Men who have sex with men (MSM) and transgender women (TGW) are considered key populations in the HIV epidemic in many countries [1, 2]. They have a 25–34 times higher risk of HIV infection than men and women in the general population [3]. HIV pre-exposure prophylaxis (PrEP) is an effective prevention tool that should be offered to high-risk individuals [4].

PrEP effectiveness was established ten years ago, and its use is predicted to substantially decrease the number of new HIV infections [5]. The daily oral use of the pills combined with tenofovir and emtricitabine is highly effective in preventing sexual exposure to HIV infection and injection drug use (at least 74%) [6,7,8].

Although MSM and TGW are considered target populations for PrEP use, significant challenges exist in accessing and retaining the use of prophylaxis by these groups [9]. These include difficulties in obtaining funding for healthcare [10], lack of adequate guidance on the prevention of sexually transmitted infections (STIs) [11], low self-perception of HIV risk [12], and family issues such as lack of communication on sex and sexuality as well as lack of family support [13, 14]. Furthermore, stigma and discrimination associated with HIV and AIDS and the use of prophylactics are barriers to PrEP initiation among MSM and TGW [15, 16].

The high PrEP discontinuation rates in MSM and TGW represent a significant challenge in controlling the HIV epidemic, negatively influencing PrEP coverage [17]. A cohort study conducted in Brazil with adolescent MSM and TGW (aMSM and aTGW, respectively) aged 15–19 years indicated a 51.8% probability of discontinuation in the first year of PrEP use, with an increased risk of discontinuation in aTGW compared to aMSM [18]. Individual, structural, and logistic factors have been linked to the discontinuation of PrEP [19], which magnifies the challenge by requiring different approaches and involvement from other sectors. Similar barriers to PrEP access hinder PrEP continuation, which include low perception of risk for HIV [18,19,20], cost [17, 19], and difficulty navigating intricate medical systems [19]. Thus, identifying retention strategies (RS) to improve PrEP persistence among MSM and TGW is essential for controlling HIV epidemics.

Demand creation strategies (DCS) and RS for PrEP use should be developed to improve PrEP access and coverage among sexual minority adolescents. This study aimed to identify, synthesize, and determine the overall effect size through meta-analysis while critically evaluating the most effective DCS and RS for PrEP among MSM and TGW.


This systematic review and meta-analysis followed the Cochrane Guidelines for Systematic Reviews of Interventions, and was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [21, 22]. The study protocol was registered in PROSPERO (number: CRD42022323220).

Eligibility criteria

We included cross-sectional studies, cohort studies, and randomized control trials [RCTs] that followed MSM, regardless of sexual orientation (e.g., homosexual and bisexual), and TGW aged ≥ 18 years and assessed the PrEP DCS and/or PrEP RS.

Studies that did not describe PrEP DCS and/or PrEP RS, prior public protocols, retrospective studies, real-world settings, qualitative studies, reviews, case series, editorials (letters or commentaries), and those focused on assessing participants' intentions rather than actual PreP use, were excluded. Studies that focused on PrEP adoption intentions or interest in future PrEP use, participant’s awareness, knowledge and willingness of PrEP use, HIV risk perception, cases where participants were already using PrEP at the beginning of the research or had reported previous use of the medication (past year PrEP use), and studies specifically related to injectable PrEP use, were excluded. Furthermore, studies that assessed adherence as the only outcome, health economic evaluation (effectiveness and cost or modeling framework), ecologic studies, transgender male and female sex workers, or men who did not have sex with men or outside the theme were excluded.

Search strategy

To answer the question “What are the best DCS and RS for MSM and TGW on PrEP?,” We searched five independent databases, namely PubMed/Medline, Embase, Web of Science, Central ( Cochrane Library), and Latin American and Caribbean Health Science Information (LILACS) for relevant literature. Additionally, we manually searched the reference lists of the included studies.

There were no language, date, document type, publication status, or geographic restrictions in the records. The last search was conducted in April 2022 and updated on November 2022. Descriptors were identified using the Medical Subject Headings (MeSH), Descritores em Ciências da Saúde (DeCS), and Embase Subject Headings (Emtree). Subsequently, they were combined with the Boolean operator “AND”, whereas their synonyms were combined with “OR”. The following meshes formed the herein-used search strategy, which was adapted based on descriptors in each database: “Pre-Exposure Prophylaxis”; “Homosexuality, Male”; “Transgender Persons”. The search strategy adopted in each database is presented in Appendix 1.

Study selection and data extraction

Electronic search results from the defined databases were uploaded to the Rayyan Qatar Computing Research Institute [23].

The study selection and data extraction were independently performed by three investigators (NSG, GMBM, and ICNR). Any discrepancies were resolved by consensus. We adopted the following steps in the study selection: initial screening of article based on title and abstract, and thorough examination of the full-text of the selected articles. Articles that did not meet the eligibility criteria were excluded.

Information extracted from the selected studies was encoded in Excel 2019® electronic form comprising the following fields: reference, title, source, journal, impact factor, location of the study conducted, study design, follow-up period, monitoring, number of centers or health services evaluated, setting, participants’ age, population, PrEP DCS type, PrEP RS, enrolled and numbers, PrEP retention barriers, and PrEP retention facilitators.

Quality assessment

Three investigators (NSG, GMBM, and ICNR) independently assessed the risk of bias in the selected studies according to the Joanna Briggs Institute (JBI) for determining the risk of bias. The checklists included analytical cross-sectional studies, cohort studies, and RCTs ( Disagreements were resolved through discussions among the three evaluators.

The overall certainty of the body of evidence was rated using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach, while considering the overall risk of bias, consistency of effect, imprecision, indirectness, and publication bias to assess the certainty of the body of evidence [24, 25]. In the event of serious concerns in any of these domains, we rated down the quality of the evidence.


The primary outcomes were the PrEP DCS and PrEP RS. Secondary outcomes were the facilitators and barriers to the retention of this population group in PrEP services.

PrEP DCS were strategies used to increase demand by delivering positive messages about the benefits of PrEP as a component of the HIV combination, and the DCS % was calculated using the following formula: people enrolled in the study/people reached by DCS. The PrEP RS was used to keep users on the PrEP services during the study period. The RS proportion was calculated as the percentage of individuals who remained in the study/the total number study / total of individuals. For the RS proportion, only studies with follow-up assessments were considered.

The PrEP DCS were classified into three groups according to Magno et al. [26]: (1) face-to-face (i.e., peer-educator recruitment at social venues, nongovernmental organizations, and parties,direct referrals by health services; friends and/or sexual partners); (2) online (i.e., chatbot or peer-educator recruitment on social media [e.g.,, Instagram or Facebook] or dating/hook-up apps [e.g., Grindr, Tinder, Badoo, and Scruff]), and mixed when both strategies were employed.

PrEP RS was used to retain users on PrEP services during the study period, which were classified as: (1) provider counseling (i.e., face-to-face by a health professional; prevention of HIV risk counseling, distribution of condoms, lubricants, and testing for HIV or other STIs; (2) online counseling (i.e., text messages, chatbot, telephone calls, social media, and peer educator); (3) cash transfer; and (4) mixed when both strategies were employed.

Statistical analysis

We conducted a meta-analysis of the prevalence estimates that were transformed using the raw proportion (PRAW) method. The final pooled results and 95% confidence intervals (CIs) were back-transformed for ease of interpretation [27], and when the estimate for a study tended toward either 0% or 100%, the variance for that study moved toward zero. Consequently, its weight was overestimated in the present meta-analysis.

Subgroup analyses for demand creation strategies were performed considering the three strategy types (online, face-to-face, and mixed) to determine whether a strategy type could clarify our results and explain the heterogeneity. For retention, subgroup analyses were performed considering professional counseling (in-person), online counseling, and cash transfer or mixed.

A meta-regression analysis was conducted to explore the potential sources of heterogeneity for each outcome, including the study design (trial, cohort, cross-sectional), sample size (≤400, > 400), study place (Asia, Western), setting (HIV prevention and care, population), monitoring (monthly, 2–3 months, 6 months), and risk of bias (low, moderate, and high).

Forest plots were used to visually assess the pooled estimates and the corresponding 95% CIs. We calculated the Q (significance level of p<0.1) and I2 statistics, and a random-effects model was applied to assess heterogeneity.

P-values <0.05 were considered statistically significant in all analyses. Publication bias analysis was not performed if this measure was inappropriate for prevalence meta-analysis [28]. Analyses were performed in the R software, version 4.2.1 (R: A Language and Environment for Statistical Computing, Vienna, Austria), using the ‘Meta’ packages, versions 6.0-0.


Search results

Our search retrieved 1,129 studies from the four selected databases. After excluding 213 duplicate articles, 916 titles and abstracts were screened. Full-text articles of the remaining 169 records were retrieved, of which 138 were excluded (Appendices 2 and 3). Additionally, through a manual search, seven studies were selected [29,30,31,32,33,34,35] and nine studies were updated [26, 36,37,38,39,40,41,42,43]. Therefore, 46 studies conducted between 2013 and 2022 were eligible for inclusion in this systematic review [26, 29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62, 62,63,64,65,66,67,68,69,70].

Studies and users characteristics

The main characteristics of the included studies are summarized in Table 1.

Table 1 Characteristics of individual studies included on systematic review, 2023

Of the 46 included studies, 24 were cohort studies [29, 32, 36,37,38,39,40,41,42, 44, 48,49,50, 52, 53, 58,59,60,61,62,63,64,65,66], 16 were conducted through randomized and non-randomized clinical trials [30, 31, 33,34,35, 43, 45,46,47, 51, 62, 67,68,69,70] and six had cross-sectional design [26, 50, 54,55,56,57].

Nineteen studies were conducted in North America [31, 33, 34, 38, 45,46,47, 49, 51,52,53, 55, 56, 60, 61, 66, 69, 70]; seven in South America [26, 37, 44, 50, 54, 57], nine in Asia [29, 35, 40,41,42, 48, 62, 63], four in Africa [32, 58, 64, 65], and seven in Europe [30, 36, 39, 43, 59, 67, 68].

The number of health services offering PrEP included in each study ranged from 1 to 21. Six studies did not report their funding sources [40, 42, 47, 48, 62, 63].

The maximum follow-up time registered in the cohort studies and trials were as follows : 90 days [42], 120 days [41], 168 days [45], 180 days [32, 39, 40, 48, 51, 62, 63], 252 days [33], 270 days [30], 283 days [34], 336 days [37, 44, 46, 49, 50, 52], 365 days [29, 38, 47, 64,65,66], 385 days [69, 70], 620 days [59], 672 days [43], 730 days [31, 35, 53, 60, 61] and 1,095 days [36].

Twenty-four studies focused their assessments on MSM [30, 31, 37, 39,40,41,42, 45,46,47, 51,52,53, 55,56,57,58, 60, 61, 64,65,66,67,68,69,70] and nineteen focused on both population subgroups [26, 29, 32,33,34,35,36, 43, 44, 48,49,50, 50, 54, 59, 62, 63] (Table 1).

In total, 36,792 individuals were included in this review. Tables 2 and 3 describe the classifications of the studies based on PrEP DCS and RS outcomes. The most common DCS was face-to-face (n=16) (i.e., through peer educator recruitment at social venues, nongovernmental organizations, and parties; direct referrals by health services; friends and/or sexual partners), followed by online strategies (n=4), such as Chabot or peer-educator recruitment on social media [e.g., Instagram or Facebook] or dating/hook-up apps [e.g., Grindr, Tinder, Badoo, and Scruff]). The DCS that recruited the fewest participants was mixed (face-to-face and online) (n=10).

Table 2 Characteristics of recruitment and retention strategies in PrEP services
Table 3 Characteristics of retention strategies at PrEP services

The retention of health services providing PrEP was observed in 28 studies. The retention strategies were online counseling (text messages, chatbot, telephone calls, social media, and peer-educator) [39, 44, 69], mixed strategy [29, 31,32,33, 41, 45, 46, 58, 62, 64], provider counseling (face-to-face by a health professional; prevention to HIV-risk counseling, distribution of condoms, lubricants, and testing for HIV or other STIs) [30, 36,37,38, 40, 43, 49, 51,52,53, 66, 68], peer educators [63],and cash transfer [35]. The last two alone.

Quality assessment

Of the 46 studies included in this systematic review, 29 were evaluated and included in the meta-analysis. Regarding DCS outcomes, six observational studies and four trials presented a high risk of bias, seven observational studies and six trials were identified as having a moderate risk of bias, and four observational studies and two trials were identified as having a low risk of bias. Regarding retention outcome, one observational study and three trials presented a high risk of bias, 11 observational studies and six trials were identified as having a moderate risk of bias, and four observational studies and two trials were identified as having a low risk of bias. The individual studies’ risks of bias for each study are presented in Appendices 47.

Meta-analysis results

Prevalence of users enrolled in PrEP recruited by DCS

The overall prevalence of users enrolled in PrEP recruited by demand creation strategy among the entire sample evaluated was 53% (95% CI: 0.33–0.73) with a high level of heterogeneity (I 2=100%) (Fig. 1).

Fig. 1
figure 1

Forest plot of pooled proportions of MSM and TGW enrolled in PrEP DCS (n=09)

The analysis of subgroups by types of PrEP DCS for the overall population revealed that face-to-face, online, and mixed recruited 53% (95% CI: 0.33–0.74; I2=100%); 51% (95% CI: 0.00–1.00; I2=100%); 50% (95% CI: 0.21–0.79; I2=100%), of the population respectively (Fig. 2).

Fig. 2
figure 2

Forest plot of pooled proportions of PrEP DCS among the entire sample (MSM and TGW) (n=10 report and 09 studies because of difference strategies)

Of the 36 included studies, 19 assessed the percentage of MSM users enrolled in PrEP recruited. The combined proportion of MSM was 64% (95% CI: 0.54–0.74; I2=100%) (Fig. 3).

Fig. 3
figure 3

Forest plot of pooled proportions of PrEP DCS among MSM (n=19)

Among the studies which discriminated against the number of MSM users enrolled in PrEP recruited by demand creation strategy, four, thirteen, and two studies evaluated mixed, face-to-face, and online DCS, respectively. The subgroup analysis by DCS type showed that 91% of MSM (95% CI: 0.85–0.97; I2=53%) were recruited through online, 74% (95% CI: 0.56–0.91; I2=99%) through mixed, and 57% through face-to-face (95% CI: 0.46–0.68; I2=99%) strategies (Fig. 4).

Fig. 4
figure 4

Forest plot of pooled proportions of PrEP DCS among MSM (n=19)

Regarding TGW, four studies presented information on users enrolled in PrEP recruited. The pooled proportions of DCS for PrEP use among TGW was 83% (95% CI: 0.71–0.95; I 2=100%) (Fig. 5).

Fig. 5
figure 5

Forest plot of pooled proportions of PrEP DCS among TGW (n=04)

In the subgroup analysis by DCS, we observed that 85% of the TGW were recruited via mixed(95% CI: 0.60–1.00; I2=91%) and 79% via face-to-face (95% CI: 0.73–0.85) strategies (Fig. 6).

Fig. 6
figure 6

Forest plot of pooled proportions of PrEP DCS among TGW (n=04)

PrEP retention strategies

Ten studies assessed the prevalence of PrEP service retention in the overall sample. The global estimate of prevalence was 68% (95% CI: 0.51–0.85) with a high level of heterogeneity (I2=100%) (Fig. 7).

Fig. 7
figure 7

Forest plot of pooled proportions of retention to pre-exposure prophylaxis (PrEP) service among the entire sample (MSM and TGW) (n=10)

The subgroup analysis revealed a retention proportion of 57% [95% CI: 0.38–0.75] for mixed and 83% [95% CI: 0.52–1.00] for professional counseling (Fig. 8).

Fig. 8
figure 8

Forest plot of pooled proportions of retention to pre-exposure prophylaxis (PrEP) service among the entire sample (MSM and TGW) by retention strategies (n=10)

The pooled proportions of retention to PrEP service among MSM was 73% (95% CI: 0.62–0.83; I2=100%) (Fig. 9).

Fig. 9
figure 9

Forest plot of pooled proportions of retention to pre-exposure prophylaxis (PrEP) service among MSM (n=19)

Of these, 83% (95% CI: 0.80–0.85; I2=17%) were retained in PrEP provision services by online counseling; 68% (95% CI: 0.54–0.81; I 2=98%) by mixed and 74% 95% CI: 0.52–096; I2=100%) by professional counseling strategies (Fig. 10).

Fig. 10
figure 10

Forest plot of pooled proportions of retention to pre-exposure prophylaxis (PrEP) service among MSM by demand creation strategies (n=19)

Eighteen studies presented the retention data for TGW. The prevalence of retention to the PrEP service by TGW was 65% (95% CI: 0.47–0.83; I2=98%) (Fig. 11).

Fig. 11
figure 11

Forest plot of pooled proportions of retention to pre-exposure prophylaxis (PrEP) service among TGW (n=08)

In the subgroup analysis, we observed that 84% of the TGW were retained in PrEP provision services through online (95% CI: 0.64–0.95); 68% (95% CI: 0.41–0.96; I2=51.8%) through professional counseling, and 54% (95% CI: 0.23–0.84) through mixed strategies (Fig. 12).

Fig. 12
figure 12

Forest plot of pooled proportions of retention to pre-exposure prophylaxis (PrEP) service among TGW by DCS (n=08)


In the meta-regression analysis, the studies were grouped according to study design (trial, cohort, cross-sectional), sample size (≤400; >400), study place (Asia; Western), setting (HIV prevention and care; population), monitoring (monthly; 2–3 months; 6 months), and risk of bias (low, moderate, and high) (Appendix 8). The covariates for the outcome of PrEP DCS did not differ significantly. However, there was a significant difference in the study design; the longer the study duration (cohort versus cross-sectional or RCT), the lower the proportions observed.

Barriers and facilitators to PrEP retention

Of the 46 studies included, two described the barriers [37, 38], and three presented the facilitators [37, 62] (Grinztejn et al. 2018) to retaining this population in health services offering PrEP. Among the studies reporting barriers to retention in PrEP services, two focused on TGW [37, 38]. The barriers highlighted in these studies were social determinants of health [37],reporting condom-less anal sex (CAS) with partner(s) of unknown HIV status [38] and being an immigrant [38]. Studies reporting the facilitators for retaining this population in PrEP services include one study focused on TGW [37], and two on the total population (MSM and TGW) [44, 62], which highlighted facilitators, including PrEP offered at public health-care clinics in a middle-income setting [44], approach to counseling [62], multidisciplinary care [62], and gender-affirming settings [37].


Herein, we conducted a comprehensive search to identify DCS and RS with higher proportions among MSM and TGW to improve PrEP persistence, which is crucial for reducing the HIV epidemic. As a main result, online counseling had the highest proportions for DCS and RS. Meanwhile, mixed DCS and RS were the most frequent for TGW.

The COVID-19 pandemic has impacted the way interviews are conducted in the health area, causing a significant increase in the use of online approaches [73, 74]. Online research methodologies may serve as an important mechanism for population-focused data collection among young individual and have been acknowledged for their potential in investigating understudied and marginalized populations and subpopulations, permitting increased access to communities that tend to be less visible and, consequently, less studied in offline contexts [75, 76].

Online interviews present several advantages over face-to-face interviews, particularly when engaging with hard-to-reach populations such as MSM and TGW. These advantages include anonymity, instant access to services, peer-to-peer models of online outreach, and reduced barriers such as geography and time. However, online approaches require reading, technological literacy of participants, and access to technology, which may limit relationship-building between participants and researchers [77].

The COVID-19 pandemic has reinforced the need to diversify the strategies for recruiting and retaining in PrEP services [26] (Dourado et al. 2020). The pandemic context, which negatively impacts access of MSM and TGW to HIV testing and prevention services in multiple countries [78, 79], demonstrates the need to readapt strategies, aiming for more online resources due to the facilitation of communication between users and services through the use of various platforms, such as social networks, dating applications, and chatbots [26, 80].

Furthermore, although our systematic review and meta-analysis revealed an 84% DCS rate of PrEP use (95% CI: 77–91%), we observed a 62% retention rate of PrEP use (95% CI: 50–74%). HIV infection is disproportionally more frequent among MSM and TGW, and new infections are increasing in this population [81]. Therefore, PrEP is a critical prevention strategy among populations at substantial risk of HIV to reduce new infections [82]. According to RCTs results, once-daily and on-demand PrEP are effective among MSM and TGW. Nonetheless, adherence and retention to this therapy are significant challenges for effective PrEP implementation and are important determinants of the effectiveness of this pharmacotherapy in preventing HIV in clinical practice [44].

The PrEP DCS and RS were relatively similar among the populations studied. Specifically, the PrEP DCS was 92% (95% CI: 0.87–0.97) among MSM and 95% (95% CI: 0.84–1.00) among TGW; while the retention rate to PrEP service was 90% (95% CI: 0.84–0.96) and 91% (95% CI: 0.74–1.00) among MSM and TGW, respectively. The secondary outcome of this review was to assess the barriers to and facilitators of MSM and TGW retention in PrEP provision services. Four studies provided data on this outcome [37, 38, 44, 62]. Socioeconomic factors play an important role in retaining MSM and TGW in PrEP provision services. PrEP offered in public health clinics was a facilitator [44], which is an important finding, particularly in middle-income countries. A study with MSM conducted in the United States suggested that affordable PrEP and care were relevant factors for PrEP retention and continuum care [83].

Moreover, multidisciplinary care [62] and gender-affirming settings [37] appear to be facilitators, as corroborated by Rogers [83], who presented culturally tailored (LGBTQ+) clinical services as an alternative for enhancing PrEP persistence. In a qualitative study with transwomen in Brazil on barriers to and facilitators of PrEP, discrimination in the public health system (SUS) was identified as a barrier to PrEP, and misgendering was identified as a specific form of discrimination, reinforcing the findings of the studies included in this review [9]. Previous data indicate the importance of addressing the social determinants of health and economic barriers, such as the cost of PrEP medication and care, discrimination in health facilities, and the lack of multidisciplinary care. Alternative options include the provision of PrEP in public health services with a multidisciplinary care and the training of health care workers to provide gender-affirming care with sensibility.

This systematic review and meta-analysis had several strengths, including the availability of subgroup analyses by interview strategy and meta-regression to identify possible sources of heterogeneity. Nonetheless, some limitations should be considered. The risk of bias assessment showed that the main problems were related to the measurement of outcomes, participants, and study selection. Furthermore, high heterogeneity exists among the studies in the meta-analyses, which remained high after subgroup and meta-regression analyses. This high heterogeneity can be explained by differences in the study designs, selection bias in some studies, and differences in some population characteristics, such as age and educational level. Another potential limitation of our study is are the limited number of included studies focusing on the description of strategies for DCS on TGW and online strategy-isolating forms.

Raising PrEP awareness among MSM and TGW, minimizing gaps in access, and ensuring retention of PrEP services are critical issues. Offering PrEP through online DCS and RS can reach and retain high numbers of MSM and TGW, and reduce HIV incidence in these populations.