is unique and complex. Compared to other EHE jurisdictions in the continental U.S., South Florida has significant ethnic/racial diversity, a growing epidemic of substance use, high immigration rates, and low health-related resources compared to other regions of similar size (Diaz-Martinez et al., 2021; “Mental Health and Substance Use State Fact Sheets,” n.d.; Miami Matters, 2020). This combination of factors makes addressing the HIV epidemic in South Florida challenging, particularly for LSMM. For example, more than 65% of all new HIV diagnoses recorded in 2020 in Miami-Dade County, FL were among Hispanic/Latinx individuals (Florida Department of Health in Miami-Dade, 2019), a number that exceeds other regions with similar HIV incidence (e.g., Atlanta, GA – 10.8%, Baton Rouge, LA – 1.6%) across the nation (AIDSVu Share Map, 2022).2022 Furthermore, South Florida’s significant income inequality uniquely contributes to structural barriers (e.g., inconsistent health care access, failure to expand Medicaid) faced by people with HIV (PWH) in the region (Foster & Lu, 2018; Miami Matters, 2020). Given the complex and unique drivers of the HIV epidemic in South Florida, combined with its high HIV incidence among LSMM, there is an urgent need to enhance the reach of all evidence-based HIV prevention tools, including but not limited to PEP, to LSMM living in South Florida.

Despite the contributions of the limited cross-sectional research on PEP use among SMM, longitudinal research that is guided by a participatory data science approach (i.e., including individuals with lived experience in the data analysis process) (Research Priorities | NIH Office of AIDS Research, 2022; Weinstein et al., 2022) is needed to adequately address the HIV epidemic among LSMM in South Florida. Although longitudinal research has strengthened PrEP awareness research among key subgroups (Holloway et al., 2020; Mosley et al., 2018), to the authors’ knowledge, longitudinal work on PEP awareness has not yet been conducted. Given the similarities between PrEP and PEP and the fact that both make unique contributions to the EHE plan, it is important to fill this gap in the literature, particularly as it relates to potentially changing trends in PEP awareness among LSMM, an underserved group within HIV prevention research. Thus, the aims of this study were to examine whether factors associated with PEP awareness cross-sectionally may also predict PEP awareness longitudinally among LSMM living in South Florida. A longitudinal examination of PEP awareness informed by the lived experiences of LSMM will provide key information on the degree to which cross-sectional predictors (e.g., PrEP knowledge, history of PrEP use) hold over time.

Methods

Participants and procedures

This secondary analysis included LSMM (N = 290) enrolled in the DIMELO Study, an 8-month longitudinal cohort study investigating engagement in biomedical HIV-prevention and behavioral health treatment services (Harkness et al., 2021, 2022b). All participants completed their baseline assessments between February and September 2020 with the final 8-month assessment finishing in May 2021. Of the 290 LSMM who completed a baseline assessment, 261 completed a 4-month assessment, and 244 completed an 8-month assessment.

Eligible participants in the parent DIMELO Study (1) identified as Latino/Hispanic, (2) reported being a SMM (i.e., gay, bisexual, or man who has sex with men), (3) were between the ages of 18 and 60 years old, (4) lived in the greater Miami, FL area, and (5) self-reported negative or unknown HIV status at the time of their baseline assessment. There were no inclusion criteria related to sexual behavior. Participant recruitment was achieved via both active (e.g., “consent-to-contact” database, community venues and events,) and passive (e.g., listservs, social media, and snowball recruitment) methods. More information on study recruitment and retention methods can be found elsewhere (Harkness et al., 2021, 2022b). All eligible LSMM reviewed consent information prior to clicking a box on REDCap indicating their consent to participate. All assessments were available in both English and Spanish. The Institutional Review Board at the University of Miami approved all study procedures prior to data collection.

At each of the 3 timepoints (baseline, 4-month, and 8-month), participants completed an online, REDCap administered self-report assessment. Measures were either selected from the published literature or created by the senior author based on prior qualitative findings as well as relevant theories that guided the parent study: syndemic, minority stress, and intersectionality (Brooks, 1981; Crenshaw, 1991; Harkness et al., 2021; Meyer, 1995; Singer et al., 2017). Participants completed a wide range of measures; however, only the factors associated with PEP awareness in our prior cross-sectional study (Weinstein et al., 2022) were included in this longitudinal analysis.

We highlight some information about the selection of variables from the prior cross sectional work here. Specifically, the variables we previously evaluated in relation to PEP awareness in our cross-sectional work were selected based on: (1) a review of the available literature on PEP awareness, (2) stochastic search variable selection, a Bayesian variable selection approach, and (3) a participatory data science approach. The participatory component of variable selection involved members of our team’s community advisory board of LSMM reviewing all potential variables in the parent dataset and, based on their lived experience, collectively identifying the variables they believed would be most relevant to PEP awareness among LSMM. Complete details about this process are described in our cross-sectional manuscript (Weinstein et al., 2022).

Study measures

Demographics

Items from the CLaRO Measures Library and the Center for HIV and Research in Mental Health (CHARM) community survey were selected for the demographic questionnaire (e.g., age, gender, sexual orientation) completed by participants at the baseline assessment (Mitrani et al., 2017). All demographic variables were assessed at baseline.

HIV Knowledge

The 10-item HIV knowledge scale was administered across timepoints to assess participants’ overall HIV-related knowledge (Carey & Schroder, 2002; Feinstein et al., 2017). This true/false scale had response options including true, false, and don’t know. Correct responses were summed, such that higher scores reflected greater HIV-related knowledge. Example items include “A person can get HIV from oral sex” and “There is a vaccine that can stop adults from getting HIV.” The HIV Knowledge scale had adequate internal consistency at all three timepoints (Baseline: α = 0.66; 4-Month: α = 0.72; 8-Month: α = 0.72).

PrEP knowledge

The 13-item PrEP Knowledge Scale was administered across all timepoints to assess participants’ PrEP knowledge (Walsh, 2019). Participants responded to items with “true” or “false” with one point awarded for each correct answer, so that when correct responses were summed, higher scores reflected greater PrEP-related knowledge. Example items include “PrEP is a daily pill you can take to reduce your risk of becoming infected with HIV” and “You should not use PrEP if you don’t know your HIV status.” The PrEP Knowledge Scale had high internal consistency at all three timepoints (Baseline: α = 0.84; 4-Month: α = 0.83; 8-Month: α = 0.81).

PrEP use

Participants were asked to report if they had ever taken PrEP. For those who selected yes, they could indicate if they were currently taking PrEP or if they had taken it in the past. We created a binary “PrEP use” variable which was coded as “0 – no” or “1 – yes” and reflected whether a participant was currently using or had ever used PrEP in the past (yes) or never used PrEP before (no).

Perceived norms about HIV testing

Perceived community norms about HIV testing was assessed by participants responding to one question, “Approximately how many people you know have had a recent HIV test (recent means in the past 6 months)?” at all time-points (Andrinopoulos & Hembling, 2013). LSMM reporting “all or almost all” or “half” were coded as having high perceived HIV testing norms while those who responded “don’t know,” “none,” or “few” were coded as having low perceived norms.

PrEP self-efficacy

An 8-item PrEP Self-Efficacy Scale evaluated participants’ general self-efficacy for engaging with PrEP at all three timepoints (Walsh, 2019). Participants rated each item using a 5-point Likert-type scale ranging from 1 “Strongly Disagree” to 5 “Strongly Agree” with higher mean scores reflecting more self-efficacy. Example items included “how difficult would it be for you to talk with your sexual partner(s) about the decision to take PrEP?” and “how difficult would it be for you to seek out more information about PrEP to decide if it is right for you?” The PrEP Self-Efficacy Scale had excellent internal consistency at all three timepoints (Baseline: α = 0.83; 4-Month: α = 0.84; 8-Month: α = 0.87).

Identity affirmation

The 3-item affirmation subscale from the Lesbian, Gay, and Bisexual Identity Scale assessed LSMM’s degree of self-affirmation in relation to their sexual orientation across the study period (Mohr & Kendra, 2011; Vinces-Guillén, 2016). LSMM responded to the following statements – “I am proud to be LGB,” “I am glad to be a LGB person,” and “I’m proud to be a part of the LGB community,” using a 6-point-likert scale from 1 “disagree strongly” to 6 “agree strongly.” The identity affirmation subscale had exceptional internal consistency at all three timepoints (Baseline: α = 0.95; 4-Month: α = 0.95; 8-Month: α = 0.95).

PEP awareness

The main outcome for this secondary analysis was PEP awareness among LSMM. At each timepoint, participants reported PEP awareness, with the following response options: “1- I’ve never heard of PEP before today,” “2- I’ve heard about it, but I don’t really know what it is,” “3- I know a little bit about it,” “4 - I know a fair amount about it,” and “5 - I know a lot about it.” This item was developed by the senior author for the parent study.

Data analytic plan

Hierarchal linear modeling (HLM) was employed to examine longitudinal predictors of PEP awareness among LSMM. Overall, HLM allows researchers to assess both within-person and between-person effects over time, which is one of the advantages of a longitudinal analysis compared to a cross-sectional analysis. Within-person effects reflect changes within an individual across timepoints (e.g., whether an individual is reporting greater or less than their average value for that predictor at that specific timepoint). Between-person effects reflect differences between people over the course of the study (e.g., whether an individual reports an average value that is greater or less than the sample average for that predictor). Cross-sectional analyses are only able to assess whether “predictor” and “outcome” variables are associated with one another at a single point in time; and as such, cross sectional findings are often a starting point for exploring relationships among variables, whereas longitudinal analyses are typically more robust and replicable (Das, 2014). This is particularly important in the realm of HIV prevention research since understanding how certain predictors that facilitate engagement of services may change over time may directly impact the way in which interventions are designed and implemented, specifically for subgroups at increased susceptibility for acquiring HIV. For all variables, data from each of the three timepoints (Baseline, 4-month, 8-month) were used in HLM analyses.

Prior to conducting any HLM analyses, descriptive statistics were calculated for participant demographics and study variables at baseline. Due to the study’s longitudinal design, HLM was employed to account for the non-independence of data. All analyses were run using the lmer package in R version 3.5.0. All person-level continuous predictors were grand mean centered. Prior to running any analyses, we assessed for assumptions of normality of residuals, homogeneity of variance, and homoscedasticity via a visual inspection of residual plots. All relevant assumptions were met. We also tested the intercorrelation between predictor variables and found no evidence of multicollinearity (all variance inflation factors < 5).

Model-building procedure

We first tested an unconditional intercept-only model to identify the intraclass correlation at the person (Level 2) and time (Level 1) levels. We then tested two unconditional growth models that included the fixed and random effects of time. The random effect of time was retained because a significant likelihood ratio test suggested that it demonstrated better model fit than the model with the fixed effect of time. A bottom-up model building approach was then used to evaluate whether PEP awareness among LSMM changed over time and determine the predictors of PEP awareness. As described above, variables that had a significant cross-sectional relationship with PEP awareness in our prior study (Weinstein et al., 2022) were added iteratively to build a final longitudinal model. Prior to analysis, we created two variables for each predictor to parse apart the within and between effects of our predictors on PEP awareness: a variable centered within-person (Level 1) and a separate person-average variable centered between-person (Level 2). This allowed us to separately model the between- and within-person effects of each potential predictor on the outcome (Hoffman & Stawski, 2009). All Level 2 predictors were mean-centered so that they represented higher or lower values compared to the sample average.

Results

Descriptive statistics

Participants (N = 290 at baseline) ranged from 18 to 60 years old (M = 32 years, SD = 8.32 years) and most identified as gay (85.2%). A slight majority of participants were born outside of the continental U.S. (52.7%). Almost half of participants were sexually active, and 42.4% had any history of PrEP use. At baseline, most participants (67.5%) reported having little to no awareness about PEP with this number slightly decreasing to 65.1% at 4-months and 60.5% at 8-months post baseline. Additional participant demographics, including citizenship status, preferred language, education level, financial status, relationship status, health insurance, and mental health concerns, can be found elsewhere (Weinstein et al., 2022).

Unconditional model

The unconditional model demonstrated a high intraclass correlation coefficient (ICC). The high ICC (0.69) means that 69% of the variability in PEP awareness was explained by the variables assessed at the between-person level. In other words, the between-person differences amongst the predictors assessed corresponded with 69% of the total variability in PEP awareness. The unconditional growth model, or the most basic model examining the effect of time without the influence of predictors, indicated a significant difference in PEP awareness across time (β = 0.07, 95% CI [0.04, 0.11], p < 0.001) suggesting that LSMM’s PEP awareness increased over the three timepoints. A random effect of time was retained from the unconditional growth model (see Table 1).

Table 1.

Fixed effect coefficients for the overall model predicting PEP awareness (N = 290)

	Unconditional Model			Unconditional Growth Model			Final Model
Parameters	Estimate	Standard Error	Std Estimate (β)	Estimate	Standard Error	Std Estimate (β)	Estimate	Standard Error	Std Estimate (β)
Regression coefficients (fixed effects)
Intercept (γ 00 )	3.01 **	0.07	−0.009	2.79 **	0.09	−0.006	2.87 **	0.09	0.007
Time (γ10)				0.12 **	0.03	0.07	0.05	0.03	0.03
PrEP Knowledge – Within Person (γ 20 )							0.05 *	0.02	0.05
PrEP Self-Efficacy – Within Person (γ 30 )							0.39 **	0.09	0.09
HIV Knowledge – Within Person (γ40)							0.01	0.03	0.008
HIV Testing Norms – Within Person (γ50)							−0.002	0.03	−0.001
LGB Identity Affirmation – Within Person (γ60)							−0.05	0.04	−0.03
Lifetime PrEP Usage (γ01)							0.03	0.11	0.01
PrEP Knowledge – Between Person (γ 02 )							0.21 **	0.02	0.49
PrEP Self-Efficacy – Between Person (γ03)							0.04	0.11	0.02
HIV Knowledge – Between Person (γ 04 )							0.10 *	0.04	0.12
HIV Testing Norms – Between Person (γ05)							0.10+	0.05	0.08
LGB Identity Affirmation – Between Person (γ06)							−0.06	0.05	−0.05
Variance components (random effects)
Residual (eti)	0.50	0.71	--	0.41	0.64	--	0.39	0.62	--
Intercept (μ0i)	1.12	1.06	--	1.50	1.22	--	0.78	0.88	--
Time (μ1i)	--	--	--	0.08	0.29	--	0.06	0.25	--

Note Outcome for this model is PEP awareness. All Within Person variables are group-mean centered so that scores represent each participant’s deviation from their individual average for that variable. All three Between Person variables are sample-mean centered so that the scores represent each participant’s average deviation from the overall mean score of all participants in the sample. PrEP Usage is coded as a binary variable (0 = No, 1 = Yes). Greek notation reflects variable in multilevel model equations, which can be found in the Supplemental Materials section.

⁺p < 0.10,

^*p < 0.05,

^**p < 0.01

Final model predicting PEP awareness

The results of the final model that included all predictors are presented in Table 1. There were significant within-person effects of PrEP knowledge (p = 0.02) and PrEP self-efficacy (p < 0.001), such that at timepoints that participants reported greater PrEP knowledge or greater PrEP self-efficacy than their individual average across all timepoints, they also reported higher PEP awareness at that timepoint (4-month or 8-month post baseline). There was also a significant positive between-person effect of PrEP knowledge (i.e., each person’s mean PrEP knowledge across the three time points), such that those with higher average PrEP knowledge reported greater levels of PEP awareness (p < 0.001) compared to those with lower PrEP knowledge. Similarly, between-person HIV knowledge was a significant predictor in this model (p = 0.01), such that those with greater average HIV knowledge reported greater levels of PEP awareness than those with lower HIV knowledge. Finally, there was a between-person effect of perceived norms which trended towards significance (p = 0.07), such that those with high perceived norms about HIV testing reported higher PEP awareness than those with low perceived HIV testing norms. There were no other significant effects of any predictors on PEP awareness over time (all p’s > 0.05) (See Table 1).

Discussion

To our knowledge, this study is one of the first to examine predictors of PEP awareness longitudinally among LSMM in South Florida, a priority jurisdiction of the EHE plan. This study leveraged a participatory data science approach from our prior cross-sectional research to examine factors associated with PEP awareness that are salient to LSMM via their lived experiences (see Table 2 for variables significant across the the two studies). Major results of this study found that: (1) when individuals’ PrEP knowledge and PrEP self-efficacy were greater than their personal average, there was a corresponding increase in their PEP awareness (i.e., within-person effects) and (2) individuals were more aware of PEP if they had higher average PrEP knowledge and higher average HIV knowledge compared to their peers (i.e., between-person effects). Furthermore, we also found a significant random effect of time such that LSMM’s individual trajectories of PEP awareness varied over time and that the effect of LSMM individual’s PrEP self-efficacy trajectories on PEP awareness was also significant.

Table 2.

Variable selection sources, mapped to cross-sectional and longitudinal results

Variable	Variable Selection Source			Results
	Literature Review	SSVS	CAB	Cross-Sectional	Longitudinal
PrEP knowledge (objective)		✔		✔	✔
PrEP self-efficacy		✔	✔	✔	✔
HIV knowledge		✔		✔	✔
Identity affirmation		✔		✔
High perceived community norms for		✔
HIV-testing
Anticipated HIV stigma	✔		✔
Anticipated risk for acquiring HIV			✔
Sexual activity			✔
Problem solving			✔
Age		✔

The first set of findings revealed that individual changes in PrEP knowledge and PrEP self-efficacy were significantly associated with higher PEP awareness. This suggests that when LSMM had higher PrEP knowledge and PrEP self-efficacy than their own average level of knowledge or self-efficacy, they also reported higher PEP awareness. As LSMM continue to gain knowledge related to PrEP and other HIV prevention behaviors in general, it is likely that they will also learn about other HIV prevention tools such as PEP through those same psychoeducational mechanisms. Furthermore, if LSMM feel more confident in their ability to navigate PrEP, they may have a corresponding increase in PEP awareness (Weinstein et al., 2022). Moreover, it may be that those who are engaged in HIV-prevention, or who have had access to HIV-related education, are aware of different prevention tools such as PEP. Alternatively, PrEP self-efficacy might serve as a proxy for PEP self-efficacy (the two constructs might be related), which could explain the relationship between PrEP self-efficacy and PEP awareness. PrEP knowledge and self-efficacy may be modifiable leverage points to target in interventions and implementation strategies to increase PEP awareness and uptake. This also suggests that increasing engagement in HIV prevention in general may indirectly improve awareness of PEP among LSMM. As an example, when initially engaging LSMM in HIV prevention, campaigns and programs can concurrently focus on both PrEP and PEP to provide LSMM with multiple options for prevention so they can make an informed choice about the biomedical prevention approach that is best suited to their needs.

Additionally, LSMM with higher average PrEP knowledge and higher average HIV knowledge also reported higher average PEP awareness compared to their peers with lower average PrEP and HIV knowledge. LSMM who are knowledgeable about HIV prevention and PrEP more broadly, may also be well-versed in other HIV prevention tools such as PEP because they are already engaged in HIV-prevention. Supported by theories of health behavior change such as the Health Belief Model (Becker, 1974) and Theory of Planned Behavior (Fishbein & Ajzen, 1975), it is possible that LSMM with both HIV and PrEP knowledge may be more willing to learn about or adopt other HIV prevention innovations such as PEP due to their greater overall knowledge base and prior engagement in HIV-prevention (Bond & Ramos, 2019; Rogers et al., 2020). Alternatively, it may be that existing HIV prevention campaigns address HIV, PrEP, and PEP, and lead to concurrent increases in knowledge of all three topic. This, highlightsg the importance of engaging LSMM in all HIV prevention services and HIV prevention and PrEP campaigns that do not already incorporate PEP should begin to do so to promote awareness.

Although not statistically significant, the between-person effects of HIV testing norms on PEP awareness trended towards significance in our final model and was significant in our prior cross-sectional research (Weinstein et al., 2022). This finding may suggest that peer normalization of HIV prevention could be another leverage point for increasing PEP awareness among LSMM. As outlined in the Theory of Planned Behavior (Fishbein & Ajzen, 1975), community norms can play a significant role in fostering behavior change. Therefore, for LSMM in our study who reported greater than average HIV testing norms in their social networks, it is possible that frequent HIV testing within one’s social network could help to create a diffused sense of PEP awareness within that network. Furthermore, as suggested by our collaborating community advisory board (CAB), future HIV prevention interventions should capitalize on peer influence in LSMM social networks (Bertrand, 2004) in facilitating behavior change by promoting programming like “peer ambassadors” or community-wide public service announcements that normalize HIV testing and even past PEP use within LSMM communities which may in turn, bolster PEP awareness (Jaramillo et al., 2021).

In addition, there was a random effect of time within our model suggesting that trajectories of individual LSMM’s PEP awareness appeared to vary across individuals. This finding is intriguing, particularly given that the fixed effect of time was not significant after including other predictors in the model. As indicated by this finding, it is possible that there are certain subgroups of LSMM who may benefit from PEP as a form or HIV prevention more than others. For example, PEP may be especially relevant in meeting the needs of LSMM who (a) are hesitant to engage with PrEP, (b) experience condom failures, or (c) are victims of sexual coercion or assault. Therefore, future studies should explore using alternative methods, such as latent class analyses, to explore not only what predictors may contribute to the different trajectories of PEP awareness over time, but also which subgroups of LSMM may be most benefited by increased awareness of PEP.

It is important to note that there were several counter-intuitive findings observed within this study. First, there was no significant fixed effect of time after including other predictors in the model. This is surprising given the fact that participants were responding to HIV prevention questions across all assessment times which could have reasonably increased PEP awareness across time above and beyond all other potential predictors. Second, PrEP usage was not significantly associated with LSMM’s PEP awareness across the 8-month study period. This finding is curious given the significant results we detected regarding PrEP self-efficacy and knowledge. Prior literature suggests that there is a strong relationship between PrEP self-efficacy and knowledge and PrEP usage among sexual minority men (Harkness et al., 2021); therefore, it is unexpected that we did not observe a significant association between PrEP usage and PEP awareness over time. CAB members suggested that one potential reason for this non-significant finding could be that if an individual is already using PrEP (i.e., 42.4% of sample who reported any current or prior PrEP use), they would not need PEP since the two biomedical prevention tools cannot both be used in tandem. In this scenario, a person may not become aware of PEP because they are already protecting themselves against HIV acquisition and are no longer looking for and/or being educated about additional biomedical tools for HIV prevention. Furthermore, CAB members suggested that other factors may be more influential in influencing PEP awareness among LSMM over time than solely whether they have ever used PrEP. This may be the reason behind the fact that authors observed lifetime PrEP usage as a significant predictor of PEP awareness cross-sectionally at baseline, but not longitudinally.

PEP awareness among LSMM in this study was persistently low, further underscoring the need to not only scale up and disseminate PEP to LSMM, but also evaluate awareness and uptake of PEP over time. The lack of a significant main effect of time after the addition of predictors within our model suggested that even after participating in an 8-month long observational study about HIV prevention, PEP awareness remained low, and that any increases in awareness may be an artifact of study participation rather than any true relationship between predictors and the outcome of interest. Compared to baseline where 67.8% of participants reported little to no awareness of PEP, awareness only modestly improved with 65.1% and 60.5% of LSMM at 4-month and 8-month follow up respectively reporting little to no awareness of PEP. This is particularly concerning due to the fact that one of the priorities of national EHE goals is to increase awareness and uptake of biomedical HIV prevention tools like PEP among subgroups with increased susceptibility of HIV acquisition (Fauci et al., 2019). Although there are several well-established strategies to increase awareness and uptake of PrEP among marginalized groups, there are still no strategies focused specifically on PEP awareness or utilization in the CDC compendium of HIV prevention as of July 2023 (CDC, 2022). Therefore, as we approach the 20th anniversary of the FDA approval of PEP, there is an urgent need to develop multilevel strategies (e.g., peer-delivered PEP messaging via peer ambassadors to simultaneously increase PEP knowledge [individual level] and normalize PEP [interpersonal level], inclusion of PrEP self-efficacy conversations in primary care visits) to enhance the reach of PEP to key populations affected by HIV, including LSMM. Furthermore, it is essential to include both community and implementer feedback through community engaged research approaches (e.g., community advisory/collaboration boards) when generating and disseminating more effective and inclusive HIV prevention programming to enhance reach to LSMM (Weinstein et al., 2023).

The results of this study should be interpreted while considering the following limitations. First, this study did not include validated measures related to PEP due to their non-existence. For example, as opposed to using hypothesized proxies of PEP knowledge and self-efficacy (i.e., PrEP knowledge and self-efficacy), our analyses would have been strengthened with measures of PEP knowledge and self-efficacy. Relatedly, this study only examined PEP awareness and not PEP use which may have had a more direct link to factors associated with health behavior change (e.g., knowledge, self-efficacy). Therefore, future studies should first develop validated PEP specific measures and examine them in relation to PEP awareness and use among subpopulations at increased risk for HIV acquisition. Additionally, the minimal improvement in PEP awareness over time should be interpreted with caution since participating in the study itself may have led to the increased PEP awareness rather than changes in any of the evaluated predictors. Finally, the results of this study cannot be generalized to other populations as we only focused on LSMM in South Florida. It would be valuable for future research to employ a similar three-pronged variable selection approach as they examine PEP awareness both cross-sectionally and longitudinally among other groups that are impacted by HIV in other EHE priority jurisdictions, as well as LSMM living in other EHE priority jurisdictions to compare findings.

Despite these limitations, there are several notable strengths to this study. First, we used a longitudinal hierarchical linear modeling approach to assess both within and between person group differences and changes in PEP awareness over time. This allowed us to (1) quantitatively assess how predictors of PEP awareness changed over time in relation to LSMM themselves and other LSMM within the study and (2) addresses some of the limitations of our prior cross-sectional work. This is notable given that this was a large group of LSMM in a major U.S. HIV epicenter. Second, in addition to the complex quantitative techniques utilized in the study, we also used a participatory data science approach (i.e., active partnership with a CAB composed of LSMM) in our prior cross-sectional study which helped identify, from a lived experience perspective, potential factors related to PEP awareness that were in turn included in our longitudinal model. Finally, we used syndemic, minority stress, and intersectionality theories to account for the unique experiences of LSMM and help contextualize our findings.

The results from this longitudinal analysis of factors related to LSMM’s PEP awareness over time indicate that more must be done to increase PEP awareness among a subgroup with significant HIV disparities. Although authors observed significant within-person effects between both PrEP knowledge and PrEP self-efficacy and PEP awareness across time points as well as between-person effects among both PrEP and HIV knowledge and longitudinal PEP awareness, more research must be conducted to assess PEP interest in and uptake among LSMM. The importance of our findings is underscored by the fact that HIV prevention tools, such as PEP, are insufficiently scaled up and out to populations with multiple marginalized identities (e.g., racial, ethnic, and sexual minorities) despite having been an FDA approved biomedical tool for HIV prevention after a possible exposure for almost 20 years. This includes LSMM, a group that is disproportionately impacted by HIV nationally as well as locally in South Florida, a major U.S. HIV epicenter. Future studies should explore how to engage LSMM in HIV prevention, incorporate PEP information into established (and currently in development) HIV prevention interventions, and identify strategies to increase overall awareness of an effective biomedical tool for HIV prevention, particularly among marginalized groups with persistent HIV disparities.

Data availability

The data that support the findings of this study are available upon reasonable request (e.g., methodologically sound proposal and signed data use agreement) from the corresponding author (A.H.), following publication.