Recruitment of New Immigrants into a Randomized Controlled Prevention Trial: The Live Well Experience

Abstract

Minority populations are hard to reach with prevention interventions because of cultural and logistical barriers to recruitment. Understanding how to overcome these barriers is pertinent to reducing the elevated burden of obesity within these underserved communities. To inform this literature gap, we explore the processes and outcomes of recruitment for Live Well - a randomized controlled obesity prevention intervention targeting new immigrant mothers and children from Brazil, Latin America, and Haiti who were residing in the greater Somerville, MA area. We employed community-based participatory research principles to develop and implement five culturally-adapted recruitment activities (posters and flyers, media announcements, church outreach, participant referrals, and community organization partnerships) and tracked enrollment for the total and stratified samples of 406 dyads (37% Brazilian, 29% Latino, 33% Haitian). We describe how strategic partnerships were built and sustained within the intervention community, and detail the key adjustments that contributed to our success. Ultimately, community organization partnerships and participant referrals enrolled a collective majority of participants (34% and 25%, respectively); however, stratified analyses revealed variation by ethnicity: Haitian immigrants responded best to ethnic- based media announcements (44%), whereas Latino and Brazilian immigrants were most responsive to community organization outreach (45% and 38%, respectively). Implications from our findings enhance the literature on recruiting hard-to-reach communities into prevention research: some less integrated communities may respond more to grassroots activities with direct engagement, whereas communities with more social capital may be more responsive to top-down, community-wide collaborations. Furthermore, we suggest that strategic and trusting partnerships are key facilitators of recruitment, and future researchers must understand communities' culture and social networks when building relationships. Our analyses provide rare insight into best practices to overcome specific cultural barriers to recruitment which future investigators can use to better reach underserved communities with prevention research.

Background

New immigrants in the United States are disproportionately affected by overweight and obesity compared to native-born peers (Akresh, 2007; Barcenas et al., 2007; Koya & Egede, 2007; Roshania, Narayan, & Oza-Frank, 2008; Sanchez-Vaznaugh, Kawachi, Subramanian, Sánchez, & Acevedo-Garcia, 2008; Singh, Siahpush, Hiatt, & Timsina, 2011). Upon immigration they have healthier average weights; yet, for both adults (Akresh, 2007; Antecol & Bedard, 2006) and children (Strickman-Stein et al., 2010), obesity increases with length of residency in the US. Research suggests increased stress, decreased feelings of social support, and an obesogenic environment – characterized by accessible energy dense foods and few opportunities for physical activity – are all determinants of obesity within this population that must be addressed (Antecol et al.; Tovar et al., 2013). Moreover, projections that new immigrants will account for 82% of the US population growth between 2005 and 2050 (Passel & Cohn, 2008) provide impetus to better engage new US immigrants with obesity prevention interventions.

Despite this well-documented risk, new immigrants are underrepresented in the current health literature (Lindberg & Stevens, 2007; Tovar, Renzaho, Guerrero, Mena, & Ayala, 2014). Few interventions have been developed for immigrant communities, and those that have been developed, are limited by difficulties with recruitment and retention, including small and unrepresentative samples and high attrition rates (George, Duran, & Norris, 2014; Lindberg et al., 2007; Tovar, Renzaho et al., 2014; Vincent, McEwen, Hepworth, & Stump, 2013). Researchers postulate that these obstacles stem from population transiency, the lack of valid and culturally appropriate measurement tools, language barriers, limited time for research participation, and cultural barriers such as mistrust and concern about immigration status (George et al., 2014; Tovar, Renzaho et al., 2014). Additionally, the time and effort needed to recruit and retain these populations are often cost prohibitive given the current funding structure which is often restricted to short durations and limited in amounts awarded.

Current research suggests that culturally adapted approaches can help overcome recruitment barriers within diverse minority populations (Corsino et al., 2012; George et al., 2014; Martin, Negron, Balbierz, Bickell, & Howell, 2013; Martin et al., 2011; Tovar, Renzaho et al., 2014). For example, the incorporation of cultural adaptations helped to reach recruitment targets for two randomized controlled diabetes prevention trials in Mexican immigrant communities in the US (Martin et al., 2011; Vincent et al., 2013). Similarly, the incorporation of community-based participatory research (CBPR) principles is also critical when recruiting underserved populations. Under the CBPR paradigm, researchers are encouraged to develop strong partnerships with community organizations, build connections with community leaders, and directly engage community members (Leung, Yen, & Minkler, 2004). This framework has been successful for recruiting in Latino immigrant communities (M. Rodríguez, J. Rodríguez, & Davis, 2006), and provides a foundation upon which to build further evaluations in other underserved immigrant communities (Tovar, Renzaho et al., 2014).

To address this research gap, we explore the strategies and activities employed during recruitment for Live Well – a community-based randomized, controlled, obesity prevention intervention for new immigrant mother-child dyads. Live Well was based primarily in Somerville, MA – a ‘gateway’ city for new immigrants. Recruitment efforts targeted dyads from the three largest immigrant groups in Somerville: Brazilians, Latinos, and Haitians. Specifically, this article focuses on the processes and outcomes of recruitment. We describe how strategic partnerships were built and sustained within the community and detail all aspects of the recruitment plan including development, implementation, and evaluation, with an emphasis on our results. We identify culturally-specific activities for overcoming recruitment barriers within a diverse, underserved community, which we hope can help future investigators understand and reach analogous at-risk populations.

Methods

Participants

Live Well aimed to recruit 400 mother-child dyads that met the following eligibility criteria: mothers had immigrated from Haiti, Brazil, or other Spanish-speaking Latin American countries (primarily from El Salvador, Colombia, Guatemala, Dominican Republic, and Honduras), resided in the US for no more than 10 years, were 20–55 years old with a child between 3–12 years old, lived in the Greater Boston, MA area, were not pregnant or were at least 6 months postpartum. Participants had to be willing to participate in either the control group or the intervention group. Those eligible and willing would be randomly assigned to either a waitlist control group (who would receive the delayed intervention) or to a two-year lifestyle intervention comprised of a one-year popular education curriculum on nutrition and physical activity and phone-based motivational interviews and a one-year civic-engagement program (Tovar, Boulos et al., 2014). In our sample (n=406), 66% of Latino mothers reported that they migrated for economic opportunities and 16% for their family, and none reported migrating for political reasons. Similarly, 72% of Brazilian mothers reported that they migrated for economic opportunities, 10% for their family, and none for political reasons. Among Haitian mothers, 40% reported migrating for their families, 25% for other reasons, including the 2010 Haitian earthquake, 24% for economic opportunities, and few (8%) for political reasons. Additionally, a large majority of members from all three ethnic groups migrated from urban environments (65%, Latino, 89% Brazilian, 88% Haitian; Tovar, Boulos et al., 2014).

All aspects of the study design and recruitment procedures were approved and monitored by the Institutional Review Board at Tufts University.

Engaging Community Stakeholders – Forming the Live Well Steering Committee

The Greater Boston area, and in particular the city of Somerville, is home to many immigrants from these countries; and, in response to the growing immigrant population and increasing racial, linguistic, and socio-economic diversity, many community-based organizations have been established to provide social and health services to meet their needs. Rooted in CBPR principles (Leung et al., 2004), we sought to form a diverse group of community and academic partners to oversee all aspects of Live Well. Our efforts began by forming the Live Well Steering Committee in 2007, prior to the start of the grant in 2009. During this time, Tufts University researchers capitalized on strong pre-existing connections to the immigrant community in Somerville, which were formed during two previous CBPR interventions – Shape Up Somerville (Economos et al., 2007) and the Somerville Community Immigrant Worker Health Project (Panikkar et al., 2012) – and further built upon during two years of formative research (Tovar et al., 2013). From these collaborations, we identified six community partners and ensured equal representation across target ethnic groups. Participating organizations included the Brazilian Women’s Group, which represented the Brazilian community of Greater Boston; the Community Action Agency of Somerville and The Welcome Project, which primarily reached the Latino population; the Haitian Coalition of Somerville, which reached the Haitian community; the Immigrant Service Providers Group/Health, which served all immigrant groups in Somerville; and Tufts University, which represented the research community. Ultimately, the committee comprised 17 individuals from six culturally and professionally diverse partners: one project manager and three project coordinators (one representing each ethnic group), Tufts researchers, students with diverse backgrounds, support staff, and six community organization leaders.

Key responsibilities of the committee included developing the Live Well study design, eligibility criteria, and recruitment plan; overseeing the process implementation and evaluation; and making ad hoc study design adjustments to achieve recruitment goals. The committee convened at monthly in-person meetings over the course of the recruitment and intervention periods, chaired by the principal investigator, and meeting locations alternated between local community settings and the university to accommodate all stakeholders.

In addition to the broader committee, members formed smaller groups to oversee specific processes. For example, recruitment advisory subcommittee members – academic researchers, project coordinators, study staff, and community representatives – guided all aspects of recruitment implementation and analysis.

The Recruitment Plan

Formative Research

Prior to the start of the grant, we conducted focus groups with new immigrant mothers from each ethnic group to better understand our target populations (Tovar et al., 2013). Discussion topics were informed by key informant interviews with community leaders, partner organizations, and university investigators. Briefly, participants emphasized a plethora of differences in beliefs, attitudes, and behaviors related to health and nutrition in the US compared to their home countries. Response themes informed the development of the Live Well curriculum and Recruitment Plan. For example, the emergent theme of stress surrounding childcare arrangements encouraged us to offer childcare during Live Well discussion sessions.

At project onset, we conducted a literature review on common practices of minority recruitment to ensure an evidence-based approach. Overall, findings from previous work in Latino communities emphasized the use of CBPR principles: building institutional partnerships with churches, schools, and WIC centers; raising community awareness through posters and flyers, radio announcements, and presentations at local community organizations; employing bilingual research assistants from the intervention community; and encouraging existing participants to recruit within their personal networks (Deren, Shedlin, Decena, & Mino, 2005; Larkey et al., 2002; Rodriguez et al., 2006). Additional strategies included providing monetary incentives and offering free meals at program sessions (Dumka, Garza, Roosa, & Stoerzinger, 1997), and as stated, providing childcare. Furthermore, the literature also conveyed the importance of creating strong community-academic partnerships to oversee all aspects of study implementation, as well as reaching out to trusted community ‘gatekeepers’ who could refer friends, family, and peers (Levkoff & Sanchez, 2003).

These findings were analyzed by the recruitment advisory subcommittee and further circulated and discussed with steering committee members and community representatives. Ultimately, our research both informed our multi-pronged approach to recruitment and, given the dearth of formal comparisons of recruitment activities’ effectiveness and applicability in other underserved communities, served as a call-to-action for a more rigorous comparative evaluation.

Strategies

The literature we reviewed above informed our three main recruitment strategies: institutional partnerships, community awareness, and individual outreach. The corresponding goals for each strategy are summarized in Table 1. Briefly, institutional partnerships connected the research and intervention community through existing organizational networks; community awareness spread information about Live Well at the community level and shared knowledge of obesity as a public health concern; and individual outreach directly engaged individuals and encouraged recruitment through personal networks.

Table 1.

Strategies, corresponding goals, and implemented activities employed for the Live Well Recruitment Plan to recruit Brazilian, Latino and Haitian new immigrant mother-child dyads from the Greater Somerville, MA area.

Strategies	Goals	Recruitment activities
Institutional Partnerships	Create strong partnerships to enhance connection to target population	Community organization engagement
	Broadcast Live Well and generate active referrals	Church outreach
Community Awareness	Spread community knowledge of Live Well throughout the community	Poster and flyer canvasing
	Increase community understanding of obesity as a public health concern	Media-public service announcements Church outreach
Individual Outreach	Connect one-on-one with target populations	Participant referrals
	Identify lay-leaders within the community to enhance recruitment	Poster and flyer canvasing

Activities

To implement our three strategies, we employed the following five activities: First, we developed posters and flyers. All print materials emphasized building healthier families and a stronger community to appeal to the widespread new immigrant aspiration for a better life (see Figure 1). We developed content in English based on formative research by steering committee members with active input from other community partners. Project coordinators pretested preliminary materials within each community and revised accordingly (Tovar et al., 2013). In addition, project coordinators who were bilingual and fluent in English and at least one of the 3 languages (Haitian-Creole, Portuguese, or Spanish), translated all content and pilot-tested all posters and flyers within each community. To account for the varying education levels across the immigrant populations, we used images to help convey our messaging and varied these images according to ethnic group to improve cultural specificity. Posters and flyers contained information about the project, partnering community organizations, time commitment (1–2 years), compensation ($200 stipend), and a contact phone number. We developed one poster per ethnic group and distributed approximately 500 copies in each language throughout the community at strategic locations: community venues, ethnic markets, supermarkets, transit stations, childcare settings, beauty salons, restaurants, schools, and via teachers, sending flyers home with students. Staff and community partners also handed out flyers to passersby near these central locations.

Fig 1.

Sample of a multi-language Live Well recruitment poster. The poster includes study information written in each of the three languages used in Live Well, Spanish, Haitian-Creole, and Portuguese, and an image of a mother-child dyad from each of the three ethnic groups. Phone numbers have been redacted.

Second, we developed Media Public Service Announcements (PSAs) that described Live Well and provided a phone number for follow up. Steering committee members developed the initial scripts, which were translated and pretested by project coordinators and revised according to community partner and member feedback. We aired announcements twice a day during recruitment, in all three languages, on six ethnic radio stations (two per ethnicity), reaching the Greater Boston Area. Live Well project staff also participated in radio interviews which aired on the same six stations, at least one per month for each of the three ethnic groups, in all three languages.

Third, we identified and partnered with churches representing each immigrant community. Church staff made announcements at church meetings and each church held evening orientation sessions at least once per week for each of the three ethnic groups where Live Well project staff presented the project.

Fourth, we recruited through participant referrals. We encouraged enrolled participants to refer friends within their networks to either attend orientation sessions at local organizations or to call project staff directly.

Lastly, we formed strategic community organization partnerships with 16 additional community organizations to disseminate information about Live Well and to support recruitment through active referrals. Example partners included: English as a Second Language (ESL) educational sites, Women, Infants, and Children (WIC) sites, community health centers, and family programs. To initiate these partnerships, we sent letter requests for institutional support and subsequently fostered these relationships through meetings with the project coordinators. We also identified additional organizations with close connections to the immigrant community in Somerville from previous collaborations with steering committee partners and we contacted these groups directly.

Once community partnerships were formed, several activities followed. First, we distributed posters and flyers in the organizations’ community buildings. Second, organizations hosted evening orientation sessions to facilitate discussion about Live Well. At each session, attendees received information about the Live Well intervention, time commitment, and randomization processes; and, as per community representative recommendation, watched an educational film about health disparities (Unnatural Causes). Third, community partners generated active referrals. For example, leaders provided Live Well staff with a list of other organizations to contact. Others, such as the Liaison Interpreter Program of Somerville (LIPS), had a more direct engagement: high school students in the LIPS program (bilingual students who learned language interpretation skills) received training on Live Well recruitment activities to recruit within personal networks.

Study Enrollment

Participants who learned about Live Well from posters and flyers, media-PSAs, or participant referrals called project staff for eligibility assessment; if eligible, they were assigned to an upcoming measurement day and administered informed consent. Those who attended orientation events held at churches or partner community organizations were assessed for eligibility in person after each presentation and given informed consent. We made follow-up phone calls to all potential participants to ensure attendance at measurement days. We obtained informed consent from all mothers; assent from all children >7 years old; and written permission for all children <7 years old (all consent forms were translated in 4 languages). Once participants completed all baseline measurements, they were randomly assigned to the intervention or control group described above using a computer program.

Key Adjustments to Achieve Recruitment

Initial eligibility criteria were as follows: mothers were of Brazilian, Latino or Haitian descent, resided for ≤5 years in the US, were 20–55 years old with a child between 5–12 years old, had lived in the Greater Somerville, MA area, and were not pregnant or ≥ 6 months postpartum. Although initial criteria were developed based on research and input from community partners, the first four months of recruitment yielded low eligibility (16%) leading the subcommittee to discuss amendments.

Ultimately, adjustments to the eligibility criteria included changing the definition of “new immigrant” from ≤5 years to ≤10 years residency in the US (Tovar, Boulos et al., 2014); lowering the age limit of the child from 5 years old to three 3 years old, since it was less common for mothers to immigrate to the US with young children or already pregnant; and, broadening outreach to nearby communities with similar percentages of Brazilian, Latino, and Haitian residents, given the observed pattern of dispersion from Somerville to areas in Greater Boston with more affordable housing. We also amended the study design by shortening the Live Well popular education intervention curriculum to 1 year instead of 2 years, revising recruitment targets, and introducing two cohorts into the study: cohort 1 received a two-year intervention with the Live Well curriculum in year one and a focus on civic engagement during year two; cohort 2 received a one-year intervention with the Live Well curriculum only. Additional adjustments to improve community engagement included hiring more project staff from each community, employing familiar and trusted community members for street outreach, and modifying media-PSAs to highlight personal experience about obesity in America. To better accommodate the busy schedules of many participants, we also held smaller nighttime measurement sessions at local and familiar community settings where linguistic and culturally competent staff provided childcare services, and culturally-appropriate food was offered.

Evaluation

Recruitment evaluation was primarily led by members of the Recruitment Advisory Subcommittee with active input from all steering committee members. Dyads that were contacted, eligible, and enrolled were tracked in a database. Project coordinators used participant screening questionnaires to inquire how participants learned about Live Well during their initial eligibility assessment, and participant eligibility and enrollment were systematically tracked and recorded throughout recruitment. These data were subsequently stratified by ethnic group and recruitment activity to inform our ongoing process evaluation. Such findings informed our incorporation of ad hoc adjustments to the recruitment plan which helped to ensure engagement efforts were appropriately tailored to each ethnic group.

Results

The flow of participants throughout recruitment is summarized in Figure 2. Overall, recruitment spanned 1 year and 9 months (overall, September 2009-June 2011; cohort 1, September 2009 – June 2010). Despite low initial enrollment, recruitment targets (400 dyads) were ultimately reached and exceeded: 1,065 dyads were contacted by project staff, 610 were eligible, and 406 enrolled (38% of contacts). During cohort 1 recruitment, which preceded the second cohort, only 145 were enrolled, which warranted expanding the recruitment strategy. The relatively smaller sample of Haitian dyads was attributed to the occurrence of a massive earthquake in Haiti (2010). Recruitment for cohort 2 resulted in 261 more participants. Ultimately, we met our recruitment goal for enrollment in Live Well (38% of contacts). A description of the demographic characteristics of the recruited study sample is available from the corresponding author upon request.

Fig 2.

The participant flow diagram, stratified by cohort and ethnicity. More than a thousand dyads were contacted throughout recruitment for Live Well (September 2009-June 2011).

The number and percent of dyads that were contacted, eligible, and enrolled from each activity are presented for each ethnic group and the total sample (Table 2). Community organization partnerships were the most successful recruitment activity for the total sample, generating the highest percent of contacted (36%) and enrolled dyads (34%). Posters and flyers were the next most successful at generating contacts (19%); however, only a small percentage were eligible which resulted in low enrollment (11%). The opposite trends were observed for participant referrals, which accounted for only 15% of all contacts and 25% of enrolled dyads. Although fewer dyads learned about Live Well from this activity, a larger percentage were eligible and willing to participate. Parsing these data by ethnic group revealed several differences.

Table 2.

The number and percent of Brazilian, Latino, Haitian, and total mother-child dyads from the Greater Boston, MA area that were contacted, eligible and enrolled from each recruitment activity employed during recruitment for Live Well (September 2009-June 2011)

	Total		Brazilian		Latino		Haitian
	N	%b	N	%b	N	%b	N	%b
All Activitiesa
Contacted	1065		391		369		305
Eligible	610		249		201		160
Enrolled	406		152		119		135
Posters/Flyers
Contacted	202	19.0	79	20.2	88	23.8	35	11.5
Eligible	77	12.6	35	14.1	29	14.4	13	8.1
Enrolled	45	11.1	13	8.6	20	16.8	12	8.9
Media-PSAs
Contacted	167	15.7	25	6.4	41	11.1	101	33.1
Eligible	95	15.6	17	6.8	17	8.5	61	38.1
Enrolled	76	18.7	8	5.3	9	7.6	59	43.7
Churches
Contacted	153	14.4	68	17.4	30	8.1	55	18.0
Eligible	73	12.0	37	14.9	18	9.0	18	11.3
Enrolled	45	11.1	31	20.4	3	2.5	11	8.1
Participant Referrals
Contacted	164	15.4	73	18.7	61	16.5	30	9.8
Eligible	133	21.8	58	23.3	47	23.4	28	17.5
Enrolled	102	25.1	42	27.6	34	28.6	26	19.3
Community Organizations
Contacted	379	35.6	146	37.3	149	40.4	84	27.5
Eligible	232	38.0	102	41.0	90	44.8	40	25.0
Enrolled	138	34.0	58	38.2	53	44.5	27	20.0

Note. PSA = Public Service Announcements.

^aThe number of dyads that were contacted, eligible, and enrolled for each ethnic group and the total sample, respectively, as a result of all recruitment activities

^bPercentages represent the contribution of each activity to the overall sample and were calculated by dividing the (N) dyads contacted, eligible, or enrolled per activity, by the total (N) dyads contacted, eligible, or enrolled, respectively, per ethnic group

Brazilian Recruitment

Consistent with overall findings, community organization partnerships and posters and flyers generated the highest percent of contacts (27% & 20%, respectively), although posters and flyers contributed minimally to enrollment (9%). Community organization partnerships and participant referrals were the most successful enrollment activities (38% & 28%, respectively), followed by churches, which were uniquely successful among Brazilian dyads (20%): although they contributed the least to the total sample, they were among the most successful activities for Brazilian enrollment (20%).

Latino Recruitment

Similar patterns in activity success were observed in Latino and Brazilian communities: community organization partnerships and posters and flyers were the most successful at generating contacts (40% & 24%, respectively), while community organization partnerships and participant referrals contributed most to enrollment (45% & 29%, respectively). Key differences were the success of posters and flyers for Latino enrollment (17%) and the relative lack of success from churches (3%).

Haitian Recruitment

Haitian dyads varied considerably from the other ethnic groups. Media-PSAs were the most successful enrollment activity, despite marginal contributions overall (Haitian: 44%; Latino: 8%; Brazilian: 5%), and the opposite was observed for posters and flyers, which generated fewer Haitian contacts, comparatively (Haitian: 12%; Latino: 24%: Brazilian: 20%). Community organization partnerships and participant referrals were likewise successful enrollment activities (20% & 19%, respectively); although less success was gleaned from community partnerships compared with the other ethnic groups (Haitian: 20%; Latino: 45%; Brazilian: 38%). Furthermore, a higher enrollment rate (enrolled/contacted dyads) was noted from participant referrals: although only 9% of contacts were recruited through referrals, 86% (26/30 dyads) were eligible and opted to participate.

Discussion

Like previous researchers (Corsino et al., 2012; George et al., 2014; Kaufman-Shriqui et al., 2013; Martin, Swidler et al., 2011), the Live Well team encountered multiple obstacles and cultural barriers as we sought to recruit our sample. However, by following the CBPR framework, employing culturally-adapted recruitment activities, and building trust through community partnerships, we exceeded our revised recruitment targets. Furthermore, our ability to revise our recruitment strategy, introduce a second cohort, and improve cultural specificity through ad hoc adjustments such as our revised media-PSAs and inclusion of nighttime measurement sessions, held in local community settings, which provided culturally-appropriate foods, facilitated our success and should be a consideration of future researchers. Our unique opportunity to recruit three diverse immigrant groups offers valuable insight and lays the groundwork for additional research to understand recruitment within other hard-to-reach populations.

Though all participants in Live Well were new immigrants in the US, several cultural differences contributed to the varied responses to each recruitment activity. Compared to Brazilians and Latinos, Haitians had more recently immigrated to the US, were less assimilated (average 3.9 years of residency compared to 7.7 and 7.3 years for Brazilians and Latinos, respectively) (Tovar, Boulos et al., 2014), and were more marginalized given their relegation as a “triple minority group” (immigrant, primarily Black, and speaking a unique language; Belizaire & Fuertes, 2011). This cultural isolation and lack of integration may explain their heightened response to media-PSAs and participant referrals, as their trust appeared more confined within their community – in close-knit, social networks or ethnic radio stations – than in broader organizations. Furthermore, their low literacy level may have encouraged a heavier reliance on audio communications than on posters or flyers.

On the other hand, Brazilian and Latino communities responded more favorably to community organizations than participant referrals, with little success from media-PSAs. Given their longer residency in the US, it is possible that these more assimilated communities expanded their networks beyond direct social spheres. It seemed they instilled trust in local organizations, demonstrating a more fluid social network and greater integration within the broader community. Despite these similarities, Brazilians and Latinos differed in their response to church-based recruitment. Initially these activities were problematic for all due to major time and resource investment required by Live Well project staff; however, outreach was successful for Brazilian dyads during the latter phase of recruitment, suggesting connections with Brazilian clergy may be beneficial, albeit resource-intensive. These variations highlight the need for a multi-strategy approach when recruiting an ethnically diverse sample and emphasize the importance of community understanding.

Much of our success resided in strategic partnerships with community stakeholders and lay-leaders. Such collaborations enabled us to learn more about the culture, traditions, and experiences of each immigrant community, and to better reach potential participants. For example, new immigrants place a high degree of trust in ESL Institutions given shared goals of learning English; and by aligning ourselves with these institutions, we cultivated community trust. Collaborations with the LIPS program yielded a similar result and further facilitated community connections and understanding. Accordingly, we postulated that, although individuals learned about Live Well through different avenues, because we spent time learning about and investing in their trusted stakeholders and sources, we were successful at recruiting and enrolling participants.

Limitations

Several limitations must be considered when interpreting our findings. Although we recruited a diverse sample, our particular focus on new immigrant women and children from Brazil, Latin America, and Haiti may limit the generalizability of our work to other underserved communities. However, we hope future researchers will be able to use the cultural characteristics and the CBPR processes of our intervention to inform their efforts. Our inability to assess why certain community members who were exposed to our messaging chose not to call for an eligibility assessment was also a limitation. Future researchers might consider additional community-wide assessments to better understand the factors that contribute to decisions to not participate. This understanding may facilitate additional community engagement and result in larger and more representative study samples. Lastly, our prior connections to the immigrant community in Somerville, MA served as a strong foundation upon which we built community partnerships and cultivated understanding. Developing these connections de novo, however, will require vast investment of time and resources which may be difficult to cultivate given limited funding availability. Future researchers working within new immigrant communities should also consider the historical and cultural differences of the sending countries when designing and assessing outcomes. Collaborating with community members through a CBPR approach can provide valuable insight into how certain practices would be received within the context of each community. These considerations were integral to the development of the Live Well intervention (Tovar et al., 2013) and will continue to inform subsequent articles.

Conclusions

Future researchers must develop cultural awareness and understand community social networks to guide strategic partnerships and connect with community members. Our findings suggest recruitment in newer, less-integrated immigrant communities should follow a grassroots approach, directly engaging lay community leaders and partnering with ethnic institutions. Alternatively, recruitment in more assimilated communities should involve partnerships with broader organizations that connect the research and intervention communities on multiple levels. Though building this understanding requires patience and persistence, we believe researchers considering these characteristics will have an advantage over those employing generalized recruitment activities which do not account for cultural or social network differences across communities.

Recruiting minority populations is a sensitive process; and building trust, becoming part of the community, and tailoring recruitment strategies, are necessary practices. Relationships must be cultivated over time, however, and forming these foundations requires significant investment in time and resources from both the intervention and research communities. Future researchers should also take into account the cost-effectiveness of each recruitment activity when evaluating their contribution to overall recruitment goals. Given the current funding climate, such research is crucial to reducing the burden of chronic disease within underserved minority populations.