Sociocultural influences on attitudes towards pre-exposure prophylaxis (PrEP), history of PrEP use, and future PrEP use in HIV-vulnerable cisgender men who have sex with men across the U.S.

Abstract

Despite its proven effectiveness in reducing HIV transmission, pre-exposure prophylaxis (PrEP) use remains low. This study used data from a 2017–2018 U.S. national cohort to investigate social influences on PrEP experience and future PrEP use among cisgender men who have sex with men. We used descriptive statistics and multivariable logistic analyses to examine social influences (e.g., how participants heard about PrEP and number of persons they knew taking PrEP) associated with each previous PrEP use and intentions to use PrEP. Among participants who knew of PrEP, commonly reported ways of first hearing about PrEP were through social media (27.4%) and friends (26.8%). These were also cited top influences on participants’ current attitudes toward PrEP (friends 23.5%, social media 22.1%). Multivariable logistic regression analyses found that knowing more people taking PrEP was associated with increased odds of previously using PrEP and intending to use PrEP. Friends and social media were common and influential sources of information regarding PrEP. Results suggest that tapping into these social connections may effectively disseminate public health messaging about PrEP and encourage use among key populations to reduce HIV burden.

INTRODUCTION:

The burden of HIV in the United States (U.S.) is largely among young men who have sex with men (MSM)¹ and gender minorities.^2,3 For the past several years, the national estimated total new HIV infections has been approximately 40,000, with young MSM accounting for two-thirds of those new infections.¹ One of the relatively new biomedical interventions being used to prevent HIV transmission is pre-exposure prophylaxis (PrEP)— HIV prophylactic drugs that are given to individuals who are thought to be at highest risk for contracting HIV.⁴ PrEP has been found to be highly effective at reducing HIV acquisition if exposed;⁵ however, despite its proven effectiveness and support from leading public health agencies, PrEP uptake lags behind need.^6,7 Therefore, more research is needed to determine innovative mediums that can be used to disseminate positive public health messaging about PrEP.

Despite public health efforts to deliver messaging around the benefits of health interventions (including PrEP), peer-to-peer discussions and social networks (including those online), continue to play a pivotal role in providing health advice^8–13 and influencing health behaviors.^14–17 Prior research has indicated that persons seeking health advice were more likely to seek it from family or friends than medical professionals especially among persons who have lower income.¹² Further, research has also supported that seeking health information from sources beyond medical professionals can independently and positively influence engaging in healthy behavior.¹⁸ Social norms and networks provide the basis for two important theories in understanding human behavior, Social Learning Theory¹⁹ and Theory of Reasoned Action.²⁰ Both of these theories propose that behavioral norms in social groups and communities are important for exposure to certain behaviors, determining social group-approved behavior, and influencing individual to engage in those behaviors.^19,20 For example, pro-condom views among peers and those providing social support have been shown to increase condom use among MSM.^15–17 Additionally, using social network members as sources of health information can aid in interpretability of health advice. Often, people rely on the “distributed health literacy” they have access to within their social support groups to seek health information, share health knowledge, and make informed health decisions.²¹ Indeed, several studies have indicated that social connections are important sources of information on PrEP^22–24 instead of medical providers; However, learning about PrEP from a medical provider was more successful in encouraging PrEP uptake compared to social media or other sources (e.g., newspaper).²⁵ This is particularly important when considering the some-what mixed reputation of PrEP^26–29 within the LGBTQ+ community. This mixed reputation includes feeling that PrEP is a useful tool to prevent HIV,²⁸ stigma surrounding those taking PrEP (i.e. “PrEP whores”²⁹) and the potential for increased “risk” behavior,^26,27 as well as individual concerns about PrEP barriers such as accessibility, affordability, and side effects.³⁰ Given that hearing negative information about PrEP through social networks could dissuade uptake or lead to discontinuation, even for appropriate candidates, medical provider influence on PrEP use represents a key point of intervention.²⁶ Further, medical providers are gatekeepers to PrEP access. Thus, considerable effort has been made to educate providers about PrEP³¹ and in providing culturally competent and sensitive³² care for MSM seeking PrEP. However, the role that social connections play in providing information about PrEP, and influencing PrEP use and other sexual behaviors, suggests that public health messaging about PrEP should utilize peer-networks, as well as formal clinical avenues for public health message dissemination. Thus, messages should be approachable for all health literacies for both formal (i.e. medical provider) and informal (e.g., friends, social media) modes of dissemination.

For researchers and practitioners interested in harnessing peer influences to propagate healthy sexual behaviors such as PrEP use, we must first understand the direction of the influences that peers may have on PrEP use. The goal of this study was to describe the associations between social group PrEP use and attitudes toward PrEP, with individual prior PrEP use and future intentions to use PrEP among a national cohort of cisgender MSM. Additionally, we provide a description of ways in which study participants first heard about PrEP and what or whom has been most influential on their current attitudes toward PrEP.

METHODS:

The Together 5000 Cohort:

Data were collected as part of Together 5000 (T5K), a U.S. national, internet-based cohort study designed to better understand intervenable factors and factors associated with PrEP uptake among those most vulnerable for HIV seroconversion among cisgender men who have sex with men. This study was approved by the Institutional Review of the City University of New York (CUNY) Graduate School of Public Health and Health Policy. The cohort and its enrollment procedures have been fully described elsewhere.^33,34 Briefly, participants were recruited from ads appearing on men-for-men geosocial sexual networking smartphone applications between October 2017 and June 2018. Core inclusion criteria specified that participants for this study were aged 16 to 49; had at least two sex partners who identified as men in the past three months; were not currently on PrEP; lived in the U.S. or its territories; and were not known to be HIV-positive at enrollment. Other inclusion criteria were additional demographic characteristics and behaviors that would place potential participants at increased risk for HIV acquisition and make them prime candidates for PrEP use that are detailed elsewhere.^33,34 Recruitment was targeted towards cisgender men who have sex with men; however, our enrollment criteria permitted trans men and trans women who otherwise met study criteria to enroll. That being said, the number of transgender men (n = 39), transgender women (n = 37), and those identifying as another gender category (n = 66) was low, and thus insufficient for analyses in the present study.

Approximately 23,000 potential participants completed a brief online screening survey of demographic and sexual behaviors, and 8,764 of those who completed screening were eligible for enrollment. These participants were then emailed another link containing a follow-up survey that assessed additional sexual, substance use, and psychosocial factors (n = 6,270). For the present study, we limited analyses to participants who completed this secondary survey, identified as cisgender men, and had ever heard of PrEP (n = 5, 817).

Measures:

PrEP experience.

The outcomes of interest for the current analyses were PrEP experiences and intentions to use PrEP. Participants answered the following question about their PrEP status at baseline, “Have you ever been prescribed HIV medications (e.g. Truvada) for use as PrEP (pre-exposure prophylaxis)?”As mentioned in the inclusion criteria above, potential participants were only enrolled into the cohort if they were not taking PrEP at the time of enrollment, therefore participants were categorized as reporting prior but not current PrEP use or being PrEP-naïve (never having used PrEP).

Intentions to use PrEP.

Intentions to use PrEP was assessed using the following question, “PrEP is currently available with a prescription from your doctor, and research has shown that a majority of insurance companies cover most or all of the costs of PrEP. Do you plan to begin PrEP?” Participants were dichotomized as reporting that they intended to use PrEP (yes) or that they had little/no intentions of using PrEP.

Number of network members known to be on PrEP.

The main independent variables of interest were network influences that could be associated with PrEP use or intended use. Participants were asked to report the number of people—zero, one, two, three, four, and more than five—that they knew on PrEP, which was categorized into none, 1–4 people, and 5 or more people based on the distribution of responses in the sample.

Attitudes toward PrEP within networks.

Participants were also asked to respond to the following prompt, “In general, I would say the people I am close with.” Participants could choose a response from, “Are in favor of PrEP,” “Are opposed to PrEP,” “Are split evenly between being in favor of and being opposed to PrEP,” “Don’t have strong opinions either way,” and “Don’t really know what PrEP is.”

How did you first hear about PrEP?

Other sociocultural influences on PrEP that were only evaluated in descriptive analyses included how/from whom participants first heard about PrEP. Answer choices were, “as part of participation in a research study,” “through a news media source,” “through a social media source,” “through a friend,” “through my main partner,” “through a casual sex partner,” “through a family member,” “through a medical provider,” “through a community-based agency,” “other,” or “I don’t remember.” Participants could only choose one method of first hearing about PrEP.

What has been most influential on your current attitudes towards PrEP?

Along the same lines, we also asked participants to report who or what has been the most influential on participants’ current PrEP views. Participants could choose one response from the same as “how did you first hear about PrEP”.

Other covariates.

Finally, we assessed many other demographic, socioeconomic, and behavioral characteristics known a priori from the literature to be associated with PrEP use and other factors associated with HIV risk.^35–37 These factors were age, gender, race/ethnicity, sexual orientation, employment status, education status, annual income, housing instability, and engaging in recent sex work (past 3 months).

Analyses:

Our objective was to assess for associations between network influences and former and intended PrEP use. Descriptive statistics (i.e. frequencies and percentages) were used to provide a profile of study participants and to assess differences (chi-squared tests) between prior PrEP users and PrEP-naïve, as well as between participants intending to use PrEP or those with no/little intentions to use PrEP. We then used two separate, adjusted models to assess associations between our PrEP outcomes of interest, network influences, and other covariates/potential confounders. First, we used a multivariable logistic regression model to determine the association of broader network influences on prior PrEP use. Then, we used another multivariable logistic regression model to assess broader network influences on intended future PrEP use. For both regression models, we report adjusted odds ratios (aORs) and 95% confidence intervals (95% CIs). All analyses were completed using SAS 9.4.

RESULTS:

Description of Together 5000 participants:

Half (50.5%) of our participants were between 25-35 years old, with nearly one-quarter (23.5%) being 16-24 years old. Most 85.4% of all participants identified as gay, queer, or homosexual. Almost half (46.9%) of participants identified as Black or African American, Asian or Pacific Islander, or other or multiracial; and 47.9% lived in the South. Similarly, nearly half (44.8%) had at least some college education or technical training and 63.2% were employed full-time. Almost one-third (31.9%) reported an annual income of less than $20,000, and 73.3% had health insurance. Twenty (19.6%) percent had experienced housing instability in the past 5 years, 13.9% had engaged in sex work in the past three months, and 14% had been incarcerated (in lifetime). Forty percent (40.4%) did not know their HIV status at enrollment. Most (85.1%) participants had never used PrEP, but over half (53.3%) reported that they had intentions to start using PrEP (Table 1).

Table 1.

Description of study participants and bivariate associations of characteristics with PrEP experiences and intention to use PrEP, together 5000 study, 2017-2018.

	Total n = 5,817		PrEP Experience						Intend to use PrEP
			PrEP-naïve n = 4948		Former PrEP n = 869				Yes n = 3100		Little or no n = 2717
Characteristic	Frequency	(%)	Frequency	(%)	Frequency	(%)	Chi-square	P	Frequency	(%)	Frequency	(%)	Chi-square	P
Age							56.0	<.0001					30.5	<.0001
16-24 years old	1369	23.5	1243	25.1	126	14.5			786	25.4	583	21.5
25-35 years old	2935	50.5	2437	49.3	498	57.3			1583	51.1	1352	49.8
36-45 years old	1195	20.5	989	20.0	206	23.7			596	19.2	599	22.1
46+ years old	318	5.5	279	5.6	39	4.5			135	4.4	183	6.7
Race/Ethnicity							5.5	0.24					48.6	<.0001
White	3087	53.1	2624	53.0	463	53.3			1535	49.5	1552	57.1
Black or African American	626	10.8	533	10.8	93	10.7			381	12.3	245	9.0
Latino	1397	24.0	1205	24.4	192	22.1			791	25.5	606	22.3
Asian or Pacific Islander	208	3.6	167	3.4	41	4.7			135	4.4	73	2.7
Other or multiracial	499	8.6	419	8.5	80	9.2			258	8.3	241	8.9
Sexual Orientation							18.5	<.0001					1.3	0.52
Gay, Queer, Homosexual	4970	85.4	4188	84.6	782	90.0			2658	85.7	2312	85.1
Bisexual	793	13.6	711	14.4	82	9.4			417	13.5	376	13.8
Other	54	0.9	49	1.0	5	0.6			25	0.8	29	1.1
Region							69.2	<.0001					6.8	0.15
Northeast	880	15.1	686	13.9	194	22.4			436	14.1	444	16.3
Midwest	864	14.9	740	15.0	124	14.3			461	14.9	403	14.8
South	2783	47.9	2456	49.7	327	37.7			1518	49.0	1265	46.6
West	1266	21.8	1043	21.1	223	25.7			672	21.7	594	21.9
US Overseas/Territories	22	0.4	22	0.4	0	0.0			11	0.4	11	0.4
Education							48.1	<.0001					18.3	0.000
< High school diploma	121	2.1	104	2.1	17	2.0			73	2.4	48	1.8
High school diploma or GED	795	13.7	726	14.7	69	7.9			459	14.8	336	12.4
Some college or technical school training	2603	44.8	2238	45.2	365	42.0			1413	45.6	1190	43.8
College graduate +	2298	39.5	1880	38.0	418	48.1			1155	37.3	1143	42.1
Employment Status							12.7	0.005					1.6	0.65
Full-time (40 hours per week)	3676	63.2	3127	63.2	549	63.2			1975	63.7	1701	62.6
Part-time (less than 40 hours per week)	738	12.7	615	12.4	123	14.2			394	12.7	344	12.7
Working or full-time student	871	15.0	769	15.5	102	11.7			447	14.4	424	15.6
Unemployed/Other	532	9.2	437	8.8	95	10.9			284	9.2	248	9.1
Annual Income							17.7	0.0001					22.9	<.0001
Less than $20,000	1856	31.9	1629	32.9	227	26.1			1021	32.9	835	30.7
$20,000-$49,999	2447	42.1	2064	41.7	383	44.1			1352	43.6	1095	40.3
50,000+	1514	26.0	1255	25.4	259	29.8			727	23.5	787	29.0
Martial Status							5.9	0.02					1.2	0.27
Yes	749	12.9	615	12.4	134	15.4			385	12.4	364	13.4
No	5068	87.1	4333	87.6	735	84.6			2715	87.6	2353	86.6
Health Insurance							2.0	0.17					0.8	0.37
Yes	4265	73.3	3611	73.0	654	75.3			2288	73.8	1977	72.8
No/do not know	1552	26.7	1337	27.0	215	24.7			812	26.2	740	27.2
Experience Housing Instability (past 5 years)							7.6	0.01					17.1	<.0001
No/Not within last 5 years	4677	80.4	4008	81.0	669	77.0			2430	78.4	2247	82.7
Yes, within last five years	1140	19.6	940	19.0	200	23.0			670	21.6	470	17.3
Engaged in Sex Work (past 3 months)							4.7	0.03					2.5	0.12
No	5010	86.1	4282	86.5	728	83.8			2649	85.5	2361	86.9
Yes	807	13.9	666	13.5	141	16.2			451	14.6	356	13.1
Ever Incarcerated							0.1	0.79					0.5	0.47
No	5004	86.0	4259	86.1	745	85.7			2657	85.7	2347	86.4
Yes	813	14.0	689	13.9	124	14.3			443	14.3	370	13.6
People known on PrEP							222.9	<.0001					34.8	<.0001
None	1382	23.8	1290	26.1	92	10.6			648	20.9	734	27.0
1-4 people on PrEP	2539	43.7	2226	45.0	313	36.0			1367	44.1	1172	43.1
5+ people	1896	32.6	1432	28.9	464	53.4			1085	35.0	811	29.9
Favorability of PrEP within network							84.9	<.0001					282.2	<.0001
Are in favor of PrEP	3784	65.1	3123	63.1	661	76.1			2315	74.7	1469	54.1
Are opposed to PrEP	37	0.6	29	0.6	8	0.9			17	0.6	20	0.7
Are split evenly between being in favor of and being opposed to PrEP	631	10.9	537	10.9	94	10.8			237	7.7	394	14.5
Don’t have strong opinions either way	710	12.2	652	13.2	58	6.7			249	8.0	461	17.0
Don’t really know what PrEP is	655	11.3	607	12.3	48	5.5			282	9.1	373	13.7
Self-reported HIV Status							122.0	<.0001					1.7	0.19
HIV-negative	3465	59.6	2800	56.6	665	76.5			1822	58.8	1643	60.5
I don’t know: I am unsure	2352	40.4	2148	43.4	204	23.5			1278	41.2	1074	39.5
Experience with PrEP													63.3	<.0001
PrEP naive	4948	85.1							2529	81.6	2419	89.0
Yes, but not currently taking PrEP	869	14.9							571	18.4	298	11.0
Intentions to use PrEP
Yes	3100	53.3
Little or no intention	2717	46.7

Description of sociocultural influences on PrEP knowledge and views:

Participants most frequently reported first hearing about PrEP through a social media source (27.9%) or a friend (27.2%) (Table 2). Of note, only five (5%) percent of participants reported first hearing about PrEP from a medical provider. Similarly, the highest proportions of participants reported that friends (24%) and social media (22.5%) sources were the most influential on their current views of PrEP. Bivariate analyses indicated differences for both PrEP experiences (PrEP-naïve vs. prior PrEP users; p-value < .0001) and intentions to use PrEP (Yes vs. little or no intentions; p-value < .0001) between different modes of first hearing about PrEP. We also found that bivariate analyses indicated statistically significant differences for PrEP experiences (PrEP-naive vs. prior PrEP users; p-value < .0001) and intentions to use PrEP (Yes vs. little or no intentions; p-value < .0001) between the differing modes of influences on participants’ current PrEP views. That is, compared to PrEP-naïve individuals, former PrEP users were significantly more likely to say they first heard about PrEP from a medical provider (13.6%) and that medical providers were the most influential source on their current views toward PrEP (24.2%). Similarly, higher proportions of participants who intended to start using PrEP reported first hearing about PrEP from their medical provider (5.1%) and that their medical provider was most influential on their current PrEP views (10.1%) compared to participants not intending to begin PrEP.

Table 2.

How study participants reported they first heard about PrEP and what is has been most influential on their current views of PrEP, together 5000 study, 2017–2018

First hear about PrEP
				PrEP experience							Intend to use PrEP¥
				PrEP Naïve		Former PrEP					Yes		Little or no
	Total			n = 4948		n = 869					n = 3100		n = 2717
	Frequency	(%)		Frequency	(%)	Frequency	(%)	Chi-square	P		Frequency	(%)	Frequency	(%)	Chi-square	P
								206.6	<.0001						59.7	<.0001
Social media source	1625	27.9	Social media source	1493	30.2	132	15.2			Friend	933	30.1	650	23.9
Friend	1583	27.2	Friend	1347	27.2	236	27.2			Social media source	853	27.5	772	28.4
News media source	753	12.9	News media source	647	13.1	106	12.2			News media source	357	11.5	396	14.6
Other or I dont remember	695	12.0	Other or I dont remembet	583	11.8	112	12.9			Other or I dont remembet	312	10.1	383	14.1
Casual sex partner	356	6.1	Casual sex partner	305	6.2	51	5.9			Casual sex partner	195	6.3	161	5.9
Medical provider	288	5.0	As part of a research stud	186	3.8	44	5.1			Medical provider	159	5.1	129	4.8
As part of a research study	230	4.0	Medical provider	170	3.4	118	13.6			As part of a research stud	141	4.6	89	3.3
Community-based agency	212	3.6	Community-based agenc	160	3.2	52	6.0			Community-based agenq	108	3.5	104	3.8
Main sex partner	59	1.0	Main sex partner	43	0.9	16	1.8			Main sex partner	31	1.0	28	1.0
Family member	16	0.3	Family member	14	0.3	2	0.2			Family member	11	0.4	5	0.2
Most influential on PrEP views
				PrEP experience¥							Intend to use PrEP¥
				PrEP Naïve		Former PrEP					Yes		Little or no
	Total			n = 4948		n = 869					n = 3100		n = 2717
	Frequency	(%)		Frequency	(%)	Frequency	(%)	Chi-square	P		Frequency	(%)	Frequency	(%)	Chi-square	P
								284.4	<.0001						113.3	<0001
Friend	1393	24.0	Friend	1233	24.9	160	18.4			Friend	791	25.5	602	22.2
Social media source	1308	22.5	Social media source	1206	24.4	102	11.7			Social media source	667	21.5	641	23.6
News media source	777	13.4	News media source	687	13.9	90	10.4			News media source	372	12.0	405	14.9
Other or I dont remember	664	11.4	Other or I dont remembet	550	11.1	114	13.1			Medical provider	313	10.1	241	8.9
Medical provider	554	9.5	As part of a research stud	380	7.7	76	8.8			As part of a research stud	302	9.7	154	5.7
As part of a research study	456	7.8	Medical provider	344	7.0	210	24.2			Other or I dont remembet	265	8.6	399	14.7
Casual sex partner	291	5.0	Casual sex partner	259	5.2	32	3.7			Casual sex partner	180	5.8	111	4.1
Community-based agency	261	4.5	Community-based agentr	195	3.9	66	7.6			Community-based agenc	141	4.6	120	4.4
Main sex partner	97	1.7	Main sex partner	81	1.6	16	1.8			Main sex partner	57	1.8	40	1.5
Family member	16	0.3	Family member	13	0.3	3	0.4			Family member	12	0.4	4	0.2

Nearly one-quarter (23.8%) of participants reported that they did not know anyone who took PrEP (Table 1). The highest proportion (43.7%) of participants knew between one and four people taking PrEP, and 32.6% knew 5 or more people taking PrEP. Bivariate analyses indicated statistically significant associations between number of network members taking PrEP and prior PrEP use (X² = 222.9, p-value < .0001) as well as participants intending to start taking PrEP (X² = 34.8, p-value < .0001). Notably, a lower proportion of former PrEP users, compared to PrEP-naïve (10.6% vs. 26.1%) knew no one on PrEP, while a higher proportion of former PrEP users knew 5 or more people taking PrEP (53.4% vs. 28.9%). A similar pattern emerged for participants who intended to use PrEP, compared to those who did not (none: 20.9% vs. 27%; 5 or more: 35% vs 29.9%).

Very few participants (0.6%) reported that people in their networks were opposed to PrEP. Most (65.1%) reported that their networks were in favor of PrEP (Table 1). Bivariate analyses indicated statistically significant associations between PrEP favorability within their networks and prior PrEP use ( X² = 84.9, p-value < .0001) as well as intent to start taking PrEP ( X² = 282.2, p-value < .0001) for PrEP favorability. A slightly higher proportion of participants who had previously taken PrEP reported that people in their networks were “in favor” of PrEP compared to PrEP-naïve (76.1% vs. 63.1%). A significantly higher proportion of participants intending to use PrEP, reported that their networks viewed PrEP favorably compared to participants who did not intend to use PrEP (74.7% vs. 54.1%).

Sociocultural influences on PrEP experience and PrEP intentions:

Results from multivariable regression analyses are presented in Table 3 and indicated that PrEP use in participants’ networks was associated with personal PrEP experiences. Participants who knew five or more people who used PrEP had three times the odds (95% CI: 2.31-3.83; ref. knows no one on PrEP) of having used PrEP before, compared to PrEP-naïve participants. Similarly, participants who knew at least one (but fewer than five) people who used PrEP had a higher odds (aOR = 1.55, 95% CI: 1.21-2.00; ref. know no one on PrEP) of being a former PrEP user, compared to PrEP-naïve participants. Participants who had networks with no strong opinions about PrEP (aOR = 0.57, 95% CI: 0.43-0.76; ref. in favor of PrEP) or networks who did not really know about PrEP (aOR = 0.66, 95% CI: 0.48-0.91) had lower odds of previously using PrEP. Demographic factors associated with increased odds of prior PrEP use were being between the ages of 25-35 and 36-45, education status other than having a high school diploma or GED, being unemployed, experiencing housing instability in the past 5 years, and engaging in sex work in the past 3 months. Demographic factors associated with lower odds of prior PrEP use were identifying as something other than gay, queer or homosexual, region of residence other than the Northeast, and not knowing one’s HIV status.

Table 3.

Associations of sociocultural infulances on PrEP experiences and intentions to use PrEP among study participants, together 5000 study, 2017-2018

	PrEP Experience					PrEP Intentions
	(Ref = PrEP-nalve)					(Ref=Little or No Intentions)
			95% CI					95% CI
Characteristic	Estimate	aOR	LL	UL	p-value	Estimate	aOR	LL	UL	p-value
Age
16-24 years old			Ref					Ref
25-35 years old	0.44	1.55	1.22	1.95	0.0003	−0.17	0.85	0.73	-- 0.98	0.02
36-45 years old	0.48	1.61	1.24	2.11	0.0004	−0.24	0.79	0.66	-- 0.94	0.01
46+ years old	0.07	1.07	0.71	1.62	0.74	−0.44	0.65	0.49	−0.85	0.001
Race/Ethnicity
White								Ref
Black or African American						0.49	1.64	1.36	−1.97	<.0001
Latino						0.28	1.32	1.15	-- 1.51	<.0001
Asian or Pacific Islander						0.76	2.13	1.57	-- 2.90	<.0001
Other or multiracial						0.05	1.05	0.86	-- 1.28	0.64
Sexual Orientation
Gay, Queer, Homosexual
Bisexual	−0.29	0.75	0.58	0.96	0.02
Other	−0.50	0.61	0.23	1.59	0.31
Region
Northeast			Ref
South	−0.54	0.59	0.48	0.72	<.0001
Midwest	−0.38	0.68	0.53	0.88	0.004
West	−0.24	0.78	0.63	0.98	0.03
US Overseas/Territories	−12.12	<0.001	<0.001	>999	0.95
Education
< High school diploma	0.65	1.92	1.05	3.50	0.03	0.10	1.11	0.74	-- 1.67	0.62
High school diploma or GED			Ref					Ref
Some college or technical school training	0.47	1.60	1.20	2.12	0.001	−0.19	0.83	0.70	-- 0.98	0.03
College graduate +	0.57	1.76	1.32	2.35	0.0001	−0.33	0.72	0.60	-- 0.86	0.0004
Employment Status
Full-time (40 hours per week)			Ref
Part-time (less than 40 hours per week)	0.21	1.23	0.98	1.55	0.08
Working or full-time student	0.10	1.10	0.85	1.43	0.45
Unemployed/Other	0.36	1.43	1.10	1.86	0.008
Annual Income
Less than $20,000						0.07	1.08	0.91	-- 1.27	0.39
$20,000-$49.999						0.17	1.19	1.03	-- 1.37	0.02
50,000+								Ref
Experience Housing Instability (past 5 years)
No/Not within last 5 years			Ref					Ref
Yes, within last five years	0.24	1.27	1.04	1.55	0.02	0.20	1.22	1.06	-- 1.40	0.01
Engaged in sex work (past 3 months)
No			Ref
Yes	0.36	1.43	1.15	1.79	0.002
Self-reported HIV Status
HIV-negative			Ref
I don’t know; I am unsure	−0.78	0.46	0.38	0.54	<0001
Experience with PrEP
PrEP naïve						−0.52	0.60	0.51	-- 0.70	<.0001
Yes, but not currently taking PrEP								Ref
No. of people in-network who take PrEP
None			Ref					Ref
1-4 people on PrEP	0.44	1.55	1.21	2.00	0.0006	0.19	1.21	1.05	-- 1.39	0.01
5+ people	1.09	2.97	2.31	3.83	<.0001	0.24	1.27	1.09	-- 1.49	0.003
Network favorability of PrEP
Are in favor of PrEP			Ref					Ref
Are opposed to PrEP	0.58	1.78	0.79	4.01	0.16	−0.80	0.45	0.23	-- 0.88	0.02
Are split evenly between being in favor of and being opposed to PrEP	−0.18	0.83	0.65	1.07	0.14	−0.97	0.38	0.32	-- 0.45	<.0001
Don’t have strong opinions either way	−0.56	0.57	0.43	0.76	0.0002	−1.00	0.37	0.31	-- 0.44	<.0001
Don’t really know what PrEP is	−0.41	0.66	0.48	0.91	0.01	−0.72	0.49	0.41	-- 0.58	<.0001

Similarly, participants who knew between one and four people on PrEP (aOR: 1.21, 95% CI: 1.05-1.39 vs. knowing no one on PrEP) and participants who knew and 5 or more people on PrEP (aOR = 1.27, 95% CI: 1.09-1.49) had significantly higher odds of reporting intentions to start PrEP. Additionally, participants who had networks that were opposed to PrEP (aOR = 0.45, 95% CI: 0.23-0.88 vs. in favor of PrEP), had split opinions about PrEP (equally in favor and opposed; aOR = 0.38, 95% CI: 0.32-0.45), did not have strong opinions (aOR = 0.37, 95% CI: 0.31-0.44), or did not know what PrEP was (aOR = 0.49, 95% CI: 0.41-0.58) had lower odds of reporting intentions to start PrEP. Demographic factors associated with increased odds of intending to start PrEP included younger age, being a racial or ethnic group other than White, reporting an annual income between $20,000-$49,999, having reported housing instability in the past 5 years, and reporting previous experience with PrEP. Finally, demographic factors associated with lower odds of intending to start PrEP were older age and higher education status.

DISCUSSION:

Our goal was to provide a description of sociocultural influences on PrEP knowledge, attitudes toward PrEP, and subsequent PrEP behaviors. We found that social connections (i.e. social media and friends) were important and influential conduits for PrEP information—both in terms of first hearing about PrEP as well as influencing current attitudes toward it. We also found that knowing more people on PrEP and having a social network with favorable opinions about PrEP were associated with both prior PrEP use (compared to being PrEP-naïve) and intentions to use PrEP in the future. These findings reaffirm the importance of considering sociocultural influences on PrEP use behaviors, in addition to access and provider influences among cisgender men in the U.S. Our findings continue to support these promotion avenues as potential opportunities for PrEP interventions.

One of the major contributions of this study is to highlight the importance of social connections and social media in the dissemination of PrEP information. Among our participants, higher proportions of participants reported first hearing about PrEP through their friends or social media rather than from a medical provider. We also found the same pattern when participants reported who or what was most influential on their current views about PrEP. This indicates that providers are not the first nor primary sources of PrEP information. Healthcare providers are gatekeepers to PrEP access and have generally been considered the first points of contact for PrEP information.³¹ However, our data suggest that social networks may also play a significant role. Efforts to educate providers about PrEP dissemination and providing culturally appropriate PrEP-related care, ensure that information PrEP candidates receive from providers is accurate.³² The same cannot be said for the information that is passed through social network or social media connections. This is particularly important when we consider the relatively low coverage of PrEP for HIV prevention, as well as stigma that has and continues to circulate about PrEP and persons who use PrEP.^26–29 Not only does stigma impact uptake, but negative social group views on PrEP can impact the adherence of those who are currently taking PrEP.²⁶ An additional layer of misinformation impeding PrEP uptake and adherence efforts as HIV prevention are the misinformation campaigns have run rampant on social media in recent months advertising against brand name PrEP and encouraging users to join product liability lawsuits.³⁸ Thus, it is critical to ensure that messages promulgated in social networks and on social media are truthful and approachable for all health literacy levels.

Our study found that knowing more people on PrEP increased the odds of our participants reporting having prior experience taking PrEP and reporting that they intended to start using PrEP in the future. This is important, as it suggests a positive dose response due to increased exposure to PrEP-using peers and is encouraging given the gradual increases in PrEP use among MSM.³⁹ However, our study also suggests that it is not just the quantity of people known to be using PrEP that is influential, but also how they view PrEP. Favorable views of PrEP among participants’ peers increased the odds of having prior experience using PrEP compared to participants who had networks who did not know of PrEP or had no strong views on PrEP. Similarly, participants had a higher odds of intending to use PrEP if they viewed PrEP favorably compared to any other view point (unfavorably, split evenly between favorable and not, no strong opinions, and not knowing about PrEP). These findings help to support the PrEP stigma literature^26–29 that the reputation of PrEP is an important factor in patients’ PrEP decision-making.

Public health campaigns may have contributed to some of the stigma surrounding PrEP in describing PrEP as a prevention method targeted to “high risk persons.”^26,31 However, more recently campaigns have considered these implications and focused on developing inclusive and sex-, sexuality-, and gender-positive health campaign messages about PrEP.⁴⁰ Reducing general stigma around sexual identity and behaviors is also important for patient-provider communications about sexual health and PrEP. Participants in other studies have voiced concerns about discussing sexual orientation, sexual behaviors, and PrEP with their healthcare providers.⁴¹ Reducing social stigma around PrEP and creating culturally sensitive and competent providers may be an overwhelming and complex task in some areas of the U.S. However, as public health practitioners develop PrEP prevention messages and interventions, these real-world experiences and social contexts should be considered. Further, knowing how potential candidates are hearing about PrEP and what is influencing their decisions on PrEP can help us, as public health practitioners, be more strategic in designing prevention programs.

Social networks and media platforms offer efficient and cost-effective opportunities to conduct sexual health research, disseminate PrEP-positive messaging, and accurate public health information.^42–44 Notably, social networks have successfully been used to promote HIV prevention and risk reduction leveraging the Diffusion of Innovations theory⁴⁵ often through tapping into peer opinion leaders.^46–49 For example, the Harnessing Online Peer Education (HOPE) study successfully leveraged peer opinion leaders and online social media technologies to increase HIV testing⁵⁰ as well as being deemed an acceptable and feasible method to reduce prescription drug abuse.⁵¹ As with increasing HIV testing, reducing HIV risk behavior, and reducing prescription drug abuse, identifying MSM community leaders who could be used to disseminate an accurate and positive PrEP messaging intervention—or building partnerships and sharing resources with key activists and community groups already working to promote PrEP as HIV prevention—would be next steps building off the current study’s findings.

Our findings should be understood in light of their limitations. First, this is an internet-based study and subject to potential fraudulent or duplicate participants; however, we took substantial measures (e.g., blocking multiple survey attempts from the same IP address) to prevent illegitimate participants and to remove those that were discovered.^33,34 Next, this study primarily collected self-reported data, including data collected on sensitive topics, like se work and sexual behavior, which may not accurately represent behavior due to social desirability bias. Additionally, due to the design of the study, we are unable to assess current PrEP use and rather focus on past and future intentions to use PrEP. However, we will be able to assess sociocultural influences on past, present, and future PrEP in upcoming study assessments as we monitor PrEP uptake in our cohort. This will also allow us to perform longitudinal assessments of sociocultural influences moving beyond the current cross-sectional analysis. Despite these limitations, our study illustrates the importance of considering social contexts in designing and disseminating PrEP educational and intervention materials.

CONCLUSION:

Much research and effort to increase PrEP uptake has focused on providers’ role in these processes, yet our findings highlight most individuals are first learning about PrEP and attribute their current views about PrEP to their social networks and social media. These results emphasize the importance of social connections as health advice sources further supported by our regression analyses. Knowing higher numbers of people taking PrEP was associated with past PrEP use or intentions to use PrEP in the future. Similarly, participants’ network member opinions were also associated past PrEP use or future intentions. These findings illustrate the importance of the behaviors and opinions of social connections within their network and the value of incorporating network theories into HIV prevention interventions.

Figure 1.

Geographic distribution of study participants who completed baseline screening and additional measures survey, Together 5000 Study, 2017-2018, n = 6,270.

Public health significance:

In addition to provider-patient communication about PrEP and PrEP eligibility, social connections continue to be impactful sources of PrEP information. Health officials should leverage social connection resources to increase PrEP awareness and utilization.