128. Maximizing Participant Recruitment Through Community Partner Engagement: Lessons from a Multi-State Randomized Controlled Trial

Evaluations of behavioral health interventions typically require a rigorous randomized controlled trial (RCT) to establish whether they are effective. Participants must be drawn from relevant populations, involving partnerships with community-based organizations (CBOs). However, the strict control needed for RCTs may challenge many aspects of CBO structure and operations, jeopardizing these partnerships and, ultimately, whether the evaluation can be conducted at all. Here we describe our evolving partnerships with CBOs during a three-year RCT testing the effectiveness of a behavioral intervention. After recruitment challenges during the first year, we drew on an implementation science perspective to identify structural barriers for our partners’ involvement, iteratively testing engagement strategies to improve collaborations and participant recruitment. The RCT tested the effectiveness of a sexual-health prevention intervention, Your Move, against an attention-control nutrition intervention, Eat Smart, between July 2016 - June 2019 (ClinicalTrials.gov #NCT02993185). Both interventions are medically accurate, group-implemented educational interventions designed to improve adolescent girls’ decision making. They are parallel in structure, format and time commitment. Groups of participants were randomly assigned to either Your Move or Eat Smart after informed consent. We partnered with local non-profit and educational organizations to provide facilitators to deliver the interventions and coordinate with youth-serving CBOs to identify and recruit groups of female adolescents for participation. Through our partner organizations we drew participants from CBOs across six states with varying missions including charter schools, after-school programs, summer programs, and faith-based organizations . We tested engagement strategies across five domains: (1) contract and payment structure with partner organizations, (2) contract and payment structure for CBO recruitment partners, (3) structure of payments for individual participants, (4) procedures for data collection, and (5) procedures for intervention delivery. Outcome measures for engagement strategy success were recruitment of participants, cohort groups, and CBO partnerships across three recruitment years. In year 1 we partnered with 11 CBOs to recruit 12 cohorts with 99 individual participants. Between years 1 and 2 we made changes in all five domains of engagement identified above. In year 2 we partnered with 31 CBOs to recruit 59 cohorts with 431 individual participants. The only substantive change made between years 2 and 3 was the further reduction of implementation partners (from 4 to 3) to keep the most productive. In year 3 we partnered with 22 CBOs to recruit 33 cohorts with 271 individual participants. The changes we made to our engagement strategies appeared to improve recruitment of partners and participants. We invested more resources in fewer partners, giving them more autonomy and accountability. This arrangement created a more trust-based relationship: that we would pay our partners in intervals rather than by milestones, and that they would provide their promised recruitment within the contract period. This meant giving up some bureaucratic control of the services agreement. By using an implementation science approach to better understand our community partners’ barriers to engagement with our university’s preferred contractual model, we were able to create a model better suited to their organizational structure and greatly improve recruitment and partnerships.