Socio-behavioral correlates of pre-exposure prophylaxis use and correct adherence in men who have sex with men in West Africa.

BACKGROUND: Multiple barriers compromise pre-exposure prophylaxis (PrEP) engagement (i.e., use and adherence) in men who have sex with men (MSM). In low/middle-income countries, little is known about PrEP engagement in this population. In West Africa, the CohMSM-PrEP study was one of the rare interventions providing PrEP to MSM. We estimated PrEP use and correct adherence rates in CohMSM-PrEP, together with associated factors over time. 
METHODS: CohMSM-PrEP recruited MSM in four community-based clinics in Mali, Côte d'Ivoire, Burkina Faso, and Togo. Quarterly follow-up included collecting socio-behavioral data, and providing a comprehensive HIV prevention package, PrEP (daily or event-driven), and peer educator (PE)-led counselling. Using repeated measures, multivariate generalized estimating equations models were used to identify factors associated with self-reported i) PrEP use and ii) correct PrEP adherence during participants' most recent anal intercourse (defined as four pills/week for daily users and 2 + 1 + 1 for event-driven users).
RESULTS: Five hundred twenty participants were included with a median follow-up time of 12 months (IQR 6-21). Of the 2839 intercourses declared over the follow-up period, PrEP use was self-reported for 1996 (70%), and correct PrEP adherence for 1461 (73%) of the latter. PrEP use was higher in participants who also attended participating clinics outside of scheduled visits (adjusted odds ratio (aOR) [95% Confidence Interval, CI], p-value; 1.32[1.01-1.71], 0.040), and in those who practiced condomless anal sex (1.86[1.54-2.24], < 0.001). Correct adherence was higher in those who often contacted PE outside of scheduled visits (2.16[1.01-4.64], 0.047) and in participants who adopted receptive/versatile sexual positions with stable partners (1.36[1.03-1.81], 0.030). Instead, after an interaction effect between financial situation and regimen was tested, it was lower in event-driven users with a difficult/very difficult financial situation (comfortable/just making ends meet & daily, 4.19[2.56-6.86], < 0.001; difficult/very difficult & daily, 6.47[4.05-10.30], < 0.001; comfortable/just making ends meet & event-driven, 1.63[1.22-2.17], 0.001), and in participants who felt alone (0.76[0.58-0.99], 0.042).
CONCLUSIONS: Community-based clinic attendance and PE contact outside of scheduled visits were both associated with higher PrEP engagement, but some socially and economically marginalized participants struggled with adherence. As scale-up continues in West Africa, we recommend implementing community-based interventions and providing extra support for vulnerable users to ensure adequate PrEP engagement.

Introduction

For men who have sex with men (MSM), multiple individual, social and structural barriers compromise pre-exposure prophylaxis (PrEP) engagement (i.e., use and correct adherence) [1-4], which can leave them exposed to HIV infection. While these barriers have been widely studied in high-income countries [5-19], very little is known about PrEP engagement in MSM in low- and middle-income countries, a setting where roll-out is comparatively limited [20-22] and where barriers to engagement can be exacerbated because of socially and culturally hostile contexts for MSM [23-25].

In West Africa, the HIV epidemic concentrated in MSM has persisted for over a decade. Prevalence in MSM is more than ten times higher (15.9%) than in the general population (1.2%) [26]. Although PrEP scale-up is urgent to contain the epidemic there, national programs in the region have been slow to implement it [27]. Structural barriers, economic constraints and biological factors disadvantage MSM and explain, in part, this concentrated epidemic and the delay in PrEP roll-out [28, 29]. Furthermore, legal barriers such as discriminatory policies and laws against homosexual behaviors reinforce a culture of widespread same-sex stigma [30-32]. At the institutional level, political stakeholders attribute less financial resources to HIV programs dedicated to LGBTQ populations [33], which limits the provision of adapted healthcare for MSM and PrEP scale-up for this population [28, 34–38]. These legal and financial barriers are compounded by heteronormative cultural values, which can lead to family rejection, social isolation and other forms of societal marginalization for MSM [30, 39]. Indeed, this complex legal and cultural context regarding MSM not only limits their access to tailored HIV prevention and care, including PrEP, but also research and clinical data on this population and their behaviors [29, 40–42], including PrEP engagement.

As in other regions, one of the biggest concerns for HIV prevention stakeholders in West Africa is poor PrEP effectiveness because of low PrEP use and adherence [43, 44]. In all income settings, individual level barriers such as the fear of side effects [5, 45–47], psychoactive substance use [5, 6], and low perceived risk of HIV infection [7-11] have all been associated with poorer PrEP-related outcomes and PrEP discontinuation. Also in all income settings, reporting higher levels of HIV risk behaviors has been associated with higher PrEP use and/or adherence [5, 6, 12–16, 48–50]. In low- and middle-income countries only, receiving partner, familial, and social support [51-55] was strongly associated with PrEP engagement, while in high-income settings having a friend [17] or partner [16] on PrEP was a facilitator. In all income settings, at the community level, medical mistrust because of experiences of homophobia, intersectional stigma [18, 23], PrEP-related stigma [24, 25, 51, 54, 56, 57], being perceived as HIV positive [53, 58], and considered at risk of HIV infection [54, 59], are all major psychosocial barriers to PrEP use and adherence by MSM. Finally, the effects of structural barriers on PrEP-related outcomes in different income settings worldwide, for example socioeconomic strain [6, 9, 16] and societal-level stigma [2, 19], have also been studied.

In places like West Africa, where there is the potential for multiple intersecting barriers to impact MSM PrEP engagement, community involvement and social network interventions could be the key to ensuring its success. To date, the impact of these types of interventions on PrEP use and/or adherence has only been quantitatively measured in studies focusing on marginalized MSM in the United States [17, 60–62]. Some of these studies involved multi-modal interventions, making it difficult to establish a clear link between these interventions and improved PrEP engagement.

One demonstration project contributing data on this topic in West Africa was CohMSM-PrEP [63]. MSM were provided with PrEP (daily and event-driven) and peer education in MSM-friendly community-based clinics. Initial findings showed that the introduction of PrEP and the use of peer-based outreach over time influenced the type of participant who enrolled in the study [64], and that PrEP uptake was associated with an 80% reduction in the risk of new HIV infections. However, it was also found that PrEP engagement decreased over time [65], and that event-driven PrEP users experiencing socioeconomic strain, as well as other socially vulnerable participants, had a higher risk of being ineffectively protected against HIV [66].

The aim of the present study was to report the rates of PrEP use and correct PrEP adherence (defined as four pills per week for daily users and 2 + 1 + 1 for event-driven users) in CohMSM-PrEP, as well as associated factors over time. We studied multiple cohort-related factors, including MSM-friendly clinic attendance and PE contact, and other sociobehavioral correlates of PrEP engagement. We also explored whether the correlates of PrEP use were similar to those of adherence. Our main goal was to better understand the cohort’s initial findings in a context where PrEP roll-out continues in national HIV/AIDS plans and programs in West Africa. To our knowledge, this is the first quantitative study to explore the correlates of PrEP engagement in MSM in Sub-Saharan Africa (SSA).

Methods

Study design

In November 2017, the CohMSM-PrEP prospective cohort study was initiated to assess the acceptability and feasibility of PrEP for MSM as part of a comprehensive sexual health prevention package in community-based clinics in West Africa (Burkina Faso, Côte d’Ivoire, Mali, and Togo). The four study sites were MSM-friendly clinics run by community-based organizations: Centre Oasis run by the Association African Solidarité (AAS) in Ouagadougou (Burkina Faso); Clinique Confiance, run by the association Espace Confiance in Abidjan (Côte d’Ivoire); Clinique de Santé Sexuelle des Halles run by the Association pour la Résilience des Communautés pour l’Accès au Développement et à la Santé – ARCAD Santé PLUS (formerly ARCAD-SIDA) in Bamako (Mali); and Centre Lucia, run by Espoir Vie Togo (EVT) in Lomé (Togo). Previously, in these same sites, the feasibility and acceptability of implementing HIV prevention and care services for MSM were studied in the CohMSM cohort [67]. Apart from PrEP, the comprehensive sexual health prevention package was the same in both cohort studies.

MSM were informed about the CohMSM-PrEP study at the clinics through a specific network of community-based organizations and peer educators (PE). CohMSM participants wishing to continue follow-up in CohMSM-PrEP (therefore wishing to take PrEP) and new ‘potential’ participants had to meet the same eligibility criteria (a comparison of the two cohorts has been previously described [64].

Participants were eligible if they were 18 years or older, HIV-negative (status confirmed at study enrollment), MSM (defined as reporting at least one episode of anal intercourse [insertive or receptive] with another man in the six months preceding enrollment), and reported any of the following HIV at-risk criteria: (i) non-virally suppressed seropositive sexual partner (male or female), (ii) condomless anal or vaginal sex with multiple partners in the previous six months, (iii) a history of sexually transmitted infection (STI) in the previous six months, (iv) post-exposure prophylaxis use in the previous six months, or (v) requesting PrEP.

Medical staff collected clinical data at each quarterly follow-up visit (scheduled or not), including PrEP regimen and quarterly HIV and STI testing results. Trained research assistants administered standardized face-to-face questionnaires at enrollment and every three months thereafter, which collected data on individual characteristics, sexual behaviors, psychosocial factors, substance use, and PrEP and condom use.

During these same quarterly follow-up visits, participants were provided free clinical examinations, PrEP (event-driven or daily), HIV testing, screening and treatment for other STI, condoms and lubricants, and tailored prevention counseling. The latter was provided by PE and focused on PrEP adherence, risk reduction strategies (including condom use, switching between regimens, abstaining from certain high-risk activities, and other risk-reduction strategies), encouraging testing, and program retention. Specific adherence-based counseling was provided monthly during the first three months and every three months thereafter. PrEP was prescribed as follows: daily (one pill per day) or event-driven (2 + 1 + 1 dosing; i.e., 2 pills between 2–24 h before sex [1 if PrEP taken the previous day] followed by 1 pill 24 h and another 48 h after the first pill[s]). At each follow-up visit, in concertation with study doctors and/or PE, participants could decide to switch PrEP strategies or stop PrEP (temporarily or permanently) depending on their needs. Participants were screened for HIV using national algorithms (Abbott Determine HIV 1/2 assay [Abbott Laboratories, Chiba, Japan] and, if the result was positive, SD Bioline HIV-1/2 3.0 [SD, Gyeonggi-do, South Korea] or First Response HIV-1/2 assay [Premier Medical Corporation, Mumbai, India]) and those diagnosed HIV positive during follow-up were invited to initiate antiretroviral treatment immediately.

Study population

The analyses for the present study included CohMSM-PrEP participants enrolled between November 2017 and November 2020 who had at least one available questionnaire from M3 to M36, since there was no pill intake at baseline. We excluded data for participants who declared no male partners in the previous three months (stable or casual). As the study focused on PrEP-based HIV prevention, participants who seroconverted during follow-up were censored at their seroconversion date.

Outcomes

The study’s two research outcomes were PrEP use and adherence during participants’ most recent anal intercourse (receptive and/or insertive) with male partners only (stable or casual), as self-reported in participants’ quarterly questionnaires.

Specifically, PrEP use was defined as declaring PrEP use alone or in combination with condoms, irrespective of adherence (= 1). No PrEP use was defined as declaring condom use only or neither PrEP nor condom use (= 0).

The adherence outcome measured whether PrEP adherence was ‘correct’ (= 1) or ‘incorrect’ (= 0). Adherence was considered correct for daily users, if they had taken at least four pills in the week before their most recent intercourse because the literature has shown that such levels confer protection against HIV infection for MSM [50, 68]. For event-driven users, it was considered correct if they had taken PrEP as prescribed (2 + 1 + 1) [69, 70]. Incorrect PrEP adherence was defined as all other pill taking combinations or taking any pills before or after sex. Only measures of most recent intercourse where PrEP use was declared were included for this outcome.

Covariates

Covariates in the present analysis included:

Sociodemographic and socioeconomic characteristics. Age (continuous), country-fixed effects (Mali, Burkina Faso, Togo and Côte d’Ivoire), employment status (employed vs unemployed), perception of financial situation (‘comfortable’ and ‘just making ends meet’ vs ‘difficult’ and ‘very difficult’).

Cohort or PrEP-related characteristics. Recruitment type (ex-CohMSM participant vs. new participant), chosen PrEP regimen (event-driven or daily), PrEP use (‘easy’ and ‘very easy’ vs ‘difficult’ and ‘very difficult’), attended clinic outside of scheduled visits (‘yes’ vs ‘no’), contacted a peer educator outside of scheduled visits (‘never,’ ‘sometimes’ [i.e., 4 or less times a month] ‘often’ [i.e., 2 or more times a week]), and follow-up time (3–6 months, 7–12 months, > 12 months).

MSM identity and psychosocial aspects. Self-defined sexual orientation (‘heterosexual’, ‘homosexual/gay/trans [i.e., transsexual or transgender],’ or ‘bisexual’), self-defined gender identity (‘man/boy’ vs ‘both a man and a woman,’ ‘more a woman,’ and ‘neither a woman nor a man’), overall perceived homosexuality stigma score (based on 11 homosexuality-related stigma questions where overall low (high) stigma was defined as < median (vs ≥ median) [71]), and feeling alone (‘no’ vs ‘sometimes’ and ‘every day’).

Sexual behaviors and substance use. Recreational drug use in the previous month (no use vs used at least one of the following: cannabis/marijuana/hash/weed, Tramadol, cocaine/crack, heroin/other opioids, medicines used for recreational purposes, Kenacort/other injectable skin lightening products, other), condomless anal sex (CAS) during most recent intercourse (‘yes’ vs ‘no’), a receptive position during most recent anal intercourse (‘yes’ vs ‘no’), transactional sex in the previous three months (‘never’ vs ‘rarely,’ ‘often’ and ‘always’ vs ‘not concerned’), number of sexual intercourses with stable and/or casual partners in the previous month (categorized into ‘none or no partner’ vs ‘1–4’ vs ‘ ≥ 5’) and sexual position with stable male partner in the previous three months (‘no stable partner’ vs ‘exclusively insertive’ vs. ‘receptive or versatile’).

Statistical analyses

The generalized estimating equation (GEE) method, using repeated measures and a binary logistic distribution function, was used to identify factors associated with PrEP use first, and then factors associated with correct PrEP adherence. For both outcomes, all univariate and multivariate models were adjusted for country-fixed effects and other confounders. All covariates with a p-value of 0.20 or less in the univariate analysis were eligible for the multivariate model. In line with the present study’s objectives, specific interactions between covariates were tested for during multivariate model specification, including PrEP regimen and perception of financial situation. The forward selection technique was used to construct the final multivariate model. The Quasi-likelihood Information Criterion (QIC) was used to verify the goodness-of-fit of the chosen model. All analyses were performed using STATA 14.0 statistical software.

Ethical considerations

The study protocol was approved by the national ethics committees of Togo (N°338/2017/MSPS/CAB/SG/DGAS/DPML/CBRS), Mali (N°2017/113/CE/FMPOS), Burkina Faso (N°2017–7-105), and Côte d’Ivoire (N°088/MSHP/CNER-kp). All participants provided written informed consent.

Results

Sample characteristics

Of the 632 CohMSM-PrEP participants enrolled between November 2017 and November 2020, 520 (82%) had at least one available questionnaire from M3 to M36 (2839 observations), and were included in the present study (Fig. 1). Excluded participants (n = 112) either did not have a male sexual partner in the previous three months or had not yet received their M3 follow-up questionnaire.

Fig. 1

CohMSM-PrEP study population flowchart with research outcomes: i) PrEP use and ii) PrEP adherencea. aDuring most recent anal intercourse with a male sexual partner; follow-up from November 2017-November 2020 (M0-M36); Median follow-up time = 12 months, IQR [6–21]. bPresent study population by PrEP regimen: 26.9% (765/2839) daily, 72.9% (2069/2839) event-driven and 0.2% (5/2839) missing. cPrEP use by PrEP regimen: 79.1%, (605/765) of daily users and 67.0% (1386/2069) of event-driven users. dCorrect adherence by PrEP regimen: 88.8% (537/605) of daily users and 66.7% (925/1386) event-driven users. PrEP: pre-exposure prophylaxis

After a median follow-up time of 12 months (IQR 6–21), of the 2839 most recent anal intercourses declared over the follow-up period, PrEP use (outcome 1) was declared for 1996 (70%) (Fig. 1). The remaining 843 intercourses (30%) were not protected by PrEP and were excluded from the PrEP adherence (outcome 2) assessment. Correct PrEP adherence accounted for 1461 (73%) of the 1996 intercourses where PrEP was used.

Of the 2839 most recent anal intercourses over the follow-up period, 26.9% (765/2839) concerned participants on daily PrEP, 72.9% (2069/2839) concerned participants on event-driven PrEP, and 0.2% (5/2839) had missing regimen data (Fig. 1). Four-fifths of daily users declared PrEP use during their most recent anal intercourse (79.1%, 605/765), compared to 67.0% (1386/2069) of event-driven users. When PrEP was used, 88.8% (537/605) of daily users declared correct adherence, compared to 66.7% (925/1386) of event-driven users.

Table 1 describes the characteristics of the entire study sample at baseline. Mean age was 25.6 years (standard deviation, SD = 5.8). Most participants (39.8%, 207/520) came from Mali, 23.2% (121/520) from Togo, 18.5% (96/520) Burkina Faso, and 18.5% (96/520) Cote d’Ivoire. Half (50%, 260/520) were employed, and 56.2% (292/520) perceived their financial situation as difficult/very difficult at baseline.

Table 1

Characteristics of study participants at baseline N = 520

Variable	n (%) or mean (SD)
Ageb (in years)	25.6 (5.8)
Country of inclusion
Burkina Faso	96 (18.5)
Cote d’Ivoire	96 (18.5)
Mali	207 (39.8)
Togo	121 (23.2)
Employment statusc
Employed	260 (50.0)
Unemployed	252 (48.5)
Perception of financial situationc
Difficult or very difficult	292 (56.2)
Comfortable or just making ends meet	220 (42.3)
Recruitment type
Ex-CohMSM participant	290 (55.8)
New participant	230 (44.2)
Chosen PrEP regimenad
Event-driven	389 (74.8)
Daily	129 (24.8)
Using PrEP isae
Difficult or very difficult	68 (13.1)
Easy or very easy	432 (83.1)
Attended clinic outside of scheduled visits
Yes	205 (39.4)
No	315 (60.6)
Contacted PE outside of scheduled visitsf
Never	355 (68.3)
Sometimes	143 (27.5)
Often	13 (2.5)
Self-defined sexual orientationc
Heterosexual	14 (2.7)
Homosexual. gay. trans	202 (38.8)
Bisexual	296 (56.9)
Self-defined gender identityc
Man or boy	307 (59.0)
Both a man and a woman; more a woman; neither a woman or a man	205 (39.4)
Perceived homosexuality stigma scoref
Low (< median)	251 (48.3)
High (≥ median)	217 (41.7)
Feeling aloneh
Yes	290 (55.8)
No	176 (33.8)
Recreational drug use in the previous monthh
Yes	53 (10.2)
No	413 (79.4)
Condomless anal sex during most recent intercoursei
Yes	162 (31.2)
No	348 (66.9)
Receptive position during most recent anal intercoursej
Yes	258 (49.6)
No	261 (50.2)
Transactional sex in the previous three months
Never	219 (42.1)
Rarely. often. always	99 (19.0)
Not concerned	202 (38.9)
Number of sexual intercourses with stable partner in previous monthk
None or no stable partner	229 (44.0)
1–4	244 (46.9)
5 +	43 (8.3)
Number of sexual intercourses with casual partner(s) in previous monthc
None or no casual partner	254 (48.8)
1–4	228 (43.8)
5 +	36 (6.9)
Sexual position with stable male partner in the previous three monthsl
No stable male partner	152 (29.2)
Exclusively inserted	146 (28.1)
Receptive or versatile	209 (40.2)

SD Standard deviation, CohMSM Cohort of MSM, PrEP Pre-exposure prophylaxis, PE Peer educator

aPrEP-related outcomes were only available from the M3 follow-up questionnaire since there was no pill intake at M0

bRange = 18 – 57, 11 (2.1%) missing values

c8 (1.5%) missing values

d2 (0.4%) missing values

e20 (3.8%) missing values

f9 (1.7%) missing values

g52 (10.0%) missing values

h54 (10.4%) missing values

i10 (1.9%) missing values

j1 (0.2%) missing values

k4 (0.8%) missing values

l13 (2.5%) missing values

Ex-CohMSM participants comprised 55.8% (290/520) of the study sample. Two-fifths (39.4%, 205/520) of the sample attended their study clinic outside of scheduled visits. In terms of contacting PE outside of scheduled visits, 68.3% (355/520) never contacted them, 27.5% (143/520) contacted them sometimes, and 2.5% (13/520) often. Most participants chose event-driven PrEP (74.8%, 389/520), and most (83.1%, 432/520) found it easy/very easy to use PrEP.

At baseline, most of the study sample self-identified as bisexual (56.9%, 296/520), followed by homosexual/gay/trans (38.8%, 202/520), and 2.7% heterosexual (14/520). Fifty-nine percent (307/520) self-identified as male, and 39.4% (205/520) as both a man and a woman, more a woman, or neither a woman nor a man. Approximately half (48.3%, 251/520)) of the participants had a low perceived homosexuality stigma score, while 41.7% (217/520) had a high score. Over half (55.8%, 290/520) felt alone at baseline and 10.2% (53/520) had used recreational drugs in the previous month.

With regard to stable male partners, 44% (229/520) of the participants had no stable male partner at baseline or no sexual intercourse with their partner in the previous month, while 46.9% (244/520) and 8.3% (43/520) had 1–4 and 5 + intercourses, respectively. With respect to casual partners, 48.8% (254/520) had no partner or no sexual intercourse in the previous month, while 43.8% (228/520) and 6.9% (36/520) had 1–4 and 5 + intercourses, respectively. In terms of intercourse with stable partners only in the previous three months, most participants had receptive or versatile intercourse (40.2%, 209/520), 28.1% (146/520) reported exclusively insertive intercourse, and 29.2% (152/520) were not concerned as they had no stable partner. Casual transactional sex in the previous three months was reported by 19% (99/520) of participants, while 42.1% (219/520) did not report it, and 38.9% (202/520) were not concerned as they did not have casual partners. During their most recent anal intercourse with either partner type, 31.2% (162/520) of the study sample had CAS, and almost half (49.6%, 258/520)) of these intercourses were receptive.

PrEP use: univariate and multivariate analyses

Table 2 shows the univariate analysis for PrEP use (outcome 1). In univariate analysis, PrEP use during the most recent intercourse was more likely in daily PrEP users (p = 0.011), participants attending their clinic outside of scheduled visits (p = 0.037), and recent recreational drug users (p = 0.045). It was also more likely in those who had CAS (p < 0.001), those declaring receptive anal intercourse (p < 0.001), and having sexual intercourse with a stable male partner in the previous month (1–4, p < 0.001; 5 + , p = 0.003). Furthermore, when an interaction effect between perceived financial situation and PrEP regimen was tested, daily users with a comfortable or just making ends meet situation were more likely to declare PrEP use (p = 0.036). In contrast, being followed up for more than 12 months (p = 0.001), finding PrEP difficult or very difficult to use (p < 0.001), and self-identifying as male (p = 0.014) were negatively associated with PrEP use. Finally, a positive trend existed between PrEP use and often contacting PE outside of scheduled visits (p = 0.094), recent transactional sex and PrEP use (p = 0.081), and daily users with a difficult or very difficult financial situation (p = 0.085), while a negative trend existed between PrEP use and having had 5 + sexual intercourses with casual male partners in the previous month (p = 0.078).

Table 2

Univariate and multivariate analyses of factors associated with PrEP use (n = 520, 2839 measures)a and with correct PrEP adherence (n = 465, 1995 measures)b

Variables	PrEP use			PrEP adherence
	n(%)c	Univariate	Multivariate	n(%)d	Univariate	Multivariate
		OR[95% CI], p-value	OR [95% CI], p-value		OR [95% CI], p-value	OR [95% CI], p-value
Age		1.01 [0.99–1.03], 0.528	1.01 [0.99–1.03], 0.448		1.02 [1.00–1.05], 0.043
Country of inclusion
Burkina Faso	344 (17.2)	0.82 [0.57–1.16], 0.261	0.89 [0.60–1.30], 0.532	246 (16.8)	0.84 [0.57–1.24], 0.386	0.71 [0.46–1.09], 0.116
Cote d’Ivoire	250 (12.5)	0.46 [0.33–0.65], < 0.001	0.42 [0.29–0.62], < 0.001	162 (11.1)	0.56 [0.37–0.84], 0.006	0.45 [0.28–0.71], 0.001
Togo	558 (28.0)	0.85 [0.62–1.17], 0.327	0.70 [0.50–0.99], 0.045	418 (28.6)	1.00 [0.70–1.42], 0.997	1.02 [0.69–1.49], 0.937
Mali	844 (42.3)	1 (ref)	1 (ref)	636 (43.5)	1 (ref)	1 (ref)
Employment status
Employed	1176 (58.9)	1.12 [0.90–1.38], 0.316		881 (60.3)	1.09 [0.84–1.40], 0.503
Unemployed	802 (40.2)	1 (ref)		570 (39.0)	1 (ref)
Perception of financial situation
Comfortable or just making ends meet	928 (46.5)	0.99 [0.80–1.21], 0.892		750 (51.3)	1.27 [0.99–1.62], 0.050
Difficult or very difficult	1050 (52.6)	1 (ref)		701 (47.9)	1 (ref)
Perception of financial situation & PrEP regimen type
Comfortable or just making ends meet & Daily	249 (12.5)	1.48 [1.03–2.13], 0.036		221 (15.1)	4.87 [3.00–7.89], < 0.001	4.19 [2.56–6.86], < 0.001
Difficult or very difficult & Daily	354 (17.7)	1.31 [0.96–1.79], 0.085		314 (21.5)	5.53 [3.54–8.63], < 0.001	6.47 [4.05–10.30], < 0.001
Comfortable or just making ends meet & Event-driven	676 (33.9)	0.98 [0.77–1.23], 0.839		480 (32.8)	1.58 [1.20–2.08], 0.001	1.63 [1.22–2.17], 0.001
Difficult or very difficult & Event-driven	694 (34.8)	1 (ref)		436 (29.8)	1 (ref)	1 (ref)
Follow-up time
> 12 months	933 (46.7)	0.71 [0.58–0.86], 0.001		735 (50.3)	1.69 [1.36–2.11], < 0.001	1.80 [1.42–2.28], < 0.001
7–12 months	479 (24.0)	0.91 [0.75–1.11], 0.345		343 (23.5)	1.36 [1.09–1.70], 0.007	1.48 [1.14–1.92], 0.003
0–6 months	584 (29.3)	1 (ref)		384 (26.3)	1 (ref)	1 (ref)
Recruitment type
Ex-CohMSM participant	1279 (64.1)	0.83 [0.64–1.07], 0.151	0.74 [0.56–0.98], 0.034	938 (64.2)	0.90 [0.68–1.21], 0.503	1.21 [0.88–1.65], 0.243
New participant	717 (35.9)	1 (ref)	1 (ref)	524 (35.8)	1 (ref)	1 (ref)
Chosen PrEP regimen
Daily	605 (30.3)	1.38 [1.08–1.77], 0.011		925 (63.3)	4.21[2.97–5.96], 0.000
Event-driven	1386 (69.4)	1 (ref)		537 (36.7)	1 (ref)
Using PrEP is
Difficult or very difficult	166 (08.3)	0.62 [0.49–0.78], < 0.001	0.64 [0.51–0.82], < 0.001	101 (6.9)	0.69 [0.50–0.95], 0.025
Easy or very easy	1829 (91.6)	1 (ref)		1360 (93.0)	1 (ref)
Attended clinic outside of scheduled visits
Yes	1058 (53.0)	1.30 [1.02–1.68], 0.037	1.32 [1.01–1.71], 0.040	667 (45.6)	0.80 [0.61–1.06], 0.123
No	938 (47.0)	1 (ref)	1 (ref)	795 (54.4)	1 (ref)
Contacted PE outside of scheduled visits
Often	61 (03.1)	1.69 [0.91–3.12], 0.094		46 (3.2)	1.55 [0.78–3.11], 0.214	2.16 [1.01–4.64], 0.047
Sometimes	618 (31.0)	1.01 [0.82–1.25], 0.901		427 (29.2)	0.79 [0.62–1.01], 0.060	0.82 [0.63–1.07], 0.140
Never	1298 (65.1)	1 (ref)		977 (66.8)	1 (ref)	1 (ref)
Self-defined sexual orientation
Heterosexual	37 (1.9)	1.24 [0.61–2.54], 0.552		24 (1.6)	0.95 [0.43–2.07], 0.891
Homosexual. gay. trans	818 (41.0)	0.97 [0.78–1.21], 0.820		608 (41.6)	1.13 [0.87–1.47], 0.364
Bisexual	1123 (56.3)	1 (ref)		819 (56.0)	1 (ref)
Self-defined gender identity
Man or boy	1186 (59.4)	0.76 [0.61–0.94], 0.014	0.74 [0.59–0.94], 0.012	870 (59.5)	0.97 [0.75–1.24], 0.789	0.89 [0.67–1.19], 0.445
Both a man and a woman; more a woman; neither a woman nor a man	792 (39.7)	1 (ref)	1 (ref)	581 (39.7)	1 (ref)	1 (ref)
Perceived homosexuality stigma score
Low (< median)	967 (48.5)	0.88 [0.71–1.09], 0.246		726 (49.7)	1.25 [0.97–1.61], 0.082
High (≥ median)	966 (48.4)	1 (ref)		688 (47.1)	1 (ref)
Feeling alone
Yes	1042 (52.2)	0.94 [0.77–1.16], 0.572		745 (51.0)	0.76 [0.60–0.98], 0.032	0.76 [0.58–0.99], 0.042
No	865 (43.4)	1 (ref)		650 (44.5)	1 (ref)	1 (ref)
Recreational drug use in the previous month
Yes	184 (9.2)	1.47 [1.01–2.14], 0.045	1.49 [1.01–2.19], 0.044	136 (9.3)	1.12 [0.73–1.72], 0.597
No	1723 (86.3)	1 (ref)	1 (ref)	1259 (86.1)	1 (ref)
Condomless anal sex during most recent intercourse
Yes	843 (42.3)	1.92 [1.60–2.29], < 0.001	1.86 [1.54–2.24], < 0.001	612 (41.9)	0.93 [0.75–1.15], 0.503
No	1153 (57.7)	1 (ref)	1 (ref)	850 (58.1)	1 (ref)
Receptive position during most recent anal intercourse
Yes	1024 (51.3)	1.48 [1.21–1.82], < 0.001		757 (51.8)	1.20 [0.94–1.52], 0.142
No	972 (48.7)	1 (ref)		705 (48.2)	1 (ref)
Transactional sex in the previous three months
Not concerned	463 (23.2)	1.19 [0.90–1.59], 0.228		350 (23.9)	1.15 [0.83–1.61], 0.396
Rarely, often, always	445 (22.3)	1.24 [0.97–1.56], 0.081		332 (22.7)	1.11 [0.85–1.47], 0.442
Never	1088 (54.5)	1 (ref)		780 (53.4)	1 (ref)
Number of sexual intercourses with stable partner in previous month
5 +	168 (8.4)	1.59 [1.17–2.17], 0.003	1.50 [1.08–2.07], 0.014	113 (7.7)	0.74 [0.52–1.04], 0.086
1–4	1158 (58.0)	1.51 [1.29–1.78], < 0.001	1.45 [1.23–1.72], < 0.001	870 (59.5)	1.07 [0.87–1.31], 0.540
None or no stable partner	662 (33.2)	1 (ref)	1 (ref)	476 (32.6)	1 (ref)
Number of sexual intercourses with casual partner(s) in previous month
5 +	104 (5.2)	0.73 [0.51–1.04], 0.078		66 (4.5)	0.65 [0.43–0.99], 0.046	0.56 [0.34–0.92], 0.022
1–4	730 (36.6)	0.87 [0.73–1.03], 0.109		527 (36.1)	0.99 [0.80–1.23], 0.941	0.98 [0.76–1.28], 0.904
None or no casual partner	1159 (58.1)	1 (ref)		866 (59.2)	1 (ref)
Sexual position with stable partner in the previous three months
No stable partner	371 (18.6)	0.88 [0.70–1.10], 0.267		268 (18.3)	1.13 [0.85–1.49], 0.411	1.09 [0.78–1.51], 0.620
Receptive or versatile	894 (44.8)	1.10 [0.90–1.36], 0.350		675 (46.2)	1.30 [1.02–1.67], 0.033	1.36 [1.03–1.81], 0.030
Exclusively insertive	707 (35.4)	1 (ref)		503 (34.4)	1 (ref)	1 (ref)

GEE Generalized estimating equation, OR Odds ratio, aOR Adjusted odds ratio, CI Confidence interval, ref reference, CohMSM Cohort of MSM, PE Peer educator

aGEE, binary logistic distribution; adjusted for country-fixed effects, age and recruitment type, QIC = 3138

bGEE, binary logistic distribution; adjusted for country-fixed effects, recruitment type and self-defined gender identity, QIC = 2014

cPrEP use accounted for 1996 measures; no PrEP use accounted for 843 measures; the difference between the total sample and the total number of measures for each variable corresponds to missing values

dCorrect PrEP adherence accounted for 1462 measures; incorrect PrEP adherence accounted for 534 measures; the difference between the total sample and the total number of measures for each variable corresponds to missing values

The results from the multivariate analysis of factors associated with PrEP use are also shown in Table 2. Age and recruitment type were confounders in the multivariate model specification and were therefore adjusted for in the final multivariate model in addition to country-fixed effects. PrEP use was more likely in the following participants: (i) those who attended the clinic outside of scheduled visits (Adjusted Odds Ratio, aOR [95% Confidence Interval, CI], p-value; 1.32 [1.01–1.71], 0.040), (ii) recent recreational drug users (1.49 [1.01–2.19], 0.044), (iii) participants who had CAS during their most recent sexual intercourse (1.86 [1.54–2.24], < 0.001), (iv) participants who had 1–4 intercourses with their stable partner in the previous month (1.45 [1.23–1.72], < 0.001) and (v) those who had 5 + intercourses with their stable partner in the previous month (1.50 [1.08–2.07], 0.014). Instead, PrEP use was less likely in participants who found it difficult or very difficult to use PrEP (0.64 [0.51–0.82], < 0.001) and those who self-identified as male (0.74 [0.59–0.94], 0.012).

PrEP adherence: univariate and multivariate analyses

Table 2 also shows the univariate analysis for PrEP adherence (outcome 2). As participants’ age (p = 0.043) increased, so did the likelihood of having correct PrEP adherence. Correct adherence was also more likely in participants with 7–12 (p = 0.007) and > 12 months of follow-up (p < 0.001), in daily PrEP users (p < 0.001) as well as in those declaring a receptive or versatile sexual position with their stable male partner in the previous three months (p = 0.033). Instead, correct PrEP adherence was less likely in participants who found it difficult or very difficult to use PrEP (p = 0.025), those who felt alone (p = 0.032), and those who had 5 + sexual intercourses with a casual partner in the previous month (p = 0.046). Furthermore, when an interaction effect between perceived financial situation and PrEP regimen was tested, event-driven users with a difficult or very difficult financial situation were also less likely to have correct adherence (daily PrEP users with a comfortable financial situation or just making ends meet, p < 0.001; daily PrEP users with a difficult or very difficult financial situation, p < 0.001; and event-driven users with a comfortable financial situation or just making ends meet, p = 0.001). A negative trend was observed between correct PrEP adherence and having 5 + sexual intercourses with stable male partners (p = 0.086), as well as between correct adherence and a comfortable or just making ends meet financial situation (p = 0.050), while a positive trend was seen between correct adherence and high levels of PE contact (p = 0.060) and a low perceived homosexuality stigma score (p = 0.082).

Multivariate results are also shown in Table 2. Self-defined gender identity and recruitment type were confounders during the multivariate model specification, and were therefore adjusted for in the final multivariate model in addition to country-fixed effects. Correct PrEP adherence was more likely in the following participants: (i) participants who were followed for 7–12 months (1.48 [1.14–1.92], 0.003), (ii) those followed for more than 12 months (1.80 [1.42–2.28], < 0.001), (iii) participants who often contacted PE outside of scheduled visits (2.16 [1.01–4.64], 0.047), and (iv) those declaring a receptive or versatile sexual position with stable partners (1.36 [1.03–1.81], 0.030). Instead, correct PrEP adherence was less likely in participants who felt alone (0.76 [0.58–0.99], 0.042), those who declared 5 + intercourses with casual partners (0.56 [0.34–0.92], 0.022), and in event-driven users with a difficult or very difficult financial situation (daily PrEP users with a comfortable/just making ends meet financial situation, 4.19 [2.56–6.86], < 0.001; daily users with a difficult/very difficult situation, 6.47 [4.05–10.30], < 0.001; and event-driven users with a comfortable/just making ends meet situation, 1.63 [1.22–2.17], 0.001).

Discussion

In the present study on MSM participants enrolled from 2017 to 2020 in the CohMSM-PrEP study in West Africa, of the 2839 most recent anal intercourses declared, PrEP use was self-reported for 1996 (70%), and correct PrEP adherence for 1461 of the latter (73%). These two PrEP engagement outcomes were mainly linked to cohort-related characteristics and participants’ sexual behaviors. In contrast, only PrEP adherence was associated with participants’ social and economic vulnerabilities. Understanding the factors separately associated with PrEP use and adherence will help guide PrEP rollout as it becomes more widely available, as it is important to understand not only who uses PrEP, but also who uses it correctly.

One of the most important results from the present study was the effect of certain cohort-related characteristics on both outcomes. Specifically, we found that attending participating MSM-friendly clinics outside of scheduled visits and contacting PE outside of scheduled visits were associated with PrEP use and correct adherence, respectively. In terms of other studies, a lack of MSM-friendly clinics is a known barrier to accessing HIV prevention services [72-74]. One important aspect of MSM-friendly clinics is community engagement, for example the use of PE. Studies conducted prior to the PrEP era, showed that community engagement and social network interventions improved MSM access to HIV services, and HIV prevention, care and risk outcomes [1, 75–77]. However, few data exist on the effect of these community-based aspects on PrEP engagement in MSM, especially in West Africa and SSA as a whole. Involvement in the MSM community has previously been shown to play a crucial role in effective HIV protection, through correct PrEP adherence and/or condom use in the CohMSM-PrEP cohort [66], and through protective PrEP concentrations in the iPrEx OLE trial [78]. The association we found between attending clinics outside of scheduled visits and PrEP use suggests that the provision of comprehensive prevention services in MSM-friendly sexual health clinics promotes PrEP use. The second finding-the association between often contacting PE outside of scheduled visits and correct adherence-shows that peer education does indeed facilitate correct adherence. At the structural and community-level, we recommend that future PrEP programs ensure MSM have access to friendly clinics where they can go to as often as possible, and where they will not feel ashamed. Furthermore, to aid PrEP users at the individual level, we recommend strengthening the role of PE in PrEP programs especially concerning adherence support.

Both PrEP engagement outcomes also had behavioral type correlates. Specifically, we found PrEP use was linked to CAS and more frequent intercourse with stable partners, while correct adherence was associated with taking a receptive or versatile sexual position with stable partners. Both PrEP use and correct adherence have been associated with engaging in high-risk behaviors [12–15, 48–50]. The prevention-effective adherence paradigm [79] was developed to describe this phenomenon where PrEP users judge their level of risk and adapt their PrEP use and adherence accordingly, in order to ensure HIV protection. It is widely recognized in MSM that condomless receptive anal intercourse has the highest risk of HIV exposure [80], so we can assume that PrEP users would be especially mindful of taking it and adhering correctly to it when engaging in this behavior, but less so for insertive intercourse. This hypothesis was reflected in our results where participants who reported versatile or receptive sexual positions were more likely to report correct adherence. In terms of the differences found between partner types, intercourse with stable partners might be more regular and/or easier to schedule than causal intercourse, so it would be simpler to incorporate PrEP into users’ routines. We recommend incorporating behavioral interventions into PrEP programs to help users develop their ability to accurately evaluate their own HIV risk and adapt their PrEP use to it.

Although the two PrEP engagement outcomes studies here shared certain cohort-related and behavioral characteristics, an important difference existed: the effect of social and economic vulnerabilities was only observed for PrEP adherence. A lack of social and familial support has been shown to impede PrEP adherence for MSM [51-55], especially in low- and middle-income countries [81]. We found that socially isolated participants struggled with adherence as a whole, while those under socioeconomic strain struggled to achieve correct event-driven adherence. It is widely recognized in the literature that event-driven PrEP is harder to adhere to than daily PrEP [45, 65, 69, 70, 82–86]. This finding reinforces prior research by our team on participants in CohMSM-PrEP which found that the compounding nature of socioeconomic strain (i.e., lower socioeconomic status) with event-driven PrEP contributed to ineffective HIV prevention [66]. Structural socioeconomic barriers including unstable housing [6], low income [16] unemployment [9], and transport problems [87] are all associated with low PrEP adherence and/or discontinuation. We assert that for our study sample, these structural factors directly influenced individual health literacy and in turn participants’ ability to understand and adapt to the event-driven PrEP regimen. Although CohMSM-PrEP participants received a comprehensive sexual health package free of charge, as well as transport reimbursement to their study clinic (US$5), these measures seem to have been insufficient in the light of the extreme structural (social, legal, and cultural) barriers for MSM in West Africa. This highlights the need for PrEP programs to directly address socioeconomic barriers to PrEP engagement and to ensure users have a sufficient social support network in place.

It is important to note that the rates of PrEP use and adherence found in the present study (70% and 73%, respectively) were similar to those found in the IPERGAY trial, which tested the efficacy of event-driven PrEP in France and Canada [69, 70]. They reported 71.6% of participants used PrEP during their most recent intercourse and 68.5% of the latter declared correct adherence [88]. However, PrEP use and adherence in the present study were considerably lower than those reported in the PREVENIR cohort, which studied daily and event-driven PrEP roll-out in “real-world” conditions in the Ile-de-France region [89]. Indeed, 85.1% of participants reported PrEP use and 97.5% of them used it correctly [89]. One reason our findings reflect those of IPERGAY more than PREVENIR has to do with the state of PrEP roll-up, which is similar in West Africa now as it was when IPERGAY began in 2015 [69, 70, 90]. Indeed, the PREVENIR cohort exists in a context where PrEP has been widely adopted and there are many resources for PrEP uptake and adherence [89, 90]. Other studies among MSM in a more similar socioeconomic context, i.e. SSA, were found and reported varying degrees of PrEP engagement [48, 86, 91, 92]. For example, among Zimbabwean MSM aware of PrEP, 76.4% of them had used it recently [91]. In terms of correct adherence rates in Kenyan MSM, they varied from 0–14.5% when measured with protective drug concentrations [48, 92] and from 68–83% when measured with a medication event monitoring system [86]. However, these figures are not easily comparable with our study because measurements of PrEP use and adherence were wildly different.

This present study has limitations. First, social desirability bias may have led to over reporting of PrEP outcomes and underreporting of other sensitive topics. To minimize this bias, research assistants were trained to administer all questionnaires during the follow-up period, the assumption being that this repeated and regular contact with participants would help build a trustful relationship over time. Second, the study population comprised a convenience sample of MSM already attending community-based clinics or recruited via peer networks; accordingly, it was not necessarily representative of the local MSM population. However, the differences found in a previous analysis comparing ex-CohMSM participants with newly enrolled participants, suggest that the addition of PrEP to the original project’s (i.e., CohMSM) sexual health prevention offer, and the continued use of peer-based outreach over time helped reach a new profile of MSM less connected to these clinics [64]. Third, the fact that we used only self-reported data to assess PrEP adherence may be considered a less reliable strategy than objective biological measures (plasma, urine, hair, etc.). Nevertheless, multiple studies have shown that drug intake can be accurately predicted by self-reported PrEP outcomes [93-95]. This is especially relevant in low-resource settings, where implementing biological measures is logistically and/or financially complicated [96]. Moreover, in an attempt to preserve the accuracy and objectivity of these self-reported PrEP measures, the CohMSM-PrEP questionnaire was modeled after other cohorts [83, 88, 97] and consequently, our results were easily comparable with other studies and reaffirmed these measures’ utility when studying PrEP adherence [65, 70].

In terms of the study’s strengths, our results contribute to the limited literature on MSM engagement in PrEP in SSA, especially in West Africa where PrEP was only very recently added to a select number of national HIV/AIDS programs. Furthermore, the data from the present study reinforce the relatively little evidence on the positive effect of community-based approaches on PrEP engagement, and advocate the incorporation of PE and the use of MSM-friendly clinics in PrEP implementation in low- and middle-income settings, including West Africa, where MSM are highly stigmatized.

Conclusions

In a cohort of MSM on PrEP in West Africa, community-based clinic attendance and contact with PE outside of scheduled visits were associated with higher PrEP engagement. However, some socially and economically marginalized participants struggled with adherence. As scale-up continues in West Africa, we recommend implementing community-based interventions and providing extra support for vulnerable PrEP users to ensure adequate PrEP engagement.