Preferences Across Pre-Exposure Prophylaxis Modalities Among Young Men Who Have Sex with Men in the United States: A Latent Class Analysis Study

Abstract

Access to daily oral pre-exposure prophylaxis (PrEP) is suboptimal among young cisgender men who have sex with men (YMSM) in the United States. Next-generation modalities that do not involve daily oral regimens may mitigate some of the barriers to PrEP use. We identified latent classes of YMSM based on health care decision-making patterns and examined associations between latent classes and access to health care and PrEP modality preferences (i.e., daily and event-driven oral, rectal douches, broadly neutralizing antibodies, subcutaneous implants, and an injectable). Between October 2020 and June 2021, we administered an online survey to 737 YMSM. Latent class analysis (LCA) identified groups of YMSM based on communication with providers, stigma and mistrust in health care, and autonomy in sexual health decisions. Logistic regression examined associations between class membership and health care access, and exploded logit regression examined associations between class membership and ranked PrEP modality preferences. LCA identified three classes: shared decision-making (high communication with providers and high autonomy); provider-led decision-making (high communication and low autonomy); and patient-driven decision-making (low communication and high autonomy). Shared decision-making was associated with higher access to health care in comparison with the other classes. Across all classes, YMSM preferred daily oral PrEP over all next-generation PrEP modalities. Preferences for daily oral PrEP over next-generation PrEP modalities were particularly marked among the patient-driven decision-making class. Shared decision-making is associated with access to health care and HIV prevention and higher acceptability of next-generation PrEP modalities, and should be considered as part of future interventions to promote use of daily oral and next-generation PrEP.

Introduction

Young cisgender men who have sex with men (YMSM) in the United States are at increased vulnerability to HIV transmission due to a combination of cognitive-developmental, psychosocial, and structural factors.^1,2 HIV risk among this population is compounded by suboptimal engagement in HIV prevention services, including pre-exposure prophylaxis (PrEP).^3,4 Daily oral PrEP was FDA-approved in 2018 for youth at increased risk for HIV acquisition.⁵ However, less than one-sixth of PrEP-indicated youth 16–24 years are taking the medication, about half as much as individuals >24 years.⁶

Communication about sexual health with providers remains challenging for sexual minority youth. Many YMSM do not feel comfortable discussing sexuality with health care providers, hindering access to HIV/STI testing, PrEP, and other preventive services (e.g., HPV vaccination and cancer screening).^7

–11 Hesitancy to communicate with providers stems from previous experiences of stigma in health care and low competence to provide care for lesbian, gay, bisexual, transgender, and queer (LGBTQ) patients among health care providers, leading to medical mistrust.^10,11

Youth who are not out to their family may be even less likely to communicate about sexuality with providers, under fear of disclosure of these conversations to family members.^7,12,13 Limited communication with providers, poor access to accurate medical information, and discrimination in health care limit patient autonomy,^14,15 and contribute to suboptimal access to HIV prevention and PrEP.^16

–19 Conversely, open communication allows for decreased informational and power asymmetry between patients and providers, enabling patient autonomy and shared decision-making.²⁰

Next-generation PrEP may address some of the challenges to PrEP uptake among YMSM. Bimonthly injectable cabotegravir was recently FDA-approved in the United States,⁵ and event-driven oral PrEP is recommended as an option for MSM by the CDC.⁵ In addition, other modalities such as subcutaneous implants, intravenous broadly neutralizing antibodies (bnAb), and rectal douches are being developed.^21,22 Products found to be efficacious will expand the biobehavioral HIV prevention toolkit and give YMSM options that are more congruent with their lifestyles, preferences, and needs.

However, to close the gap between PrEP indication and uptake among YMSM, existing barriers to health care, including communication with providers and health care stigma, must be addressed. Our previous research showed that discussing next-generation PrEP modalities with providers may be more challenging than discussing daily oral PrEP.¹² Different PrEP modalities may also bring about distinct concerns about confidentiality (e.g., seeing subcutaneous implants through the skin vs. having pill bottles at home), and can expose YMSM to stigma, including from family members.^12,23 Acceptability of next-generation PrEP modalities may also be influenced by level of autonomy in health care decision-making and dispositions toward medical innovations (e.g., early vs. late adopters).^24
–26

Therefore, understanding how health care decision-making patterns among youth shape preferences across HIV prevention options is crucial for designing health services for YMSM's needs. We sought to identify latent classes of YMSM based on characteristics of sexual health decision-making and experiences accessing health care; and to examine associations between profiles of YMSM and engagement in health care and HIV prevention, and preferences across PrEP modalities.

Methods

Participants and procedures

Between October 2020 and June 2021, we administered an online survey to 737 YMSM recruited from social media (e.g., Facebook), targeted e-mail lists, and, for individuals ≥18 years old, dating apps (e.g., Jack'd). Inclusion criteria included being 15–24 years old, assigned male sex at birth and identifying as male at enrollment, self-reported HIV-negative or status unknown, reporting same-sex attractions or consensual sexual behaviors with another man in the past 6 months, living in the United States, and speaking English fluently.

Individuals completed an online screening survey to determine eligibility. Research staff verified CAPTCHA validation, time to complete screening survey, IP address, phone number, and e-mail address to identify fraudulent and duplicate responses.^27
–29 Eligible individuals received an individualized link and password to the survey. The online survey asked participants to rank preferences for six different PrEP modalities (daily oral, event-driven oral, injectable, bnAb infusions, subcutaneous implant, and rectal douche) and contained questions on sociodemographic characteristics, access to health care, health care decision-making, and HIV risk. We sent up to 11 e-mail reminders to enhance response rates. The survey was designed and pilot-tested with a youth advisory board (YAB) of sexual and gender minority youth to be completed in ∼45 min.

Individuals received $40 for participation. All participants provided online informed consent/assent. A waiver of parental consent was obtained for individuals 15–17 years old. The Institutional Review Board at University of Pennsylvania approved this study.

Measures

Informed by our previous study with YMSM,¹² latent class analysis (LCA) included six indicators across three domains: sexual health communication, stigma and mistrust in health care, and autonomy in sexual health decisions (Table 1).

Table 1.

Items Included in Latent Class Analysis

		Response options	Coding
	Domain 1: Sexual health communication
1	Does your health care provider know your sexual orientation?	Yes/No	Yes = 1No = 0
2	I believe that my primary health care provider should know everything about my sex life to take good care of me	SD/D/A/SA	A/SA = 1D/SD = 0
	Domain 2: Stigma and mistrust in health care
3	“I have been mistreated by healthcare providers” OR “I have been ignored by healthcare providers”	Never/Rarely/Often/Always	Rarely, Often, Always to Either = 1Never to both = 0
4	When health care providers make mistakes, they usually cover it up	SD/D/A/SA	A/SA = 1D/SD = 0
	Domain 3: Autonomy in sexual health decision
6	I would rather have my primary health care provider make the decisions about what's best for my sexual health than to be given a whole lot of choices	SD/D/N/A/SA	D/SD = 1N/A/SA = 0(reverse-coded)
7	The important medical decisions about my sexual health should be made by my health care provider, not me	SD/D/N/A/SA	D/SD = 1N/A/SA = 0(reverse-coded)

A, agree; D, disagree; N, neither agree nor disagree; SA, strongly agree; SD, strongly disagree.

Sexual health communication with providers

Sexual health communication with providers was assessed using two binary-coded items: (1) having a provider who knows about the participants' sexual orientation and (2) affirmative response to question asking if provider “should know everything about their sex life to take good care of them” (adapted from Flynn et al.³⁰).

Stigma and mistrust in health care

Stigma in health care was measured with two items from the Experiences of Healthcare-Related Sexual Orientation Stigma survey³¹ assessing whether participants had ever been “mistreated” or “ignored” by health care providers due to their sexual orientation. We assessed mistrust in health care with one item from the Global Medical Mistrust scale.³¹ Participants who agreed with the statement, “When healthcare providers make mistakes they usually cover it up,” were considered to endorse medical mistrust. Given low sample variability in responses to other items in the scale (∼80% endorsement) and resulting poor latent class delineation, we selected this strongly worded item as a conservative measure of mistrust in health care.

Autonomy in sexual health decisions

We measured autonomy in sexual health care by adapting validated items on autonomy in general health care to assess whether participants preferred to have their providers make important sexual health decisions for them and whether they would like to be given several choices from which to pick.³⁰ Items were reverse-coded so that participants who reported disagreeing with the statements were considered to endorse autonomy in sexual health decisions.

Access to health care and HIV prevention services

Considering different confidentiality concerns for youth accessing HIV prevention through their parents’/guardians' insurance versus their own,^12,32 we considered three insurance status categories (having own insurance, covered under parents’/guardians’, and uninsured). We also assessed whether participants believed health care services met their health needs, answered on a 5-point scale and further dichotomized (“always”/“usually” vs. “sometimes”/“rarely”/“never”). We also assessed self-reported HIV test (ever), PrEP awareness (aware/unaware), and PrEP use (currently/previously/never).

Preferences across PrEP modalities

Participants were provided with written descriptions of six different PrEP modalities developed through cognitive interviews with YMSM and YAB consultations.³³ Participants were then asked to rank the following PrEP modalities from 1 (most preferred) to 6 (least preferred): daily oral, event-driven oral, bimonthly intramuscular injections, yearly subcutaneous implants, bimonthly intravenous bnAb infusions, and event-driven rectal douches.

We also assessed sexual identity, educational attainment, food insecurity, and ZIP code.

Analytic plan

We identified latent classes of YMSM based on indicators of sexual health communication with providers, stigma and mistrust in health care, and autonomy in sexual health decisions. Starting with a one-class model, we fit models with increasingly larger number of classes until models were underidentified. We then compared fit indices (adjusted likelihood ratio chi-square goodness-of-fit test, Bayesian information criterion, consistent Akaike's information criterion, and the adjusted Lo–Mendell–Rubin and bootstrap likelihood ratio tests) and substantive interpretability across candidate models to select our final model.³⁴ We considered class-specific item probabilities <30% or >70% as thresholds for class homogeneity and between-class item endorsement odds ratios (ORs) >5 or <0.2 for class separation.³⁴ In all subsequent analyses, we assigned individuals to their most likely class membership based on posterior class probabilities, an approach with good performance if model entropy ≥0.8.³⁵

We used multinomial logistic regression to examine associations between sociodemographic variables (independent variables) and most likely class membership (dependent variable). We then used logistic regression to examine associations between most likely class membership (independent variable) and measures of engagement in health care (dependent variables). Finally, we used exploded (i.e., rank-ordered) logit regression to examine the association between most likely class membership (independent variable) and ranked preferences across PrEP modalities.

Exploded logit regression estimates the relative preference for a given PrEP modality compared with a reference category based on sets of ranked items, rather than just the top-ranked choice.³⁶ Exponentiated coefficients were interpreted as ORs of choosing a given PrEP modality over daily oral PrEP (reference category). We also calculated model-estimated probabilities of each PrEP modality being ranked first.³⁷ In exploded logit models, the effect of covariates on ranked modality preferences is evaluated with interaction terms covariate*modality, and overall significance of the relationship is evaluated with likelihood ratio tests.³⁶ Adjusted regression models controlled for age, race/ethnicity, and sexual identity.

Mplus 8.7 was used for LCA and SAS 9.4 was used for descriptive statistics and regression models.

Results

We invited 895 eligible individuals to participate, and obtained N = 737 valid survey responses (response rate = 82.3%). Age, race, and PrEP awareness did not differ significantly between eligible individuals who completed the survey from those who did not (not shown). Participants' mean age was 21.13 years [standard deviation (SD) = 2.27]. About half (56%) was White, 19% identified as Black, 12% as Asian, 13% Multi-racial/Other, and 24% were Latino. Most participants identified as gay (80%) and 42% had a higher education degree. About one-third (34%) of participants were in the South, 28% in the Northeast, 19% in the West, and 18% in the Midwest. Table 2 shows additional sample characteristics.

Table 2.

Sociodemographic Characteristics and Measures of Access to Health Care (N = 737)

	Total (N = 737)	Class 1 (35%): Shared decision-making	Class 2 (25%): Provider-led decision-making	Class 3 (41%): Patient-driven decision-making	p
Age (M, SD)	21.13 (2.27)	21.57 (2.07)	20.84 (2.52)	20.94 (2.23)	<0.001
Age group, n (%)					0.01
15–17	62 (8)	10 (4)	23 (13)	28 (9)
18–20	200 (27)	65 (26)	50 (28)	85 (28)
21–24	475 (65)	179 (71)	109 (60)	187 (62)
Race					<0.001
White	412 (56)	164 (64)	66 (36)	182 (61)
Black	139 (19)	36 (14)	59 (32)	44 (15)
Asian	91 (12)	26 (10)	27 (15)	38 (13)
Multi-racial/other	95 (13)	29 (11)	30 (17)	36 (12)
Latinx, n (%)	174 (24)	53 (21)	50 (28)	71 (24%)	<0.001
Sexual identity, n (%)					0.13
Gay	587 (80)	212 (84)	142 (78)	233 (78)
Bisexual	111 (15)	28 (11)	32 (18)	51 (17)
Queer	29 (4)	9 (4)	5 (3)	15 (5)
Other	9 (1)	5 (2)	3 (2)	1 (0.3)
Educational attainment, n (%)					<0.001
High school or less	174 (24)	37 (15)	64 (35)	73 (24)
Some college or vocational school	250 (34)	87 (34)	55 (30)	108 (36)
College degree or higher	311 (42)	129 (51)	63 (35)	311 (42)
Food insecurity (past 3 months), n (%)	134 (18)	43 (17)	49 (27)	42 (14)	0.001
Region, n (%)					0.001
South	251 (34)	68 (27)	69 (27)	114 (45)
Northeast	209 (28)	95 (45)	48 (23)	66 (32)
West	137 (19)	41 (30)	37 (27)	59 (43)
Midwest	134 (18)	50 (37)	25 (19)	59 (44)

Latent classes

Fit indices and substantive interpretation favored a three-class solution with entropy = 0.81 (Table 3). Class 1 (35% of participants) was characterized by moderate-to-high endorsement of communication indicators and high endorsement of autonomy indicators. Given high communication and high autonomy, we labeled Class 1 “shared decision-making.” Class 2 (25%) presented moderate-to-high endorsement of communication indicators, but low endorsement of autonomy items, being labeled “provider-led decision-making.” Finally, Class 3 (41%) showed low-to-moderate endorsement of communication indicators and high endorsement of autonomy indicators and was labeled “patient-driven decision-making.” All classes presented moderate heterogeneous endorsement of stigma and mistrust in health care, and these indicators could not adequately separate classes (between-class ORs ranging 0.58–2.2; Table 4). Figure 1 shows proportion of indicator endorsement across classes.

FIG. 1.

Model-estimated item probability profile by class.

Table 3.

Latent Class Analysis Model Fit Indices of Candidate Models

Classes	Log-likelihood	Parameters in model	Adjusted χ ² LRT (df), p value	BIC	CAIC	LMR-LRT, p value	BLRT, p value
1	−2837.84	6	285.11 (57), p < 0.00001	5715.30	5721.30	—	—
2	−2740.26	13	89.94 (50), p = 0.0005	5566.35	5579.34	191.04, p < 0.00001	p < 0.00001
3	−2719.52	20	48.47 (43), p = 0.26	5571.10	5591.10	40.59, p = 0.0002	p < 0.00001
4	−2713.03	27	35.49 (36), p = 0.49	5604.33	5631.33	12.71, p = 0.06	p ≈ 1.00
5^*	—	—	—	—	—	—	—
6^*	—	—	—	—	—	—	—

Boldface indicates best-fitting model for each fit index measure. For the BIC, a difference <5 was not considered relevant and both two- and three-class solutions were considered adequate.³⁴

Models with five and six classes were empirically underidentified and model fit indices are not shown.

BIC, Bayesian information criterion; BLRT, bootstrap likelihood ratio test; CAIC, consistent Akaike's information criterion; df, degrees of freedom; LMR-LRT, adjusted Lo–Mendell–Rubin likelihood ratio test; LRT, likelihood ratio test.

Table 4.

Proportion of Indicator Endorsement and Between-Class Odds Ratios Across Classes

Three-class solution (entropy = 0.81)	Overall	Class 1: Shared decision-making (35%)	Class 2: Provider-led decision-making (25%)	Class 3: Patient-driven decision-making (41%)	OR_1:2	OR_1:3	OR_2:3
Domain 1: Sexual health communication with health care providers
Has provider who knows about sexual identity	44%	100%	41%	0%	Max	Max	Max
Believes providers should know everything about sexuality to take good care of them	58%	62%	74%	44%	0.58	2.07	3.58
Domain 2: Stigma and mistrust in health care
Experienced stigma attributed to sexual orientation in health care	38%	50%	35%	31%	1.82	2.20	1.21
Believes providers try to cover up when make mistakes	52%	43%	57%	56%	0.58	0.61	1.04
Domain 3: Autonomy in sexual health decisions
Believes that patients should be given choices and decide about what is best for their sexual health	63%	88%	4%	90%	183.7	0.82	0.01
Believes patients should make important medical decisions about sexual health	66%	88%	21%	86%	27.4	1.25	0.05

Boldface indicates probability of endorsement ≥70% or ≤30% or between-class ORs ≥5.0 or ≤0.2, considered to be good measures of homogeneity and class separation, respectively.³⁴

OR, odds ratio.

Individuals in the shared decision-making class were older than those in the provider-led and patient-driven decision-making classes (p < 0.001); individuals in the provider-led decision-making class were less likely to be White (p < 0.001), had lower educational attainment (p < 0.001), and were more likely to report food insecurity (p = 0.001) compared with those in the shared and patient-driven decision-making classes. Shared decision-making was more common among individuals in the Northeast (45%), whereas patient-driven decision-making was more common in the South (45%; p = 0.001) (Table 2).

Access to care and HIV prevention services

Most participants had health insurance (61% covered under their parents’/guardians' plan and 30% had their own health insurance), 84% had a regular health care provider, and 81% reported their health needs were met (Table 5). Compared with the shared decision-making class (class 1), individuals engaging in provider-led (class 2) and patient-driven decision-making (class 3) were less likely to have health needs met [adjusted odds ratio (aOR) = 0.32, p < 0.001; and aOR = 0.42, p < 0.001, respectively], have their own health insurance (aOR = 0.12, p < 0.001; and aOR = 0.12, p < 0.001, respectively) or be covered by their parents' insurance (aOR = 0.12, p < 0.001; and aOR = 0.16, p < 0.01, respectively), and have a primary care provider (p < 0.001, Table 5).

Table 5.

Associations Between Latent Class Membership and Engagement in General Health Care and HIV Prevention Services

Most likely class membership	Health needs met (81%)		Insurance status^a				HIV test (ever; 79%)		PrEP awareness (95%)		PrEP use^b
	Health needs met (81%)		Parents' insurance (61%)		Own insurance (30%)		HIV test (ever; 79%)		PrEP awareness (95%)		Current (19%)		Past (9%)
	OR	aOR	OR	aOR	OR	aOR	OR	aOR	OR	aOR	OR	aOR	OR	aOR
Class 1: Shared decision-making	REF	REF	REF	REF	REF	REF	REF	REF	REF	REF	REF	REF	REF	REF
Class 2: Provider-led decision-making	0.33^***	0.32^***	0.10^***	0.12^***	0.10^***	0.12^******	0.23^***	0.26^***	0.19^***	0.20^**	0.40^**	0.43^**	0.47^*	0.46^*
Class 3: Patient-driven	0.43^***	0.42^***	0.17^**	0.16^**	0.11^***	0.12^***	0.19^***	0.21^***	0.58	0.61	0.24^***	0.26^***	0.36^***	0.37^**

All individuals in the shared decision-making class (Class 1) reported having primary care providers, and, therefore, maximum likelihood-based logistic regression models with this dependent variable did not converge and are not shown; 77% of individuals in Class 2: Provider-led decision making and 75% of individuals in Class 3: Patient-driven decision-making reported having a primary care provider (χ² _(df ₌ ₂₎ = 71.41, p < 0.001).

Reference group: Not insured.

Reference group: No current or past PrEP use.

p < 0.05; ^** p < 0.01; ^*** p < 0.001.

aOR, adjusted odds ratio; PrEP, pre-exposure prophylaxis.

Adjusted models control for age, race/ethnicity, and sexual identity.

Lifetime HIV testing and PrEP awareness were prevalent in our sample (79% and 95%, respectively). Less than one-third reported ever taking PrEP (28%) and 19% were currently taking PrEP (Table 5). In comparison with shared decision-making, provider-led and patient-driven decision-making classes were less likely to have been tested for HIV (aOR = 0.26, p < 0.001; and aOR = 0.21, p < 0.001, respectively; Table 5). Provider-led decision-making was also associated with lower PrEP awareness compared with shared decision-making (aOR = 0.20, p < 0.001; Table 5).

In multinomial regression models, both provider-led and patient-driven decision-making were associated with lower past PrEP use (aOR = 0.46, p < 0.05; and aOR = 0.37, p < 0.01, respectively) and lower current PrEP use (aOR = 0.43, p < 0.01, and aOR = 0.26, p < 0.001; Table 5). There were no significant differences between provider-led and patient-driven decision-making in access to general health care and HIV prevention services, with the exception of PrEP awareness, which was greater among the patient-driven decision-making group (aOR = 3.08, p < 0.01, not shown).

Preferences across PrEP modalities

Overall, across all six PrEP modalities included in the study, daily oral PrEP had the highest probability of being ranked first [26%; mean ranking (Mr) = 2.71, SD = 1.48], followed by event-driven oral (23%; Mr = 2.82, SD = 1.59), intramuscular injections (22%; Mr = 2.91, SD = 1.54), subcutaneous implants (12%; Mr = 3.74, SD = 1.64), bnAb infusions (10%; Mr = 4.17, SD = 1.46), and rectal douches (6%; Mr = 6.4, SD = 1.53). Next-generation modalities were less likely to be preferred over daily oral PrEP, although this relationship was only marginally significant between event-driven and daily oral PrEP (p = 0.06, not shown). Figure 2 shows the predicted rankings of PrEP modalities.

FIG. 2.

Predicted ranking of different PrEP modalities across classes. PrEP, pre-exposure prophylaxis.

There were differences in PrEP preferences across classes (omnibus Wald χ ² _(df=10) = 23.34, p < 0.001), driven mainly by differences between shared and provider-led decision-making. The odds of choosing any next-generation modality over daily oral PrEP was lower among individuals in the provider-led decision-making group compared with individuals in the shared decision-making group (all adjusted ps < 0.05). Although the odds of preferring next-generation PrEP modalities over daily oral PrEP were of consistently lower magnitude in the patient-driven versus shared decision-making classes, this difference only reached statistical significance for PrEP implants (adjusted p < 0.05) (Table 6).

Table 6.

Preferences for Pre-Exposure Prophylaxis Modalities Across Classes

	Class 1: Shared decision-making		Class 2: Provider-led decision-making			Class 3: Patient-driven decision-making
	OR	aOR	OR	aOR	Class 1 versus 2 (p of interaction term in adjusted models)	OR	aOR	Class 1 versus 3 (p of interaction term in adjusted models)	Class 2 versus 3 (p of interaction term in adjusted models)
Daily oral PrEP	REF	REF	REF	REF	REF	REF	REF	REF	REF
bnAb	0.45^***	0.41^***	0.33^***	0.29^***	0.043^*	0.36^***	0.33^**	0.14	0.44
Rectal douche	0.28^***	0.22^***	0.19^***	0.14^***	0.013^*	0.25^***	0.19^***	0.39	0.07
PrEP implant	0.67^***	0.67^***	0.32^***	0.31^***	<0.0001^***	0.46^***	0.46^*	0.01^*	0.02^*
Injectable PrEP	1.00	1.05	0.64^***	0.66^*	0.008^**	0.83	0.86	0.20	0.10
Event-driven	1.00	0.96	0.67^**	0.64^**	0.015^*	0.94	0.89	0.63	0.04^*

Omnibus test for overall differences in preferences across classes: Wald χ ² _(df=10) = 23.34, p < 0.01 (unadjusted model); Wald χ ² _(df=10) = 22.17, p = 0.01 (adjusted models).

p < 0.05; ^** p < 0.01; ^*** p < 0.001.

bnAb, broadly neutralizing antibodies.

Adjusted models control for age, race/ethnicity, and sexual identity.

In stratified analyses, individuals in the shared decision-making class were less likely to prefer PrEP implants (OR = 0.67, p < 0.001), bnAb (OR = 0.45, p < 0.001), and rectal douches (OR = 0.28, p < 0.001) over daily oral PrEP, but rankings of injectable and event-driven oral PrEP were not different from daily oral PrEP (OR = 1.00 for both). Preferences among the patient-driven decision-making class were similar to that of the shared decision-making, with lower ORs indicating lower preferences for next-generation modalities in this class: daily oral PrEP was significantly preferred over implants (OR = 0.46, p < 0.001), bnAb (OR = 0.36, p < 0.001), and rectal douches (OR = 0.25, p < 0.001), but not event-driven oral (OR = 0.94, p = 0.56) and injectable (OR = 0.83, p = 0.06).

Among individuals in the provider-led decision-making class, all next-generation modalities were preferred less compared with daily oral PrEP: event-driven oral (OR = 0.67, p < 0.01), injectable (OR = 0.64, p < 0.001), bnAb (OR = 0.33, p < 0.001), implants (OR = 0.32, p < 0.001), and rectal douches (OR = 0.19, p < 0.001).

Adjustment for covariates did not meaningfully change directionality and magnitude of differences across classes. Table 6 and Fig. 3 show preferences for different PrEP modalities across classes.

FIG. 3.

Forest plot of adjusted odds of choosing PrEP modalities over daily oral PrEP (reference group). aOR, adjusted odds ratio.

Discussion

In this study, we identified three subgroups of YMSM based on sexual health communication and autonomy in sexual health decision-making. Notably, stigma and mistrust in health care were prevalent across all classes (reported by 31–57% of participants); however, these variables did not distinguish between classes. This finding supports prior studies^12,31,38,39 suggesting that structural discrimination against sexual minority individuals and the resulting medical mistrust remain major barriers to care among this group, and need to be addressed.

Bilateral communication and patient autonomy are key characteristics of shared decision-making.^20,40,41 Communication about sexual practices is a crucial element of PrEP indication evaluations and is needed to provide accurate information about HIV prevention options.⁴² However, communication alone does not necessarily lead to shared decisions, and, due to power imbalances in patient–provider relationships, providers may dominate the decision-making process even when communicating with patients.^20,43 For example, when discussing antiretroviral adherence challenges, providers often take a directive approach, ask few open-ended questions, and may fail to engage in problem-solving with patients.⁴⁴ Conversely, shared decision-making is patient-centered and emphasizes both communication and patient autonomy, enabling patients to make sexual health decisions that are best aligned with their preferences and needs.

Compared with provider-led decision-making (high provider communication but low patient autonomy) and patient-driven decision-making (high autonomy but low communication), shared decision-making was associated with several measures of health care access (e.g., insurance status, HIV testing, and PrEP use). Shared decision-making may lead to empowered patients well-positioned to navigate the health care system, communicate with providers, and access care. As such, shared decision-making has been advocated for in HIV treatment and prevention care, including PrEP.^45

–48

However, empirical evidence on the benefits of interventions to nurture shared decision-making skills among providers is inconclusive.⁴⁹ Beyond targeting attitudes and skills of patient and providers only, future interventions may benefit from also addressing organizational aspects of health services that may facilitate shared decision-making in care (e.g., sufficient time for medical visits, private environments, continuity in patient–provider relationships).⁴³ We encourage further studies to explore shared decision-making and the potential role of the clinic environment in the context of HIV prevention and PrEP, particularly as more PrEP options become available and HIV prevention decision-making becomes more complex.

For the past decade, several public health campaigns and interventions have raised awareness and knowledge about the efficacy, safety, and affordable ways to access daily oral PrEP among potential users.^50
–52 Conversely, given the investigational status of several of the next-generation PrEP modalities, less information about product characteristics is available, which may limit acceptability of these modalities. For example, some YMSM expressed concern about decreased efficacy and potential long-term toxicity of next-generation compared with daily oral PrEP before data on the efficacy and toxicity of these modalities was available as trials were underway.^53
–55

Moreover, although there have been advances in ensuring affordable access to daily oral PrEP,⁵⁶ insurance coverage, cost assistance programs, and out-of-pocket costs for next-generation modalities among YMSM have not been determined, and may influence modality preferences.^12,23 Future research should continue to investigate product preferences as more data on these modalities become available and feature in health education campaigns targeting potential users. Previous studies showed that social media advertisements, at times with misleading information, have played a role in daily oral PrEP users' preferences for tenofovir alafenamine over tenofovir disoproxil fumarate, and similar processes may happen with regard to next-generation modalities in the near future.^57,58

YMSM in both the provider-led and patient-driven decision-making groups were consistently less likely to choose next-generation over daily oral PrEP compared with individuals in the shared decision-making group. It is possible that barriers to acceptability of next-generation products (e.g., concerns about product efficacy, side effects, and cost) are less salient among individuals engaging in shared decision-making, who may be better connected to health care, more comfortable relying on providers and clinic staff for navigating insurance claims and cost-assistant programs, and more trusting in health care and research institutions.

This finding may indicate that subgroups of YMSM who are already engaging in shared decision-making with providers and have better access to HIV prevention services may be the ones most likely to initiate next-generation PrEP. This is of concern because youth who are not being reached by existing services may continue to face challenges to access next-generation PrEP products, which could limit the role of these technologies in curbing HIV transmission and widen PrEP access disparities.

Individuals in the provider-led decision-making class exhibiting low autonomy in sexual health decisions were particularly less likely to choose next-generation PrEP. This class was also younger, less educated, and more likely to be non-White, live in the South, and experience food insecurity compared with the other classes. It is possible that concerns about next-generation PrEP are greater among this class, leading to lower willingness to adopt these modalities. Since communication with providers is already in place, engaging providers may be viable to generate demand for next-generation PrEP modalities among this group.

Overall, however, PrEP readiness among providers caring for youth is low,^59,60 and next-generation PrEP modalities may further increase provider education needs. Therefore, it is important that educational resources for providers keep up with the ongoing development of new PrEP modalities. Provider training and education programs should consider developmental and sociocultural aspects in fostering shared decision-making and patient autonomy with diverse YMSM, including younger and socially disenfranchised youth.

Incorporating decision aids into PrEP care may also facilitate patient–provider conversations and contribute to shared informed decisions that are congruent with patients' values and needs.^45,61 Decision aids have been shown to increase patient knowledge and satisfaction across a variety of clinical contexts,^62,63 and can be integrated into medical visits and be available online, similar to contraception decision tools.⁶⁴ Importantly, since the effects of decision aids on patient knowledge may not be sustained,⁶⁴ frequent reassessment of PrEP indications and modality preferences should be incorporated into clinical follow-up of youth receiving HIV prevention services. Future intervention research examining how to best tailor content to meet the needs of these three latent classes of youth may inform future efforts to promote uptake of next-generation PrEP.

Our study has several limitations. First, although we used culturally appropriate, comprehensible descriptions of PrEP modalities developed specifically for use with YMSM,³³ data were collected before any of the next-generation PrEP modalities had been approved for use in the United States. Therefore, YMSM's preferences for these modalities are hypothetical and likely to change as these products are introduced for clinical use and more evidence about their efficacy is made available. Future studies should continue to explore preferences for PrEP products after they reach the market. Second, we evaluated a nonexhaustive list of next-generation PrEP modalities, and cannot draw inferences regarding other PrEP products being developed (e.g., subcutaneous injections and long-acting oral pills).⁶⁵ Third, even though we recruited a geographically and racially/ethnically diverse sample of YMSM in the United States, our online convenience sample limits the generalizability of our findings (e.g., individuals not engaging with YMSM-related content online).

Despite these limitations, we offer insights into subgroups of YMSM with respect to patterns of health care decision-making and how that may influence access to care and HIV prevention choices, which may inform the future implementation of PrEP services targeting youth. Identifying subgroups of YMSM based on their interactions with health care providers and sexual health services sheds light on youth who may be facing challenges to accessing general health care and HIV prevention services. Although there has been some enthusiasm and optimism regarding the potential of next-generation PrEP modalities to provide at-risk individuals with alternative HIV prevention options,^66,67 wide implementation of these modalities in the United States and globally will depend on addressing barriers to PrEP situated at health system and provider levels. Shared decision-making may be a promising framework to address these multi-level barriers and promote engagement in the PrEP care continuum, while also developing YMSM's strengths and autonomy.

Footnotes

Acknowledgments

The authors would like to thank research participants, staff, and the Youth Advisory Board at University of Pennsylvania for their support.

Authors' Contributions

P.K.V. and R.D. contributed to data collection and analysis and writing of findings; J.A.B. and K.B.B. contributed to the conceptualization of the study, data collection and analysis, and writing of findings; W.Y.L. contributed to data collection and writing of findings; D.T.D.S., D.O., and J.R. contributed to writing of findings; L.H.-W. and K.H.M. contributed to the conceptualization of the study and writing of findings.

Author Disclosure Statement

K.H.M. received unrestricted research grants from Gilead and Merck; on the scientific advisory board of Gilead, Merck, and ViiV. K.B.B. received unrestricted research grants from Merck.

Funding Information

This study was made possible through support from the National Institute of Child Health and Human Development (NICHD; U19HD089881). D.T.D.S. is supported by the Agency for Healthcare Research and Quality (AHRQ; T32HS026116-03). J.R. is supported by the National Institute of Mental Health (NIMH; K23MH123335-01).

References

Mustanski

, Newcomb

, Du Bois

, Garcia

, Grov

. HIV in young men who have sex with men: A review of epidemiology, risk and protective factors, and interventions. J Sex Res, 2011; 48:218–253.

Macapagal

, Moskowitz

, Li

, et al. Hookup app use, sexual behavior, and sexual health among adolescent men who have sex with men in the United States. J Adolesc Health, 2018; 62:708–715.

Hojilla

, Hurley

, Marcus

, et al. Characterization of HIV preexposure prophylaxis use behaviors and HIV incidence among US adults in an Integrated Health Care System. JAMA Netw Open, 2021; 4:e2122692.

Moskowitz

, Moran

, Matson

, Alvarado-Avila

, Mustanski

. The PrEP cascade in a national cohort of adolescent men who have sex with men. J Acquir Immune Defic Syndr, 2021; 86:536–543.

Pre-exposure Prophylaxis for the Prevention of HIV Infection in the United States—2021 Update. 2021. Available at: https://www.cdc.gov/hiv/pdf/risk/prep/cdc-hiv-prep-guidelines-2021.pdf [Last accessed: February 2, 2022].

Core indicators for Monitoring the Ending the HIV Epidemic Initiative (Preliminary Data): National HIV Surveillance System Data Reported Through June 2021; and Preexposure Prophylaxis (PrEP) Data Reported Through March 2021. 2021. Available at: https://www.cdc.gov/hiv/pdf/library/reports/surveillance-data-tables/vol-2-no-5/cdc-hiv-surveillance-tables-vol-2-no-5.pdf [Last accessed: December 21, 2021].

Stupiansky

, Liau

, Rosenberger

, et al. Young men's disclosure of same sex behaviors to healthcare providers and the impact on health: Results from a US national sample of young men who have sex with men. AIDS Patient Care STDs, 2017; 31:342–347.

Fisher

, Fried

, Macapagal

, Mustanski

. Patient–provider communication barriers and facilitators to HIV and STI preventive services for adolescent MSM. AIDS Behav, 2018; 22:3417–3428.

Holloway

, Tan

, Gildner

, et al. Facilitators and barriers to pre-exposure prophylaxis willingness among young men who have sex with men who use geosocial networking applications in California. AIDS Patient Care STDs, 2017; 31:517–527.

10.

Dean

, Meyer

, Robinson

, et al. Lesbian, gay, bisexual, and transgender health: Findings and concerns. J Gay Lesbian Med Assoc, 2000; 4:102–151.

11.

Qiao

, Zhou

, Li

. Disclosure of same-sex behaviors to health-care providers and uptake of HIV testing for men who have sex with men: A systematic review. Am J Mens Health, 2018; 12:1197–1214.

12.

Valente

, Bauermeister

, Lin

, et al. Next generation pre-exposure prophylaxis for young men who have sex with men: Lessons from system and provider-level barriers to oral PrEP. AIDS Behav, 2022; 26:3422–3435.

13.

Phillips

, Raman

, Felt

, Han

, Mustanski

. Factors associated with PrEP support and disclosure among YMSM and transgender individuals assigned male at birth in Chicago. AIDS Behav, 2019; 23:2749–2760.

14.

Ubel

, Scherr

, Fagerlin

. Empowerment failure: How shortcomings in physician communication unwittingly undermine patient autonomy. Am J Bioeth, 2017; 17:31–39.

15.

Schmitz

, Tabler

. Health services and intersections of care: Promises and pitfalls experienced by LGBTQ + Latino/a emerging adults. J LGBT Youth, 2021; 18:1–22.

16.

Quinn

, Dickson-Gomez

, Zarwell

, Pearson

, Lewis

. “A gay man and a doctor are just like, a recipe for destruction”: How racism and homonegativity in healthcare settings influence PrEP uptake among young Black MSM. AIDS Behav, 2019; 23:1951–1963.

17.

Schwartz

, Grimm

. Stigma communication surrounding PrEP: The experiences of a sample of men who have sex with men. Health Commun, 2019; 34:84–90.

18.

Cahill

, Taylor

, Elsesser

, Mena

, Hickson

, Mayer

. Stigma, medical mistrust, and perceived racism may affect PrEP awareness and uptake in black compared to white gay and bisexual men in Jackson, Mississippi and Boston, Massachusetts. AIDS Care, 2017; 29:1351–1358.

19.

Brooks

, Nieto

, Cabral

, Landrian

, Fehrenbacher

. Delivering PrEP to adults with “low” or “no” HIV risk and youth: Experiences and perspectives of PrEP providers. Cult Health Sex, 2022; 24:95–108.

20.

Charles

, Gafni

, Whelan

. Shared decision-making in the medical encounter: What does it mean? (or it takes at least two to tango). Soc Sci Med, 1997; 44:681–692.

21.

Mayer

, Allan-Blitz

L-T

. PrEP 1.0 and beyond: Optimizing a biobehavioral intervention. J Acquir Immune Defic Syndr, 2019; 82 Suppl 2:S113–S117.

22.

Bauermeister

, Tingler

, Dominguez

, et al. Acceptability of a dapivirine/placebo gel administered rectally to HIV-1 seronegative adults (MTN-026). AIDS Behav, 2022; 26:1333–1346.

23.

Macapagal

, Nery-Hurwit

, Matson

, Crosby

, Greene

. Perspectives on and preferences for on-demand and long-acting PrEP among sexual and gender minority adolescents assigned male at birth. Sex Res Social Policy, 2021; 18:39–53.

24.

Rogers

EM.

Diffusion of Innovations. The Free Press: New York, NY; 1995.

25.

Greenhalgh

, Robert

, Macfarlane

, Bate

, Kyriakidou

. Diffusion of innovations in service organizations: Systematic review and recommendations. Milbank Quart, 2004; 82:581–629.

26.

Dearing

, Cox

. Diffusion of innovations theory, principles, and practice. Health Aff, 2018; 37:183–190.

27.

Bauermeister

, Pingel

, Zimmerman

, Couper

, Carballo-Diéguez

, Strecher

. Data Quality in web-based HIV/AIDS research: Handling invalid and suspicious data. Field Methods, 2012; 24:272–291.

28.

Guest

, Adam

, Lucas

, et al. Methods for authenticating participants in fully web-based mobile app trials from the iReach Project: Cross-sectional Study. JMIR Mhealth Uhealth, 2021; 9:e28232.

29.

Teitcher

JEF

, Bockting

, Bauermeister

, Hoefer

, Miner

, Klitzman

. Detecting, preventing, and responding to “fraudsters” in internet research: Ethics and tradeoffs. J Law Med Ethics, 2015; 43:116–133.

30.

Flynn

, Smith

, Vanness

. A typology of preferences for participation in healthcare decision making. Soc Sci Med, 2006; 63:1158–1169.

31.

Eaton

, Driffin

, Kegler

, et al. The role of stigma and medical mistrust in the routine health care engagement of black men who have sex with men. Am J Public Health, 2015; 105:e75–e82.

32.

Mullins

TLK

, Zimet

, Lally

, Kahn

, Interventions

AMTNfHA

. Adolescent human immunodeficiency virus care providers' attitudes toward the use of oral pre-exposure prophylaxis in youth. AIDS Patient Care STDs, 2016; 30:339–348.

33.

Biello

, Valente

, Lin

, et al. PrEParing for NextGen: Cognitive interviews to improve next generation PrEP modality descriptions for young men who have sex with men. AIDS Behav, 2022; 26:1956–1965.

34.

Masyn

KE.

Latent Class Analysis and Finite Mixture Modeling. Oxford University Press: New York, NY;, 2013:551.

35.

Clark

, Muthén

. Relating Latent Class Analysis Results to Variables Not Included in the

Analysis

. 2009. Available from: http://www.statmodel.com/download/relatinglca.pdf

36.

Allison

, Christakis

. Logit models for sets of ranked items. Sociol Methodol, 1994:199–228.

37.

Gardiner

, Luo

Logit Models in Practice: B, C, E, G, M, N, O…. Washington: SAS Global Forum; 2011.

38.

Graham

, Berkowitz

, Blum

, et al. The Health of Lesbian, Gay, Bisexual, and Transgender People: Building a Foundation for Better Understanding. Washington, DC: Institute of Medicine;, 2011; 10:13128.

39.

Edeza

, Karina Santamaria

, Valente

, Gomez

, Ogunbajo

, Biello

. Experienced barriers to adherence to pre-exposure prophylaxis for HIV prevention among MSM: A systematic review and meta-ethnography of qualitative studies. AIDS Care, 2021; 33:697–705.

40.

Charles

, Gafni

, Whelan

. Decision-making in the physician–patient encounter: Revisiting the shared treatment decision-making model. Soc Sci Med, 1999; 49:651–661.

41.

Peek

, Lopez

, Williams

, et al. Development of a conceptual framework for understanding shared decision making among African-American LGBT patients and their clinicians. J Gen Intern Med, 2016; 31:677–687.

42.

Nunn

, Brinkley-Rubinstein

, Oldenburg

, et al. Defining the HIV pre-exposure prophylaxis care continuum. AIDS, 2017; 31:731.

43.

Joseph-Williams

, Elwyn

, Edwards

. Knowledge is not power for patients: A systematic review and thematic synthesis of patient-reported barriers and facilitators to shared decision making. Patient Educ Couns, 2014; 94:291–309.

44.

Laws

, Beach

, Lee

, et al. Provider-patient adherence dialogue in HIV care: Results of a Multisite Study. AIDS Behav, 2013; 17:148–159.

45.

Valente

, Wu

, Cohen

, Caskey

, Meyers

. Behavioral and social science research to support development of educational materials for clinical trials of broadly neutralizing antibodies for HIV treatment and prevention. Clin Trials, 2021; 18:17–27.

46.

Calabrese

, Magnus

, Mayer

, et al. “Support your client at the space that they're in”: HIV pre-exposure prophylaxis (PrEP) prescribers' perspectives on PrEP-related risk compensation. AIDS Patient Care STDs, 2017; 31:196–204.

47.

Taylor

, Fornos

, Tarbutton

, et al. Improving HIV care engagement in the south from the patient and provider perspective: The role of stigma, social support, and shared decision-making. AIDS Patient Care STDs, 2018; 32:368–378.

48.

McNulty

, Acree

, Kerman

, Williams

HHS

, Schneider

. Shared decision making for HIV pre-exposure prophylaxis (PrEP) with black transgender women. Cult Health Sex, 2022; 24:1033–1046.

49.

Légaré

, Adekpedjou

, Stacey

, et al. Interventions for increasing the use of shared decision making by healthcare professionals. Cochrane Database Syst Rev, 2018; 7:CD006732.

50.

Sophus

, Mitchell

. A review of approaches used to increase awareness of pre-exposure prophylaxis (PrEP) in the United States. AIDS Behav, 2019; 23:1749–1770.

51.

Kudrati

, Hayashi

, Taggart

. Social media & PrEP: A systematic review of social media campaigns to increase PrEP Awareness & uptake among young Black and Latinx MSM and women. AIDS Behav, 2021; 25:4225–4234.

52.

Finlayson

, Cha

, Xia

, et al. Changes in HIV preexposure prophylaxis awareness and use among men who have sex with men—20 urban areas, 2014 and 2017. MMWR Morb Mortal Wkly Rep, 2019; 68:597–603.

53.

Patel

, Crane

, López

, et al. Pre-exposure prophylaxis for HIV prevention preferences among young adult African American men who have sex with men. PLoS ONE, 2018; 13:e0209484.

54.

Calder

, Schieffer

, Bryndza Tfaily

, et al. Qualitative consumer research on acceptance of long-acting pre-exposure prophylaxis products among men having sex with men and medical practitioners in the United States. AIDS Res Hum Retroviruses, 2018; 34:849–856.

55.

Biello

, Hosek

, Drucker

, et al. Preferences for injectable PrEP among young U.S. Cisgender men and transgender women and men who have sex with men. Arch Sex Behav, 2018; 47:2101–2107.

56.

Ready

Set

, PrEP. 2021. Available at: https://www.hiv.gov/federal-response/ending-the-hiv-epidemic/prep-program [Last accessed: May 20, 2022].

57.

Grov

, Westmoreland

, D'Angelo

, et al. Marketing of tenofovir disoproxil fumarate (TDF) lawsuits and social media misinformation campaigns' impact on PrEP uptake among gender and sexual minority individuals. AIDS Behav, 2021; 25:1396–1404.

58.

D'Angelo

, Westmoreland

, Carneiro

, Johnson

, Grov

. Why are patients switching from tenofovir disoproxil fumarate/emtricitabine (Truvada) to tenofovir alafenamide/emtricitabine (Descovy) for pre-exposure prophylaxis?. AIDS Patient Care STDs, 2021; 35:327–334.

59.

Hart-Cooper

, Allen

, Irwin

, Scott

. Adolescent health providers' willingness to prescribe pre-exposure prophylaxis (PrEP) to youth at risk of HIV infection in the United States. J Adolesc Health, 2018; 63:242–244.

60.

Mullins

TLK

, Zimet

, Lally

, Xu

, Thornton

, Kahn

. HIV care providers' intentions to prescribe and actual prescription of pre-exposure prophylaxis to at-risk adolescents and adults. AIDS Patient Care STDs, 2017; 31:504–516.

61.

Krakower

, Mayer

. Engaging healthcare providers to implement HIV pre-exposure prophylaxis. Curr Opin HIV AIDS, 2012; 7:593–599.

62.

Stacey

, Légaré F, Lewis K, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Database Syst Rev, 2017; 4:Cd001431.

63.

Coronado-Vázquez

, Canet-Fajas

, Delgado-Marroquín

, Magallón-Botaya

, Romero-Martín

, Gómez-Salgado

. Interventions to facilitate shared decision-making using decision aids with patients in primary health care: A systematic review. Medicine (Baltimore), 2020; 99:e21389-e.

64.

Jones

, Allison

, Perry

. Effectiveness of contraceptive decision aids in adolescents and young adults: A systematic review. J Pediatr Adolesc Gynecol, 2022; 35:7–17.

65.

The Years Ahead in Biomedical HIV Prevention Research. 2021. Available at: https://www.avac.org/infographic/years-ahead-hiv-prevention-research [Last accessed: December 29, 2021].

66.

Castor

, Meyers

, Allen

. The only way is up: Priorities for implementing long-acting antiretrovirals for HIV prevention and treatment. Curr Opin HIV AIDS, 2020; 15:73–80.

67.

Thoueille

, Choong

, Cavassini

, Buclin

, Decosterd

. Long-acting antiretrovirals: A new era for the management and prevention of HIV infection. J Antimicrob Chemother, 2022; 77:290–302.