Peer Influence on Motivation to Use Pre-Exposure Prophylaxis Among Latino Sexual Minority Men in Miami,Florida: A Network Autocorrelation Model

Abstract

Despite the availability of pre-exposure prophylaxis (PrEP), Latino sexual minority men (LSMM) continue to experience disparities in PrEP uptake and subsequently, HIV vulnerability. Social network norms are an underutilized solution to increase PrEP uptake. We used a peer influence model (network autocorrelation model) to examine the role of social network descriptive norms (i.e., actual behaviors) surrounding PrEP use. A total of 11 sociocentric networks of 13 friends (n = 143 LSMM) were recruited into our study from 2018 to 2019 in South Florida. Most participants were in PrEP pre-contemplation (n = 44), and almost one-third of our sample were using PrEP (n = 38). Three network autocorrelation models were estimated using an empirically informed Bayesian analysis. We found a positive association between participants’ Motivational PrEP Cascade (MPC) position and their network members’ (friends’) cascade position based on three different measures of connection even when accounting for PrEP knowledge: friendship (ρ = 0.22; 95% CIa = 0.01–0.42), emotional closeness (ρ = 0.24; 95% CI = 0.03–0.44), and frequency of interaction (ρ = 0.22; CI = 0.03–0.42). Our findings highlight that an individual’s progress in the MPC may be influenced by their network members’ progress in the MPC, suggesting that LSMM using PrEP may serve as role models to their peers for PrEP use due to descriptive norms. Our findings further suggest that PrEP interventions for LSMM along the MPC can be implemented at the social network level.

Introduction

In 2021, approximately 1.1 million individuals in the United States were living with HIV.¹ ^,Of particular concern is the disproportionate burden of HIV-related morbidity and mortality among racial, ethnic, sexual, and gender minority subpopulations. In 2021, Latino individuals and men who have sex with men accounted for 29% and 67%, respectively, of the nearly 36,000 new HIV infections in the United States.² If current rates persist, one in five Latino sexual minority men (LSMM) will be diagnosed with HIV by age 49 years.³ In Miami-Dade County, Florida, a domestic epicenter of the HIV epidemic, the HIV incidence rate was 37.3 per 100,000 in 2019 and 30.3 per 100,000 in 2021 among persons aged ≥13 years.⁴ Latino individuals represented 60.4% and sexual minority men (SMM) represented 69.8% of new diagnoses in MDC in 2021.⁴

Pre-exposure prophylaxis (PrEP) is an effective biomedical preventive tool that reduces the risk of infection by over 99% when adherence is followed.^5
–7 However, LSMM have low PrEP uptake despite having high PrEP awareness and acceptability.⁸ Despite many Latino men qualifying for PrEP, there continues to be a large discrepancy in the percentage of these men who actually receive a PrEP prescription compared to White men.⁹ The Motivational PrEP Cascade (MPC)¹⁰ can be used to understand this discrepancy between awareness and uptake. This framework, informed by the Transtheoretical Model of Change and the Theory of Planned Behavior,^11,12 includes five stages to describe an individual’s willingness and intentions for PrEP uptake, engagement, and persistence. The precursory phase for entry into the cascade is objective identification, which is a combination of an individual’s awareness and perception of HIV risk, knowledge of HIV status, and clinical indication for PrEP based on current Centers for Disease Control and Prevention (CDC) PrEP guidelines. Individuals eligible for PrEP may then move forward or backward through the following stages: PrEP precontemplation (unwilling to take PrEP or do not perceive themselves to be an appropriate candidate for PrEP), PrEP contemplation (willing to take PrEP but no means or plans to do so), PrEParation (have plans and means to obtain PrEP but have not yet received PrEP prescription), PrEP action (have spoken to doctor and obtained PrEP prescription), and PrEP maintenance (adhering to PrEP and following PrEP care guidelines).¹⁰

Recent studies have utilized the MPC to characterize PrEP outcomes among SMM to identify where along the cascade individuals may be lost to care or stop progressing.^10,13

–16 Together, these studies consistently suggest that most PrEP-eligible individuals do not progress past the early stages of the MPC. This may be a limitation of the MPC as it gives little attention to the social context and subjective norms (i.e., beliefs of others and the desire to comply with others) that may impact an individual’s willingness and intentions to engage in PrEP care.¹⁷ Thus, we sought to fill this gap by using social network methods to examine the importance of social norms and peer influence on LSMM progression within the MPC.

Individuals can be considered within their social context. Social network support and social network connectedness have been shown to reduce barriers to engaging with PrEP and other HIV prevention services.^18

–21 For example, LSMM may be more likely to engage in PrEP care if they hear about PrEP from a social network member (i.e., a friend or another LSMM).²² Qualitative research suggests that LSMM’s PrEP use is also swayed by social norms, including descriptive norms or observed behaviors within a social network.¹⁷ Greater social network connectedness to the SMM community and Latino group social network membership endorsement are also associated with PrEP initiation.²³ Furthermore, SMM with more social network connections are more likely to engage in PrEP services, necessitating the importance of considering individuals within their social context.²¹ Thus, to examine the influence of the social context, or peer influence, social network analyses such as the autocorrelation model can be used.²⁴ To our knowledge, the present article is one of the first to explore peer influences of an individual’s MPC position, focusing on a population at high vulnerability to HIV (LSMM) and using a Bayesian sociocentric network approach. The overall aim of the present analysis is to characterize LSMM’s stage in the MPC and describe how their stage relates to the stages of those in their friendship network using the peer influence/network autocorrelation model. The underlying mechanism of social influence in the context of this study is hypothesized to be communication. Previous studies have explored discussions and encouragement surrounding PrEP among networks of friends and sexual partners and the relationship of these discussions with various dyadic characteristics (i.e., closeness, frequency of contact).^25

–28 This work highlights the need for additional analyses to further investigate PrEP communication within social networks as a mechanism for social influence in increasing uptake, adherence, and persistence in biomedical HIV prevention services.

Methods

Participants, setting, and recruitment

Data for this study were collected from August 2018 to January 2020. A total of 11 networks composed of 13 individuals (n = 143 LSMM) were recruited into our cross-sectional social network study. We partnered with a South Florida community-based organization (CBO) that prioritizes LSMM sexual health to recruit participants using a respondent-driven sampling (RDS) approach. A detailed description of our RDS recruitment approach can be found elsewhere.^{25
–27,29
–31} To summarize the RDS approach, 11 participants were recruited as seed individuals, and each seed referred their friends to the study, who in turn referred their friends to the study. Participants met the following inclusion criteria: identifying as Latino, self-reporting HIV-negative status, and having recommended indications for PrEP use based on the CDC guidelines.³² Study staff obtained informed consent from participants and then conducted a 2-hour long assessment in English or Spanish using tablets and paper-based sociocentric surveys.

Measures

Sociodemographic measures included age (in years), race (White, Black, Multi-racial, Another race), marital status (single, married), education (high school degree or less, some college, college degree, or more), and county of residency (Miami-Dade County, Broward County). Knowledge of PrEP effectiveness was measured with the question, “How effective is PrEP when taken as prescribed in preventing HIV?” with options, “I don’t know,” “Less than 50% effective,” “Less than 75% effective,” “Less than 90% effective,” and “90% or more effective.” Participants who selected the option “90% or more effective” were scored “1,” while all other responses were scored “0.”

Stage on the MPC was calculated using Parson’s MPC approach,¹⁰ with participants placed based on their current PrEP use at six stages. In order, the stages were: objective identification (this included all participants within our sample as they were identified as being at high vulnerability to HIV in accordance with CDC guidelines),³³ PrEP precontemplation (unwilling to take PrEP), PrEP contemplation (willing to take PrEP but no means or plans to do so), PrEParation (have plans and means to obtain PrEP but have not yet received PrEP prescription), PrEP action (have been prescribed PrEP), and PrEP maintenance (adherence to PrEP in the past 3 months).

Network-level measures included declared friendship with another participant in their recruited network (a binary indicator), their emotional closeness to the declared friends (a 5-point scale from 0 = “Not at all close” to 5=“Extremely close”), and their frequency of interaction with the declared friends (a 4-point scale from 0=“Never” to 4=“Almost every day”).

Descriptive analysis

We assessed associations between sociodemographic characteristics and MPC stages using chi-squared tests, one-sample t-tests, and analysis of variance tests. We constructed three adjacency matrices of the connections between respondents and friends based on friendship, emotional closeness, and interaction frequency. We used each of these in a network autocorrelation model to assess the relationship between a respondent’s stage on the MPC and the positions of the respondent’s friends on the cascade. Each model was estimated with Bayesian methods specifying an informed prior distribution for the autocorrelation coefficient $ρ$ that parametrizes the strength of peer influence on an individual. Network autocorrelation models estimate the manner and degree to which network connections affect some outcome variable, such as stage on the MPC, and have been used in a variety of applications.^34
–36 Our models are categorized as “network effects” models, one of Leenders’ four ways that a correlation could arise between an individual’s opinion/behavior and the opinions/behaviors of their peers.³⁷ A precondition of social influence, as noted by Leenders, is the awareness of the ego of the cascade position of their alters. In our use, an individual’s position on the cascade can be modeled as a function of individual-level covariates that impact the cascade stage as well as the cascade stages of those to whom they have a tie (i.e., friendship connection). Previous work has shown that LSMM discuss and encourage PrEP use among their friendship and sexual networks, indicating that LSMM have access to this knowledge about their friends and sexual partners.^25,27,29 Within the context of this study, survey respondents were asked about the frequency of previous PrEP conversations with each of their friends. Such conversations may represent opportunities for knowledge transfer and encouragement among network members concerning PrEP use.

Three models explore variations in measures of contact with associates. Model A examines an individual’s cascade position and the weighted average of their friends’ cascade positions, where each friend has the same weight. Model B weighs the positions of associates by the level of emotional attachment, so the positions of associates to whom the individual has close emotional attachment have greater weight. Model C weighs the positions of associates by the frequency of interaction; the positions of associates to whom the individual has frequent interaction have greater weight. In each model, the individual’s PrEP knowledge score is included as a covariate. The assumed prior distribution of the autocorrelation coefficient is based on Dittrich, Leenders, and Mulder’s work, in which they explored peer influence in more than 50 published studies, finding a mean for the distribution of 0.36, with a standard deviation of 0.19.³⁸

The network effects autocorrelation model is expressed by the equation: $y = ρ W y + β x + ε$

In this equation, $W$ is the adjacency matrix that captures influence. It is a row-normalized version of the raw adjacency matrix, that is, the rows sum to 1.0, a convention that captures how influence is divided up among the contacts of each respondent. The variable $y$ is the outcome variable of interest (i.e., cascade score) and $x$ is the respondent-level covariate that is hypothesized to impact the cascade stage independent of the scores of one’s friends. We used knowledge of PrEP effectiveness as a covariate in this model. Due to power restrictions, we were unable to include additional covariates. We also report 95% credible intervals, which represent the probability that the true estimate lies within a given interval.³⁹ All analyses were conducted using the R environment.^40,41

Ethical approval

Each participant received a $50 gift card as compensation for their time during interviews and a $10 gift card for each friend referral who enrolled in the study. Study staff obtained written informed consent from participants. Ethical approval was obtained from the University of Miami Institutional Review Board. We used the STROBE guidelines to report our study findings.⁴²

Results

Among seeds, five were current PrEP users and six were not using PrEP at the time of data collection. Most participants were aged between 25 and 29 years, White (71%), and single (87%). Additional descriptive statistics of participants can be found in Table 1 and elsewhere.^{25
–27,29
–31} Most participants were in the PrEP pre-contemplation stage (n = 44), and almost one-third of our sample were using PrEP (n = 38). The MPC visualized as a stacked bar plot can be found in Figure 1.

FIG. 1.

Bar plot of Motivational PrEP Cascade stage of Latino sexual minority men in South Florida, n = 140. PrEP, pre-exposure prophylaxis.

Table 1.

Sociodemographic Information of Latino Sexual Minority Men at Risk of HIV in South Florida, Stratified by Motivational PrEP Cascade Stage

	PrEPPre-Contemplation(n = 44)	PrEPcontemplation(n = 31)	PrEParation(n = 27)	PrEP actionand initiation(n = 9)	PrEPmaintenance(n = 29)	Overall(n = 140)	Summary teststatistics
Age group
18–24 years	6 (13.6%)	8 (25.8%)	6 (22.2%)	3 (33.3%)	4 (13.8%)	27 (19.3%)	χ2 = 20.75; Fisher’s p = 0.052
25–29 years	15 (34.1%)	14 (45.2%)	17 (63.0%)	2 (22.2%)	18 (62.1%)	66 (47.1%)
30–34 years	12 (27.3%)	6 (19.4%)	4 (14.8%)	3 (33.3%)	5 (17.2%)	30 (21.4%)
35–39 years	11 (25.0%)	3 (9.68%)	0 (0%)	1 (11.1%)	2 (6.90%)	17 (12.1%)
Race
White	35 (79.5%)	25 (80.6%)	18 (66.7%)	5 (55.6%)	18 (62.1%)	101 (72.1%)	χ2 = 18.24; Fisher’s p = 0.115
Black	2 (4.55%)	0 (0%)	2 (7.41%)	2 (22.2%)	0 (0%)	6 (4.29%)
Multi-racial	7 (15.9%)	4 (12.9%)	5 (18.5%)	2 (22.2%)	8 (27.6%)	26 (18.6%)
Other	0 (0%)	2 (6.45%)	2 (7.41%)	0 (0%)	3 (10.3%)	7 (5.00%)
Marital status
Married	3 (6.82%)	2 (6.45%)	3 (11.1%)	2 (22.2%)	7 (24.1%)	17 (12.1%)	χ2 = 6.91; Fisher’s p = 0.134
Single	41 (93.2%)	29 (93.5%)	24 (88.9%)	7 (77.8%)	22 (75.9%)	123 (87.9%)	χ2 = 6.91; Fisher’s p = 0.134
Education
College degree or more	11 (25.0%)	15 (48.4%)	11 (40.7%)	2 (22.2%)	11 (37.9%)	50 (35.7%)	χ2 = 7.86; Fisher’s p = 0.482
High school degree or less	7 (15.9%)	4 (12.9%)	3 (11.1%)	0 (0%)	4 (13.8%)	18 (12.9%)
Some college	26 (59.1%)	12 (38.7%)	13 (48.1%)	7 (77.8%)	14 (48.3%)	72 (51.4%)
Knowledge of PrEP effectiveness^*
Incorrect knowledge	17 (38.6%)	11 (35.5%)	12 (44.4%)	4 (44.4%)	3 (10.3%)	47 (33.6%)	χ2 = 9.48;Fisher’s p = 0.028
Correct knowledge	27 (61.4%)	20 (64.5%)	15 (55.6%)	5 (55.6%)	26 (89.7%)	93 (66.4%)	χ2 = 9.48;Fisher’s p = 0.028

p < 0.05.

PrEP, pre-exposure prophylaxis.

A visualization of the LSMM social networks and their stage along the MPC can be found in Figure 2. The color of the node denotes the participants’ stage in the MPC, with darker colors indicating further progression. The shape of the node denotes the seed of the network (i.e., the first person recruited for that network) and is visualized as a square with a dark border, and participants who are not seeds are visualized as a circle with a light gray outline. A line between nodes indicates a declared friendship. As depicted in Figure 2, all networks contain individuals at different stages in the MPC. However, the density varies across the networks, with some networks having ties among every friend and others with only a few friends who know one another.

FIG. 2.

Sociometric visualization of Latino sexual minority men networks and their stage in the Motivational PrEP Cascade, n = 143. PrEP, pre-exposure prophylaxis.

Correlates of MPC stage

PrEP knowledge was significantly associated with an individual’s position on the MPC (χ² = 9.48; Fisher’s p = 0.028), and participants who had correct knowledge of PrEP effectiveness were more likely to be further advanced in the MPC. For participants who had correct information about PrEP effectiveness, 29% were in the pre-contemplation stage, 22% in the contemplation stage, 16% in the preparation stage, 5% in the action and initiation stage, and 28% in the maintenance stage. For participants who had incorrect information about PrEP effectiveness, 36% were in the pre-contemplation stage, 23% in the contemplation stage, 25% in the preparation stage, 9% in the action and initiation stage, and 6% in the maintenance stage. The association between participant age and the MPC position trended toward significance (χ² = 20.75; Fisher’s p = 0.052).

Correlation of MPC position relative to friends’ positions

All correlations between participants’ MPC position and the average of their friends’ positions were positive and provided strong evidence of peer influence (Table 2). The first model (Model A) explored the relationship between the average position of the respondents’ friends and their MPC position with PrEP knowledge as a covariate (β_{PrEPKnowledge} = 0.537, CI = 0.04–0.99), yielding a positive autocorrelation coefficient ( $ρ$ _Friend = 0.217, CI = 0.01–0.42). The second model (Model B) assessed a peer’s influence weighted by emotional closeness with PrEP knowledge as the covariate (β_{PrEPKnowledge} = 0.541, CI = 0.11–0.98), yielding a positive autocorrelation coefficient (ρ _Closeness = 0.239, CI = 0.03–0.44). The final model (Model C) assessed a peer’s influence weighted by the frequency of interaction with PrEP knowledge as the covariate (β_{PrEPKnowledge} = 0.489, CI = −0.01 to 0.99), yielding a positive autocorrelation coefficient (ρ _Frequency = 0.222, CI = 0.03–0.42). For all three measures of connection, the coefficient of peer influence was positive and the 95% credible interval did not contain 0.

Table 2.

Findings from the Bayesian Analysis of the Network Effects Autocorrelation Model

Predictors and key model parameters	Estimate (95% credible interval)^a
Predictors and key model parameters	Model AFriend’s MPC position^b	Model BEmotional closeness^b	Model CFrequency of interaction^b
Intercept	1.176 (0.233–2.191)	1.113 (0.217–2.062)	1.270 (0.297–2.248)
PrEP knowledge	0.537 (0.037–0.992)	0.541 (0.109–0.978)	0.489 (−0.009–0.993)
ρ	0.217 (0.010–0.423)	0.239 (0.032–0.438)	0.222 (0.027–0.418)
Posterior probabilities^c	H1: 0.881 H2: 0.119	H1: 0.922 H2: 0.078	H1: 0.931 H2: 0.069

The intervals presented in parentheses refer to the credible interval.

Weighted average of network members.

H1: rho > 0; H2: complement.

PrEP, pre-exposure prophylaxis.

Finally, the Bayesian approach allowed us to calculate the posterior probability of a specific hypothesis, like $ρ > 0$ , versus an alternative hypothesis, $ρ \leq 0$ . In all three cases, the hypothesis $ρ > 0$ was strongly supported by the evidence. For friendship, its posterior probability was 0.88 versus 0.12 for the alternative; for emotional closeness, its posterior probability was 0.92 versus 0.08 for the alternative; and for frequency of interaction, its posterior probability was 0.93 versus 0.07 for the alternative.

Discussion

Our study used a sociocentric network approach to investigate how descriptive norms within social networks can influence LSMM’s progression in the MPC. We found that although around 25% of our sample were using PrEP, generally progression along the MPC was not as advanced as that seen in other studies. Importantly, we found that LSMM’s PrEP use was influenced by their friends’ PrEP use. More specifically, LSMM’s stage in the MPC was associated with the average stage of the participants’ friends’ position in the cascade. We also found that LSMM’s progress along the MPC was associated with correct PrEP knowledge. Furthermore, we found that emotional attachment and interaction frequency may influence progression along the MPC, with stronger emotional attachment and more frequent interactions positively weighing the strength of the autocorrelation. Altogether, the models were very similar statistically in terms of size of effects. Substantively, we might expect the model based on interaction frequency (Model C) to display the strongest effect since direct contact with an associate is intuitively a prerequisite for influence. On the other hand, frequent contact with someone to whom one does not feel especially close may not lead to more influence than less frequent contact but with someone to whom one feels especially close, an argument that weighs in favor of emotional closeness as a preferred indicator of adjacency.

A key finding of this study was the positive association between an individual’s stage on the MPC and the mean stage of their friends—an association influenced by emotional attachment and frequency of interaction. We found that a person’s cascade position was more closely associated with the cascade position of those with whom the individual perceived themselves to be closer or with whom they interacted more frequently. Our findings support previous studies that found that individual health behaviors and motivations, such as those related to the use of and adherence to PrEP among LSMM, do not occur in isolation from one’s social and structural environments. Instead, these behaviors are influenced by an individual’s relationships with others and the values or behavioral norms of such behaviors within their social networks.^43

–47 Our findings also support previous research that found that exposure to a behavior within a network is associated with the adoption of that behavior and that tie strength is associated with peer influence.^48
–50 This also aligns with previous research that while weaker ties are important for the dissemination of information, tie strength is more important to “transmit” more complex behaviors, such as PrEP use.^48,51,52

We found that around 25% of our sample was either taking PrEP or adherent to PrEP; other studies have found PrEP use to be higher with 28–42% of LSMM using PrEP,^13,23,53 with one study reporting PrEP use to be lower among LSMM with only 15–19% of the LSMM sample reporting PrEP use.^54,55 Because our study was recruited through a CBO in South Florida that prioritizes LSMM health, this could suggest that the high awareness and use of PrEP in this sample could be in part due to the efforts of this CBO. However, in the circumstance that participants of this study were not seeking services at our partner CBO, this suggests that social networks could serve as a bridge to PrEP information and access. Our findings support the original construction of the MPC, which found that once participants reached PrEP action and initiation, more participants reached PrEP maintenance.¹⁰ Future interventions can leverage the current clientele of CBOs and focus on peer navigation or influence interventions to reach their social networks.

A review by Zhang and Centola (2019) also highlighted the complexity of the spread of health-related information, beliefs, and behaviors through social networks and the importance of understanding the general mechanisms for the diffusion of these health-related contagions (e.g., pathogenic diffusion, diffusion of information and beliefs, diffusion of positive and negative health behaviors).⁴⁴ Peer influence may operate to change health outcomes through several processes such as social norm adoption, social learning, social pressure, and peer imitation.⁴⁴ Furthermore, peer influence within the MPC may have significant implications for both the stage of LSMM along the cascade and why LSMM may fall out of care. Although the specific mechanisms of peer influence for PrEP uptake, adherence, and maintenance were not assessed in this study, future studies may expand upon these findings by incorporating questions as to whether participants know the MPC stage of their friends to better understand the underlying processes of peer influence on PrEP uptake, adherence, and maintenance among LSMM and their social networks.

While this is one of the first studies to demonstrate peer influence as it relates to the MPC within friendship networks of LSMM, the MPC has been criticized previously for the lack of attention given to the social context, which may impact an individual’s willingness and intentions to engage in PrEP care.¹⁷ The findings from this study indicate that when used in conjunction with social network approaches, the MPC can be a useful tool to describe engagement with PrEP services among friendship networks of LSMM. Engagement in preventive behaviors, such as HIV testing and condom use, may, at least in part, be influenced by the consistency of such behaviors with the descriptive norms and values of one’s social network.⁵⁶ The focus of this study was on interpersonal behaviors and dynamics within LSMM friendship networks. However, progression along the PrEP cascade is heavily influenced by factors at each level of the socioecological model. In particular, structural and systemic barriers impede access and engagement in PrEP services among Latinx populations, such as competing needs (i.e., housing, food, employment),⁵⁷ costs, distance to clinics,³¹ shortage of Spanish-speaking health care providers and availability of Spanish-language informational materials,^31,57,58 immigration status and concerns,^57
–59 medical mistrust,⁵⁵ and negative PrEP messaging.⁶⁰ Because LSMM may not access sexual health services at the same rate as White SMM, often due to immigration and discrimination-related barriers and stressors,³¹ social networks have utility in reinforcing PrEP use norms and serving as a bridge to PrEP services. Future studies may build up on these findings to further explore the importance of social norms and networks as tools for increasing access to information about biomedical HIV prevention services. Network interventions along with other community-based programming (i.e., CBO-delivered PrEP telehealth⁶¹) may serve to further increase access to and engagement in PrEP services among priority populations.

Implications for public health interventions

Our findings support the utility of network interventions, defined as “purposeful efforts to use social networks or social network data to generate social influence, accelerate behavior change, improve performance, and/or achieve desirable outcomes among individuals, communities, organizations, or populations,” to increase PrEP use among LSMM.⁶² More specifically, our findings suggest that social network-based PrEP interventions can leverage emotional attachment and interaction frequency—relational attributes that bear weight on the peer influence of PrEP—relative to just having a friend in the network who uses PrEP. There are four types of social network-based interventions: identification, segmentation, induction, and alteration.⁶² Of these, HIV prevention interventions have predominantly and successfully used identification, segmentation, and induction strategies.^62,63 Research suggests that individuals who are more central within sociocentric networks and in the network core are more likely to practice HIV preventive behaviors such as HIV testing. Thus, a potential future intervention could consider altering network structures to increase social proximity of LSMM to other LSMM who have progressed along the MPC. Alternatively, efforts can be focused on supporting and providing resources to peers and trusted community leaders (i.e., individuals that are already within LSMM friendship networks) so that they can serve as sources of PrEP information, reduce stigma, and increase trust in communities that have experienced historical and systemic distrust in the health care system.²⁸ Key network members may also be instrumental in increasing PrEP use self-efficacy, frequency of PrEP-related conversations, and improving willingness to use PrEP and PrEP-related descriptive and subjective peer norms among sociocentric networks of key populations.^29,47

Limitations

This study is not without limitations. First, our findings may not be generalizable to the broader US LSMM population for several reasons: (1) Miami is heterogenous for Latino nationalities, which is not often seen throughout the rest of the nation; and (2) the study sample was identified using RDS, a nonrandom type of sampling method. However, as we have 13 networks of LSMM, there is some generalizability of our findings outside this study.

Regarding the network structure of those included in the sample, one of the networks had all friends stemming unidirectionally from the seed—an abnormal structure that was outside of the RDS recruitment approach that was used in this study. Given the cross-sectional nature of the study, causality cannot be determined. In addition, we were unable to test for changes in PrEP use or observe the directionality of peer influence; thus, our study does not ascertain whether LSMM use PrEP because their network members are using PrEP or if they are actively seeking out other PrEP users to include them in their network (i.e., differential attribution). Furthermore, as a cross-sectional study, we were unable to determine the driving force behind progression in the PrEP cascade. Our ability to understand the role of PrEP knowledge as a measure of intrinsic opinion that either influences or is influenced by progression along the cascade is also limited. In future longitudinal studies, we will examine changes in network dynamics and their association with progression in the cascade.

Conclusions

Despite the US Food and Drug Administration approving the use of oral PrEP in 2012 and injectable PrEP in 2021,⁶⁴ LSMM continue to experience HIV and PrEP-related disparities. We found that by increasing one person’s PrEP use, a program may also increase their network’s positive descriptive norms surrounding PrEP, as we found that exposure to PrEP use within a network is associated with the adoption of PrEP use. The widening gap between PrEP awareness and use in this population may be due to multi-faceted structural and individual-level barriers that prevent access to HIV prevention services. This study found that social network factors affect individual health outcomes and can be used to develop health behavior interventions. Our findings suggest that the MPC can be a useful tool to describe engagement with PrEP services among friendship networks of LSMM.

Footnotes

Acknowledgments

The authors would like to acknowledge our community partner, Latino Salud, for their continued support. The authors would also like to thank our participants without whom this work would not be possible.

Authors’ Contributions

C.H.S.: Conceptualization, methodology, investigation, data curation, writing—original draft, writing—review and editing, visualization, and project administration. L.C.: Writing—original draft and writing—review and editing. E.R.: Writing—original draft and writing—review and editing. A.J.: Writing—original draft and writing—review and editing. J.S.: Formal analysis and writing—reviewing and editing. K.S.: Writing—original draft and writing—review and editing. M.K.: Conceptualization, methodology, investigation, writing—review and editing, supervision, and funding acquisition.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported by the National Institute on Drug Abuse (#K99DA041494 and #R00DA041494; PI: M.K.), the National Institute of Allergy and Infectious Diseases (award #P30AI050409 CFAR Adelante sub-award; PI: M.K.) and the National Institute of Mental Health (#R01MH125727; PI: M.K.; P30MH116867; PI: Safren). C.H.S.’s efforts were supported by the National Institute of Minority Health and Health Disparities (#F31MD015988; PI: C.H.S.; U54MD002266 PI: Behar-Zusman), the National Institute of Drug Abuse (#R25DA050687; #T32DA031099), the National Institute of Allergy and Infectious Diseases (#T32AI114398), and the National Institute of Mental Health (#P30MH43520). The content presented in this article is solely the responsibility of the authors and does not necessarily represent the official views of NIH or its affiliate institutions.

References

Sullivan

, Woodyatt

, Koski

, et al. A data visualization and dissemination resource to support HIV prevention and care at the local level: Analysis and uses of the AIDSVu public data resource. J Med Internet Res, 2020; 22(10):e23173; doi: 10.2196/23173

Centers for Disease Control and Prevention. HIV Surveillance Report, 2021; Vol 34. 2023. Available from: https://stacks.cdc.gov/view/cdc/149071 [Last accessed: August 1, 2023 ].

Hess

, Hu

, Lansky

, et al. Lifetime risk of a diagnosis of HIV infection in the United States. Ann Epidemiol, 2017; 27(4):238–243.

Florida Department of Health. Miami-Dade county epidemiological profile, 2021. 2024. Available from: https://www.flhealthcharts.gov/ChartsDashboards/rdPage.aspx?rdReport=HIVAIDS.DataViewer&cid=0471 [Last accessed: March 20, 2024 ].

Landovitz

, Li

, Grinsztejn

, et al. Safety, tolerability, and pharmacokinetics of long-acting injectable cabotegravir in low-risk HIV-uninfected individuals: HPTN 077, a phase 2a randomized controlled trial. PLoS Med, 2018; 15(11):e1002690.

HIV Prevention Trials Network. HPTN 083: A phase 2b/3 double blind safety and efficacy study of injectable cabotegravir compared to daily oral tenofovir disoproxil fumarate/emtricitabine (TDF/FTC), for pre-exposure prophylaxis in HIV-uninfected cisgender men and transgender women who have sex with men. Emtricitabine (TDF/FTC), for Pre-Exposure Prophylaxis in HIV-Uninfected Cisgender Men and Transgender Women who have Sex with Men. 2017. Available from: https://www.hptn.org/research/studies/hptn083

Clement

, Kofron

, Landovitz

. Long-acting injectable cabotegravir for the prevention of HIV infection. Curr Opin HIV AIDS, 2020; 15(1):19–26.

Algarin

, Shrader

, Bhatt

, et al. The pre-exposure prophylaxis (PrEP) continuum of care and correlates to initiation among HIV-negative men recruited at Miami Gay Pride 2018. J Urban Health, 2019; 96(6):835–844.

Centers for Disease Control and Prevention. HIV Surveillance Supplemental Report: Monitoring Selected National HIV Prevention and Care Objectives by Using HIV Surveillance Data United States and 6 Dependent Areas, 2018; Vol 25 (2). 2020. Available from: https://stacks.cdc.gov/view/cdc/149255 [Last accessed: August 2, 2024 ].

10.

Parsons

, Rendina

, Lassiter

, et al. Uptake of HIV Pre-Exposure Prophylaxis (PrEP) in a national cohort of gay and bisexual men in the United States. J Acquir Immune Defic Syndr, 2017; 74(3):285–292; doi: 10.1097/qai.0000000000001251

11.

Ajzen

. The theory of planned behavior. Organ Behav Hum Decis Process, 1991; 50(2):179–211; doi: 10.1016/0749-5978(91)90020-T

12.

Prochaska

. Transtheoretical Model of Behavior Change. In: Encyclopedia of Behavioral Medicine. ( Gellman

. ed.) Springer International Publishing: Cham; 2020; pp. 2266–2270.

13.

Moskowitz

, Moran

, Matson

, et al. The PrEP cascade in a national cohort of adolescent men who have sex with men. J Acquir Immune Defic Syndr, 2021; 86(5):536–543; doi: 10.1097/qai.0000000000002613

14.

Dehlin

, Issema

, Eavou

, et al. The motivational PrEP cascade guides interpretation of early PrEP linkage to care for young black men who have sex with men: The case of Chicago’s PrEPLine. AIDS Educ Prev, 2019; 31(6):491–504; doi: 10.1521/aeap.2019.31.6.491

15.

Mehrotra

, Westmoreland

, Patel

, et al. Breaking inertia: Movement along the PrEP cascade in a longitudinal us national cohort of sexual minority individuals at risk for HIV. J Acquir Immune Defic Syndr, 2021; 86(5):e118–e125; doi: 10.1097/qai.0000000000002611

16.

Willie

, Stockman

, Keene

, et al. Social networks and its impact on women’s awareness, interest, and uptake of HIV Pre-exposure Prophylaxis (PrEP): Implications for women experiencing intimate partner violence. J Acquir Immune Defic Syndr, 2019; 80(4):386–393; doi: 10.1097/QAI.0000000000001935

17.

Schnarrs

, Gordon

, Martin-Valenzuela

, et al. Perceived social norms about oral PrEP use: Differences between African–American, Latino and white gay, bisexual and other men who have sex with men in Texas. AIDS Behav, 2018; 22(11):3588–3602; doi: 10.1007/s10461-018-2076-7

18.

Lelutiu-Weinberger

, Golub

. Enhancing PrEP access for black and Latino men who have sex with men. J Acquir Immune Defic Syndr, 2016; 73(5):547–555; doi: 10.1097/QAI.0000000000001140

19.

Zapata

, Petroll

, de St Aubin

, et al. Perspectives on social support and stigma in PrEP-related care among gay and bisexual men: A qualitative investigation. J Homosex, 2020; 69(2):254–276; doi: 10.1080/00918369.2020.1819709

20.

Phillips

, Neray

, Birkett

, et al. Role of social and sexual network factors in PrEP utilization among YMSM and transgender women in Chicago. Prev Sci, 2019; 20(7):1089–1097; doi: 10.1007/s11121-019-00995-6

21.

Kuhns

, Hotton

, Schneider

, et al. Use of Pre-exposure Prophylaxis (PrEP) in young men who have sex with men is associated with race, sexual risk behavior and peer network size. AIDS Behav, 2017; 21(5):1376–1382; doi: 10.1007/s10461-017-1739-0

22.

Harkness

, Satyanarayana

, Mayo

, et al. Scaling up and out HIV prevention and behavioral health services to Latino sexual minority men in South Florida: Multi-Level implementation barriers, facilitators, and strategies. AIDS Patient Care STDS, 2021; 35(5):167–179; doi: 10.1089/apc.2021.0018

23.

Robles

, Hong

, Yu

, et al. Intersecting communities and PrEP uptake among US-based latinx sexual minority men. J Racial Ethn Health Disparities, 2021; 9(6):2157–2163; doi: 10.1007/s40615-021-01154-w

24.

Valente

. Network models and methods for studying the diffusion of innovations. Models and Methods in Social Network Analysis, 2005; 28:98–116.

25.

Shrader

C-H

, Arroyo-Flores

, Skvoretz

, et al. PrEP Use and PrEP use disclosure are associated with condom use during sex: A multilevel analysis of latino MSM egocentric sexual networks. AIDS Behav, 2021; 25(5):1636–1645; doi: 10.1007/s10461-020-03080-0

26.

Shrader

C-H

, Arroyo-Flores

, Stoler

, et al. The association between social and spatial closeness with PrEP conversations among Latino men who have sex with men. J Acquir Immune Defic Syndr, 2021; 88(4):366–375; doi: 10.1097/qai.0000000000002777

27.

Kanamori

, Shrader

C-H

, Flores-Arroyo

, et al. The association between egocentric sexual networks and sexual meeting venues with PrEP conversation and encouragement for use among Latinx men who have sex with men. AIDS Care, 2022; 34(11):1420–1427.

28.

Quinn

, Christenson

, Spector

, et al. The influence of peers on PrEP perceptions and use among young black gay, bisexual, and other men who have sex with men: A qualitative examination. Arch Sex Behav, 2020; 49(6):2129–2143; doi: 10.1007/s10508-019-01593-x

29.

Kanamori

, Shrader

, Johnson

, et al. The association between homophily on illicit drug use and PrEP conversations among Latino men who have sex with men friends: A dyadic network and spatially explicit study. Arch Sex Behav, 2022; 51(5):2485–2495; doi: 10.1007/s10508-021-02131-4

30.

Kanamori

, Shrader

C-H

, St George

, et al. Influences of immigration stress and occupational exploitation on Latina seasonal workers’ substance use networks: A qualitative study. J Ethn Subst Abuse, 2022; 21(2):457–475.

31.

Shrader

C-H

, Stoler

, Arroyo-Flores

, et al. Geographic disparities in availability of Spanish-language PrEP services among Latino sexual minority men in south Florida. J Immigr Minor Health, 2022; 25(2):374–381.

32.

Centers for Disease Control and Prevention. Preexposure Prophylaxis for the Prevention of HIV Infection in the United States—2017 Update: A Clinical Practice Guideline. Centers for Disease Control and Prevention: Atlanta; 2018.

33.

US Public Health Service. Centers for disease control and prevention. Preexposure prophylaxis for the prevention of HIV infection in the United States-2021 update. 2021 Available from: https://www.cdc.gov/hiv/pdf/risk/prep/cdc-hiv-prep-guidelines-2021.pdf

34.

Gimpel

, Schuknecht

. Political participation and the accessibility of the ballot box. Polit Geogr, 2003; 22(5):471–488.

35.

Mizruchi

, Stearns

. The conditional nature of embeddedness: A study of borrowing by large US firms, 1973–1994. Am Sociol Rev, 2006; 71(2):310–333.

36.

O’Malley

, Marsden

. The analysis of social networks. Health Serv Outcomes Res Methodol, 2008; 8(4):222–269.

37.

Leenders

RTAJ

. Modeling social influence through network autocorrelation: Constructing the weight matrix. Social Networks, 2002; 24(1):21–47; doi: 10.1016/S0378-8733(01)00049-1

38.

Dittrich

, Leenders

RTAJ

, Mulder

. Bayesian estimation of the network autocorrelation model. Social Networks, 2017; 48:213–236; doi: 10.1016/j.socnet.2016.09.002

39.

Hespanhol

, Vallio

, Costa

, et al. Understanding and interpreting confidence and credible intervals around effect estimates. Braz J Phys Ther, 2019; 23(4):290–301; doi: 10.1016/j.bjpt.2018.12.006

40.

R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing: Vienna, Austria; 2019.

41.

Mulder

, Dittrich

, Leenders

RTAJ

. BANAM: Bayesian analysis of the network autocorrelation model. 2023.

42.

Cuschieri

. The strobe guidelines. Saudi J Anaesth, 2019; 13(Suppl 1):S31–s34; doi: 10.4103/sja.SJA_543_18

43.

Valente

. Social Networks and Health: Models, Methods, and Applications. Oxford University Press; Oxford; 2010.

44.

Zhang

, Centola

. Social networks and health: New developments in diffusion, online and offline. Annu Rev Sociol, 2019; 45(1):91–109; doi: 10.1146/annurev-soc-073117-041421

45.

Centola

. The spread of behavior in an online social network experiment. Science, 2010; 329(5996):1194–1197; doi: 10.1126/science.1185231

46.

Centola

. An experimental study of homophily in the adoption of health behavior. Science, 2011; 334(6060):1269–1272; doi: 10.1126/science.1207055

47.

Kelly

, Amirkhanian

, Walsh

, et al. Social network intervention to increase pre-exposure prophylaxis (PrEP) awareness, interest, and use among African American men who have sex with men. AIDS Care, 2020; 32(sup2):40–46; doi: 10.1080/09540121.2020.1739207

48.

Alexander

, Piazza

, Mekos

, et al. Peers, schools, and adolescent cigarette smoking. J Adolesc Health, 2001; 29(1):22–30; doi: 10.1016/s1054-139x(01)00210-5

49.

Shrader

C-H

, Duncan

, Chen

Y-T

, et al. Latent profile patterns of network-level norms and associations with individual-level sexual behaviors among Black sexual and gender minorities assigned male at birth: The N2 cohort study in Chicago. Arch Sex Behav, 2023; 52(6):2355–2372.

50.

Rogers

, Kincaid

. Communication Networks: Toward a New Paradigm For Research. Free Press: New York; 1981.

51.

Valente

. Network Models of the Diffusion of Innovations (Quantitative Methods in Communication Subseries). Hampton Press: Cresskill, NJ; 1995.

52.

Valente

. Social network thresholds in the diffusion of innovations. Social Networks, 1996; 18(1):69–89.

53.

Bonett

, Dowshen

, Bauermeister

, et al. Characterizing the PrEP continuum for black and latinx sexual and gender minority youth. AIDS Behav, 2022; 26(4):1211–1221; doi: 10.1007/s10461-021-03476-6

54.

Rivera

, Brady

, Blashill

. Traditional machismo, caballerismo, and the pre-exposure Prophylaxis (PrEP) cascade among a sample of Latino sexual minority men. J Sex Res, 2021; 58(1):21–28; doi: 10.1080/00224499.2020.1743961

55.

Kimball

, Rivera

, Gonzales

, et al. Medical mistrust and the PrEP cascade among Latino sexual minority men. AIDS Behav, 2020; 24(12):3456–3461; doi: 10.1007/s10461-020-02916-z

56.

Fisher

. Possible effects of reference group-based social influence on AIDS-risk behavior and AIDS-prevention. Am Psychol, 1988; 43(11):914–920; doi: 10.1037/0003-066X.43.11.914

57.

Yamanis

, Gomez

, Landis

, et al. Pandemic-Era policies that can help end the HIV epidemic for Latinas/os/xs. AIDS Patient Care STDS, 2023; 37(8):374–378; doi: 10.1089/apc.2023.0072

58.

Lozano

, Jaramillo

, Prado

, et al. Barriers and facilitators to PrEP use and HIV testing for subgroups of Latino sexual minority men. AIDS Care, 2023; 35(9):1329–1337; doi: 10.1080/09540121.2023.2216009

59.

Harkness

, Lozano

, Bainter

, et al. Engaging Latino sexual minority men in PrEP and behavioral health care: Multilevel barriers, facilitators, and potential implementation strategies. J Behav Med, 2023; 46(4):655–667; doi: 10.1007/s10865-022-00371-w

60.

Shorrock

, Alvarenga

, Hailey-Fair

, et al. Dismantling barriers and transforming the future of pre-exposure prophylaxis uptake in young black and Latinx sexual minority men and transgender women. AIDS Patient Care STDS, 2022; 36(5):194–203; doi: 10.1089/apc.2021.0222

61.

Andrade

, Betancourt

, Morales

, et al. A community-based pre-exposure prophylaxis telehealth program focused on Latinx sexual minority men. AIDS Patient Care STDS, 2023; 37(11):517–524; doi: 10.1089/apc.2023.0185

62.

Valente

. Network interventions. Science, 2012; 337(6090):49–53; doi: 10.1126/science.1217330

63.

Valente

. Social Networks and Health: Models, Methods, and Application. Oxford University Press: New York, NY; 2010.

64.

US Food & Drug Administration. FDA Approves First Injectable Treatment for HIV Pre-Exposure Prevention. Available from: https://www.fda.gov/news-events/press-announcements/fda-approves-first-injectable-treatment-hiv-pre-exposure-prevention