How Black and Latino young men who have sex with men in the United States experience and engage with eligibility criteria and recruitment practices: implications for the sustainability of community-based research

Abstract

Research recruitment, eligibility, and who chooses to participate shape the resulting data and knowledge, which together inform interventions, treatment, and programming. Patterns of research participation are particularly salient at this moment given emerging biomedical prevention paradigms. This paper explores the perspectives of Black and Latino young men who have sex with men (BL-YMSM) regarding research recruitment and eligibility criteria, how their experiences influence willingness to enroll in a given study, and implications for the veracity and representativeness of resulting data. We examine inclusion and recruitment as a complex assemblage, which should not be reduced to its parts. From April-July 2018, we conducted in-depth interviews with 30 BL-YMSM, ages 18–29, in New York City. Interviews were recorded, transcribed, and analyzed using the constant comparative method. Black and Latino YMSM’s responses unveiled tensions between researchers’, recruiters’, and participants’ expectations, particularly regarding eligibility criteria (e.g., age, sex frequency), assumptions about ‘risky behaviors,’ and the ‘target’ community. Men preferred peer-to-peer recruitment, noting that most approaches miss key population segments. Findings highlight the need to critically examine the selected ‘target’ community, who sees themselves as participants, and implications for data comprehensiveness and veracity. Study eligibility criteria and recruitment approaches are methodological issues that shape knowledge production and the policies and programs deployed into communities. These findings can inform how future research studies frame recruitment and eligibility in order to better meet the needs of participants and ensure future engagement.

INTRODUCTION

The sustainability of public health research is dependent upon community engagement and the continued participation of targeted populations (Chiu, 2008). This realization relies on a specific imaginary of a ‘target’ community (Oldenhof & Wehrens, 2018), which has implications for who gets recruited into a given study. In turn, those who are eligible and participate shape the resulting data and knowledge produced (Holt, 2013), which together inform interventions, treatment, and programming. However, public health researchers may lack an adequate understanding of the populations they seek to research (relying on outdated and flawed knowledge) and unintentionally reproduce these imaginaries through their sampling and recruitment strategies (Bungay et al., 2015). HIV is an important area in which to think about the ethical and epistemological implications of community engagement in biomedical research practices and knowledge production (Epstein, 1995, 1998). This paper uses HIV-focused research among Black and Latino young men who have sex with men (BL-YMSM) in the United States (U.S.) to explore how their understanding of eligibility criteria and recruitment approaches influences who engages in research and the type of knowledge produced. It also critically examines the role of the recruiter as a mediator between researchers and targeted communities, with particular attention to the tensions that surface as recruiters try to meet recruitment goals while achieving a representative sample. We argue that recruitment approaches, including eligibility criteria, the selection and participation of community members are critical dimensions of public health research. However, recruitment approaches remain under-theorized within public health scholarship and practice, and there is limited exploration of the implications of recruitment for the comprehensiveness and veracity of resulting data. Furthermore, public health research continues to a lack established guidance on best practices for promoting inclusion and diversity in processes of participant recruitment

Critical considerations in sampling and recruitment

The social sciences have grappled with longstanding concerns about how case selection, eligibility criteria, and recruitment approaches affect research and the resulting co-produced knowledge (Goldsmith et al., 2019). Even so, the issue has received scant theoretical attention in public health regarding how it shapes the research process and knowledge production (for a notable exception see (Fitzgerald, 1996)). Such inquiries are particularly salient given precision medicine and emerging biomedical prevention paradigms and modalities (Montgomery & Pool, 2017). For example, the testing and scale-up of biomedical advancements in HIV treatment and prevention (e.g., clinical trials for pre-exposure prophylaxis (PrEP) (O’Byrne et al., 2019) in pill, injectable, implant, and patch formulations will require participants’ prolonged engagement and provision of personalized medical information. The successful development of medications and interventions that can benefit population-level health has, and will continue to, rely on the inclusion of community members who are often marginalized within society more broadly (e.g., racial/ethnic and sexual minority individuals). As the nature of public health research increasingly shifts towards reliance on the sustained participation of large, often “hardly reached” populations, so must the frameworks that are employed for understanding and practicing ethically-sound recruitment approaches (Gyure et al., 2014).

Major shifts in the regulation of biomedical research have created a push towards greater inclusion of “underrepresented populations” (e.g., women, racial/ethnic minority groups) in research participation. However, as Epstein (2008) and other scholars have discussed (Epstein, 2008), these efforts are often impeded by assumptions about distinct social categories that may not align with the realities, perceptions, and desires of the communities targeted for research and intervention; research studies often rely on static categories to achieve a representative sample, even though these categories often lack nuance and community relevance (Epstein, 2004). This disconnect may both limit the accuracy and scope of research findings, and also reproduce social inequalities, such as through the reinforcement of racial stereotypes or by excluding important segments of the population that do not identify with predetermined research categories (e.g., “gay,” “visible minority”) (Bauer et al., 2020). Indeed, the ways that researchers imagine a target community and the categories that they apply (e.g., ‘Hispanic women with breast cancer’) determine recruitment methods, who is recruited, and who agrees to participate (Garcia et al., 2016). Also, little research has explored how recruiters struggle to balance the tensions to meet recruitment goals while achieving a representative sample: i.e., “ethics on the ground” versus “official ethics” (Heimer, 2013). Although public health frameworks incorporate social science perspectives (e.g., social context, questions of inclusion, and discrepancies between trial participants and those affected by trial results), these considerations are not always incorporated into research design (Mykhalovskiy et al., 2019).

HIV Health Disparities:

In 2018, 21% of the 37,832 new HIV diagnoses in the U.S. occurred among youth ages 13–24 and, if existing incidence trends continue, half of all Black MSM and over one-quarter of Latino MSM will be living with HIV in their lifetime (Centers for Disease Control and Prevention, 2018). The Center for Disease Control and Prevention (CDC) has therefore focused on BL-YMSM through their “high-impact HIV prevention” approach (Gagnon & Guta, 2012). Even so, this approach has not significantly reduced infection rates among BL-YMSM who continue to experience gaps in prevention and treatment (Kuester & Freestone, 2020). When interventions fail to produce intended results, blame is often placed on ‘culture’ or the individual, at the expense of understanding the complexities of HIV prevention (Nguyen et al., 2011).

The failure to adequately address health disparities among BL-YMSM has raised important questions about the limitations of current research frameworks and recruitment practices within the field of public health. Recruitment strategies, including eligibility categories, are well-reasoned within the context of research design, funding, and research ethics. However, they often do not represent participants’ ‘real world’ lives, which includes a complex constellation of risks (Ferlatte et al., 2018). Without careful attention to the way eligibility categories are constructed, researchers may deter individuals from participating because of confusion about eligibility criteria or discomfort with recruitment approaches. Individuals may also fear that their partners or friends will see their participation as a signal that they are deserving of a ‘high risk’ categorization (Philbin et al., 2016). The funding and institutional pressure to complete studies and achieve enrollment targets can result in researchers repeatedly using the same approaches and venues to recruit BL-YMSM. Choosing certain venues may further limit inclusion of eligible participants because attendance at bars or Pride-related events presupposes public ‘outness,’ leading to repeated recruitment and over-representation of certain community segments; this approach may also reify the notion of a singular BL-YMSM experience. In short, the veracity and quality of resulting data is limited if we repeatedly rely on the same pool of participants who prove easier to locate and thus systematically miss people who are never recruited or who, once recruited, choose not to participate.

Efforts to improve researcher approaches that better align with the realities of people’s lives has led to a rise in “community participatory” approaches. However, studies have shown that the structure of the research industrial complex can lead to data fabrication (Kingori & Gerrets, 2019). This may result from underlying power dynamics and inequalities present in the researcher/fieldworker relationship: poor working conditions, inadequate training and communication, and a lack of formal recognition of the unique contribution (e.g., knowledge, skills) and challenges (e.g., pressure to recruit an unrealistic number of participants) that the fieldworker brings to the development of research findings (Dorcas et alt. 2013). It may also stem from senior researchers’ assumptions about a fieldworker’s ties to the local community, which may be reinforced by the fieldworker’s own assertions about community affiliation in order to gain economic and social opportunities (Kingori & Garrets, 2019).

The current study builds on existing work that focuses on how research practices affect knowledge production (Epstein, 1995, 1998; Fisher et al., 2017; Flicker et al., 2015), and contributes to longstanding concerns about how case selection, the structure of eligibility criteria, the act of recruitment and who chooses to participate in a given study affect research findings. Given advances in biomedical HIV prevention and treatment, the demands placed on participants who engage in HIV-focused studies has expanded to encompass everything from observational research (i.e., tracing the outcome of a treatment over time), to intervention trials (i.e., comparing outcomes between a group exposed to a medication or intervention and an unexposed “control group”), and program rollout (i.e., assessing the efficacy of program implementation) with significant variation across study type of the risks, benefits, and circumstances of community participation. Given shifts toward long-acting modalities of HIV treatment and prevention (e.g., long-acting injectable ART and PrEP), and the increased demands on participants, we conducted a qualitative study to explore these tensions among BL-YMSM.

METHODS

We conducted in-depth interviews from April-July 2018 with 30 BL-YMSM in New York City; interviews lasted 60–90 minutes. Eligibility criteria included identifying as male; Black/African American or Latino; being 18–29 years of age; and reporting having had sex with a man. Eligibility criteria also included having been recruited for previous HIV-related prevention research studies—we define recruitment as individuals who were approached and asked if they were interested in being screened for inclusion in a study—this included individuals who had previously been recruited in other studies but who both declined or were ineligible. These studies into which they had been previously recruited encompassed behavioral-focused studies on the uptake of condoms and safer sex negotiation among BL-YMSM. A portion of the individuals we recruited had also served as recruiters and could thus speak explicitly to the tensions they experienced in serving as recruiters in the communities within which they lived. We contacted men who had been previously recruited by Columbia University HIV-related prevention studies and provided permission to be contacted. Specifically, we used a database of contact information that a colleague had generated in his HIV prevention research with BL-YMSM. When individuals were recruited into those studies, recruiters would ask if they were interested in participating in future studies, regardless of their eligibility for, or interest in, the current study.

Interview topics included a brief overview of men’s social lives (e.g., family, education, work), followed by a series of questions about their experience with HIV-related research, including preferred recruitment strategies and venues and, when applicable, their own role as recruiters. Participants provided verbal informed consent and received $50 and a roundtrip subway card. The Columbia University Institutional Review Board approved this study.

Interviews were recorded, transcribed and uploaded into Dedoose (Dedoose, 2020). Study team members analyzed interview data within and across cases using the constant comparative method (Glaser & Strauss, 1967). Three team members independently conducted line-by-line coding on five transcripts to create an initial codebook. We then generated a list of thematic codes based on existing literature and integrated those into the codebook to ensure the inclusion of both theory-based and emergent concepts (MacQueen et al., 2007). This included theories around ‘official’ versus ‘on the ground’ ethics, power dynamics with community-based research, and categories and counting (Freedman 1998; Epstein 2008; Heimer, 2013; . Team members subsequently independently cross-coded a random sample of five additional transcripts to refine the codebook. Once refined, two study team members coded the remaining 20 interviews using the codebook as a guide to ensure relevant themes (both theoretical and emergent from the data) were captured. The research team conducted a series of twice-monthly meetings to resolve discrepancies and triangulate data findings.

RESULTS

Participants’ median age was 26 (range 20–29), and the majority was employed and had health insurance (primarily Medicaid). The high level of Medicaid coverage is a consequence of men being underemployed or employed in low-wage jobs (e.g., fast food or coffee shop) that did not provide health insurance. The sample ranged in education level, income, and sexual identity (Table 1). Two primary areas of tension emerged: 1) misunderstanding of eligibility criteria and 2) suggestions regarding recruitment.

Table 1:

Demographic Characteristics of Black and Latino YMSM

Characteristic	Total N=30
Age	26 (median)
18–24	8
25–29	22
Sexual Identity
Gay	19
Queer	5
Bisexual	1
Straight	1
No Label	4
Race (self-report)
Black	18
Latino	5
Mixed	7
Income
Monthly Range	$400–9000
Education
<High School	3
High School/vocational	8
Some College	8
College or more	11
Employed
Yes	24
No	6
Insurance Status
Public	18
Private	10
Uninsured	2

(Mis)understandings and problematic categorization of eligibility criteria

Race, age and sexuality

While men understood the need for eligibility criteria (e.g., “that’s how you get an unbiased result”) (27 y/o, Latino, MSM) as a necessary condition of rigorous public health research, they also expressed frustration. Many felt that the categories used to determine participant eligibility did not align with their identities and lived realities, and often served to exclude important segments of the population. One participant described the exclusionary impact of eligibility categorization as “a form of segregation” (27 y/o, Latino, MSM). This problematic was most frequently discussed in terms of race, gender, and sexuality.

Men described how research often focused on Black men, which excluded others who might also be at risk, “Straight people get it, white people get it. Don’t single out Blacks, Black gays, or youth. [Focus on] Latinos” (22 y/o, Black, Gay). Age-based exclusion criteria also frequently failed to represent their experiences, “I don’t think everything should be aimed for youth. People over 25 need help. It’s always that 13 or 18–25 age group, [but] they’re not the only people that are HIV positive” (22 y/o, Black, Gay). Sexual orientation was another contested category: “I tell people, ‘Yeah, I have sex with men.’ But I never tell anybody I’m gay or homosexual. I never use them terms” (28 y/o, Black, MSM).

In addition to segmenting the community, some men felt that the way eligibility was determined (i.e., through screeners) created an adversarial tone that suggested inherent mistrust, “It’s about catching me in a lie or something. It’s repetitive questions asked differently” (21 y/o, Black, Gay). Men stressed the need for inclusion, but also emphasized the importance of better tailoring studies towards diverse and distinct populations (e.g., differentiating between MSM and transwomen).

Behaviors and timeframes

Men expressed frustrations with timeframes related to behavior-focused eligibility criteria (e.g., unprotected sex last three months versus six months). Men felt that these timeframes were arbitrary and could eliminate participants whose voices needed to be heard: “Studies should be for all types of men, they shouldn’t be the promiscuous one or you got to meet certain criteria. You got to be having sex with the world for them to recruit you” (22 y/o, Black, MSM). The way that studies characterized the timeframes for inclusion varied by researcher and institution, as well as type of study (e.g., behavioral versus clinical, individual versus couples etc.). This segmentation both excluded certain populations and had implications for how risk perception became (often inadvertently) communicated to prospective participants (e.g., if you are ineligible because you had unprotected sex in the last six months but not last three months then perhaps you do not need HIV prevention). It also had implications for how others viewed those being recruited: “It may put ideas in their [friends’] head about, ‘Well, I see that it says HIV and they’re recruiting for gay men, so obviously either all gay men have HIV or they’re at a higher risk’” (25 y/o, Black, Gay). It also affected how individuals viewed their own or their friends’ behaviors (e.g., the notions of promiscuity noted by the 22 year old, above).

Recruiter Dilemmas

Some men had been both participants and recruiters. Recruiters described struggles with eligibility criteria, and how it could place them at odds with the target communities, of which they were often members. The majority of men we interviewed did not understand that they might be ineligible for a given study, and recruiters described feeling that they were deceiving men about the likelihood of participation:

It didn’t work out for you and I can’t tell you why? It feels like exploitation because you didn’t get what I promised. I look like a jackass. I wasted your time. I can’t even give you an answer and it’s like I just used you

(29 y/o, Mixed-race, Pansexual).

These tensions have implications for who participates in current and future studies and recruiters’ relationships with their communities. First, the majority of men in our study who also served as recruiters described using their own social media (e.g., Grindr, Facebook, Craigslist) and personal networks to encourage study participation. While recruiter participants described recruiting people through their own social media or personal networks to be the most efficient, some expressed discomfort with this approach. Superficially, they believed it attracted potential participants who did not care about the purpose of the study and participated solely for the money. Also, recruiters felt like they were putting themselves in harm’s way by risking their own exposure to strangers, sexual predators and “creeps.”

“I understand the thing about numbers but you get the sleaziest of the sleazy people--, at a certain point it was like, “Well, maybe we should just go on Grindr and recruit people.” But that is emotionally tolling because the people on Grindr aren’t thinking that I’m here [for research]. Even if I made a fake profile I have to engage with this sketchy thing that I usually don’t do…

(29 y/o, Mixed-race, Pansexual).

Recruiters described pressure to reach certain recruitment goals, which led one to note: “I don’t know why the criteria was so strict. And that was where I was like, okay, [where are the] grey areas...” (29 y/o, Mixed-race, Pansexual). These “grey areas” may include criteria like gender, sexuality or even age that recruiters cannot dispute. While recruiters understood the need for criteria, they wanted individuals they screened to be eligible and to thus receive the remuneration: “There’s a few straight men that hang out with these guys. And we would tell them to lie…to get a little money if I know that person needs it” (29 y/o, Hispanic, Gay). Recruiters described cajoling people to complete the screener, especially since ineligible individuals often expressed frustration and unwillingness to participate in future studies. This affected recruiters’ relationships within the broader community, and made subsequent recruitment challenging because of some people’s wariness to be screened again.

This research highlights potential disconnects between the motivations that drive recruiters and the messages received by participants. Recruiters endeavored to meet research goals (e.g., number of people recruited), contribute to knowledge in ways that supported their communities and its vulnerable members, and to find participants who want to contribute “for the right reason.” However, they faced considerable challenges in reaching these objectives due to structural inequalities (e.g., poverty) and other social dynamics, such as prospective participants engaging with recruiters for ulterior motives (e.g., to “hook up”).

Research findings are likely affected by the ways recruiters navigate the “messiness” of real life circumstance. For example, they may modify established recruitment strategies in order (e.g., to avoid creating tension/conflicts within their own communities; in response to pressure from researchers to meet recruitment goals; and in order to increase their own social status via continued opportunities to be hired as a recruiter in the future. Study recruiters are rarely involved in the design of the study; they are hired once the eligibility criteria and approach are determined. Their earlier involvement might help alleviate some of these challenges and tensions between what their role is and what gets communicated to, and understood by, the research participants.

Preferred Modes of Recruitment

Many studies with BL-YMSM recruit through similar methods, including online, or at bars, parties, or Pride events. Men agreed that researchers often approached recruitment in ways that could limit whom they include, and miss certain key groups of people.

Everyone’s not an in-your-face person. Everyone’s not on the apps. Check out places that aren’t as public. Like doctor’s offices or support groups that cater to gay people. Because people that are affected are not on Grindr anymore, or don’t go to Pride. It’s about going where the people that you want to talk to are. And sometimes they’re not in gay settings

(28 y/o, Biracial, Gay).

However, men did not always agree about how they would like to be approached, highlighting the importance for overall sensitivity on the part of the recruiters.

Person-to-person recruitment

The majority preferred person-to-person recruitment, particularly because it signaled that researchers cared and were willing to invest more time. Some men expressed a preference for recruitment in ‘gay friendly’ venues: “Make an announcement where the drag shows are... Even at two in the morning…But you’ve got to let them know up front hey, you want to make quick money” (29 y/o, Hispanic, Gay). Many expressed willingness to take a survey at a bar, though felt that being approached might shift the night from socializing to research, which could be frustrating, “I was a at a bar and I took the survey, but my friends were annoyed. Because they had to wait on me to try to get in” (29 y/o, Black, Gay). Others did not want to be approached at a bar, including because it could ‘out’ them, “I think I would feel more of an invasion of privacy at a bar than on an app” (28 y/o, Black, MSM).

App or poster/flyer-based recruiting

Men had differing opinions about the effectiveness of recruiting through posters/flyers circulated on apps or public transportation:

It [Grindr] does have an influence, because if you’re in a bar, you’re having a couple drinks, and you want to hook up. You’re going to find this person talking to you about HIV annoying, and you’re going to tell them, “You know what? Get away from here”

(27 y/o, Latino, MSM).

Men had mixed feelings about the comprehensiveness of app-based recruitment, and its potential impact on data quality:

I think that whole Grinder method, will get you older men that are sitting online all day, or a young person that don’t have nothing to do, that is just coming in for a conversation versus to really channel what the study is for. Versus going to places where the age bracket or race is dominant, you get a more realistic and personal approach

(26 y/o, Black, Gay).

Men’s dislike of subway or bus ads was uniform, because they felt it would not recruit a representative sample. This sentiment suggests that while using app- and venue-based recruitment will target a certain sector, it would miss individuals “living their life” who might have qualitatively different experiences and behavior patterns (Pisani, 2008).

DISCUSSION

Our study findings reveal that research participants and recruiters continue to experience tension between desires to be included in HIV research and pressing concerns with the conditions and contingencies that their participation entails: a phenomenon that arises from a long history of targeted HIV research (Epstein 1995). Interview participants described an assemblage of factors that keep BL-YMSM in a complex and protracted engagement with public health research, including the conflict between their scientific-technical knowledge (e.g., the necessity of inclusion criteria) and their embodied (racialized and sexualized) knowledge. These findings serve as a starting point to complicate discussions about public health research and ethical engagement.

In response to parachute models of research in which findings did not come back to benefit community members, HIV activists and people living with and at risk of HIV assumed a strong role in challenging scientific authority and in influencing the HIV biomedical research trajectory through the imposition and credibility of their own “lay expertise” (Epstein, 1995). However, their status as other (at risk) created additional tensions between the desire for research to be as inclusive as possible (a community-building activity) and tailored to their immediate needs (only approach these kinds of people at these times under these circumstances). Complexity is further reflected in the role of the recruiter who is working with and of the community, yet sometimes simultaneously positioned in opposition to the community by research demands and stipulations. Such dynamics are constituted by spatial (from the nightclub, to the bus, to on-line platforms) and temporal factors (past, present, and future sexual and research encounters), which shape and reshape meaning and interpretation. The routine and mundane act of being deemed ineligible (screened out) represents an act of systematic exclusion among communities who are already marginalized.

Participants understood the purpose of eligibility criteria but frequently disagreed with its structure (e.g., specific age, race/ethnicity, or ‘risk behavior’ requirements) and referred to it as ‘arbitrary.’ Participants saw eligibility criteria as creating artificial segmentations and sub-categories (e.g., around race and gender) that did not reflect the complexity of reality, a perspective previous studies have critiqued (Mukherjea & Vidal-Oritz, 2006). HIV/AIDS has been called an “epidemic of signification” (Treichler, 1999) because it creates categories and communities such as ‘most at risk populations’ or ‘key populations’ but often ignores the fluidity of these categories in people’s daily lives (Mukherjea & Vidal-Oritz, 2006). Research has therefore both conflated populations and produced artificial categories that are then absorbed as reified labels in social spaces.

Recruiters were simultaneously constrained by the expectations of the research system and the community in which they live. Studies have acknowledged the complexity of navigating dual and often oppositional roles and described these challenges as underestimated and under-researched (True et al., 2017). Recruiters worked to provide ‘good data,’ but that goal was complicated by enrollment targets: this led to the repeated recruitment of the same individuals, which meant that entire sub-groups who were more challenging to recruit could be repeatedly excluded. Studies have described tensions around how recruiters apply eligibility criteria, and the criteria itself, as the ‘burden of compliance’ (Heimer, 2013). This is particularly salient in settings where certain criteria (e.g., age) may be less meaningful than social milestones (e.g., marriage or children) (Heimer, 2013). This has been referred to as “official ethics” versus ethics “on the ground,” (Heimer, 2013) as recruiters often struggle to balance expectations conceived by researchers and what they must navigate within the community. Heimer (2013) referred to the moment-by-moment decisions that recruiters face and described how these “cumulate over time into routinized but not fully codified ways of doing” that then inform the research process (page 374). This is particularly salient because clinical and university settings often frame recruitment as menial labor, and do not pay sufficient attention to who serves as the recruiter and implications for the type of data collected. These ‘official ethics’ bring into stark relief how the majority of research studies begin by framing an ideal research sample versus acknowledging the ‘on the ground’ ethics and messiness that might affect their sample. One option to increase the validity of the data and to ensure that only individuals who meet the eligibility criteria participate could be to create a separate ‘community engagement contingency’ fund within a study to provide remuneration to individuals who are not otherwise eligible.

Eligibility criteria may also convey messages about who is most ‘at risk’ (e.g. by race/ethnicity, age, or behaviors) and thus need to be researched. Since eligibility criteria often overlaps with racialized and non-heteronormative sex, it may associate an already marginalized community with stigmatized behaviors, such as unprotected anal sex or injection drug use (Mukherjea & Vidal-Oritz, 2006). Such criteria may also unintentionally convey that certain types of sex, or sex at certain intervals, is worthy of intervention and thus problematic. Conversely, these criteria might also inadvertently communicate that other sexual practices or patterns are lower-risk, which may be inaccurate. Men shared how their understanding of eligibility criteria, coupled with frustrations about being ineligible, meant that they were frequently wary of returning for subsequent studies. This self-exclusion further winnows the pool of future participants, and makes it unclear whether resulting data are reflective of any given community.

Men’s opinions about where and how they wanted to be recruited varied though recruitment approaches influence the resulting sample. Research found differences of nearly 20% in reported HIV testing rates depending on internet platform (Merchant et al., 2017), and internet-based versus in-person approaches have yielded significant differences based on age, race, HIV status and substance use (Chen et al., 2018). In addition, individuals who enroll in studies often fail to reflect the full landscape of BL-YMSM. This is likely because individuals who are most economically stable lack the time or willingness to participate (Pisani, 2008). In addition, the very means and assumptions about recruitment—and who is most at risk—often skew the sample to incorporate individuals who are more structurally vulnerable. In addition, individuals who are more economically and structurally insulated may hold the (potentially incorrect) assumption that their sexual practices are less risky. Such assumptions may influence their willingness to participate in HIV-prevention studies, which leads to this demographic being excluded.

There is an increasing body of social science work that shows the “diverse ways in which ethical norms are negotiated, transgressed, and transformed within the constraints of institutions and according to the interests of different actors” (page 106) (Cooper, 2018). However, this social science work is rarely incorporated into, or applied, in public health literature in ways that facilitate critical thinking around recruitment, the role of the recruiter, and the resulting knowledge production. Future research should recognize how research recruitment conveys certain messages that may be internalized by potential participants, thus reproducing ideas of stigma and stereotypes. It should also examine ways that recruitment approaches impose limitations on generating novel insight as certain individuals repeatedly participate while others never join, and how this changes across time and space.

Limitations and Strengths:

We included both men who were recruited but refused to participate in a previous study, and those who chose to participate, further widening the opinions of individuals included in this research. However, we were unable to recruit younger men (18–19 years of age) who might have different experiences with recruitment. This study was conducted in New York City, and findings might be different than in smaller cities that have fewer ongoing research studies. Future research could build on this work by exploring these tensions among rural populations or individuals recruited for research outside of HIV. Finally, we recognize that our attempt to capture the preferences and motivations of research participants and recruiters may itself reproduce the kinds of limitations and power dynamics in public health research that we are trying to challenge. However, combining the empirical and theoretical work will help advance, and unite, the fields of empirical research ethics and public health ethics.

Conclusions:

Eligibility criteria and recruitment approaches are more than a methodological issue: they shape knowledge production and the kinds of policies and programming that are subsequently deployed into these communities. This is one of the first studies to explore these tensions among BL- YMSM who are frequently recruited, and among men who serve as recruiters and must, therefore, navigate these ‘on-the-ground’ ethics in real-time. Understanding these phenomena is particularly important at this moment given the need for sustained research participation in biomedical and precision medicine research. Social science research has long stressed the need for biomedical and epidemiological studies to incorporate socio-cultural research that speaks to the complicated ways that HIV intersects with communities, and the implications for how knowledge is produced. Even so, we’ve reached a tipping point where so much research has occurred among YMSM that it has changed how they understand risk at a community level. Findings may help inform how future research studies could frame recruitment and eligibility in order to better meet the needs of participants and ensure future engagement. Doing so will also help facilitate the accuracy of the data that these studies produce and the interventions and programs that are subsequently deployed into these communities.

While this paper is not the first to point out the limitations of eligibility criteria and its impact on knowledge production, this discussion has been under-theorized, especially in public health. In addition, no changes have been made to adjust research to address these limitations. This paper demonstrates how eligibility criteria are deployed and subsequently taken up in the lives of BL-YMSM. For example, we describe what it means for a bus with an ad to pass by, or to confront an ad for research on Grindr, or to be told one’s behavior does or does not match a given ‘risky’ criteria. Including this focus on context shows how recruiters and participants engage in negotiations across time and space to produce the knowledge that then determines how communities are counted and categorized. This paper helps extend the ongoing debate at the intersection of ethics, methods and knowledge production by reviewing data from an applied study focus on HIV among a population disproportionately targeted for HIV studies.