Recruitment strategies for predominantly low-income, multi-racial/ethnic children and parents to 3-year community-based intervention trials: Childhood Obesity Prevention and Treatment Research (COPTR) Consortium

The recruitment of study participants into community-based interventions is the foundation for a successful trial. Contemporary community-based interventions often focus on economically disadvantaged, racially/ethnically diverse populations [1]. Recruitment of these populations requires an understanding of the challenges faced by potential participants, such as feelings of mistrust of medical care and medical research [1‐4], family stressors including high mobility and financial distress [5], lack of a working phone to communicate with research staff [5], inflexible work hours or lack of time due to working several jobs [5], lack of childcare or transportation [6], low literacy skills [5], and limited understanding of research purposes and processes [1]. When intervention studies require a long-term commitment (e.g., more than a year), the recruitment of participants is more challenging than for a short-term study [7].

Childhood obesity has been associated with hypertension [8], cardiometabolic abnormalities [8], cardiovascular dysfunction [9], sleep disorders [8], psychological difficulties [10, 11], and adult obesity [12]. The total direct medical costs of childhood obesity have been estimated at $14.1 billion annually [13]. Childhood obesity is an important public health problem, and is up to three times more prevalent among low-income, racial/ethnic minority children [14]. Despite its importance, our recent systematic review of published pediatric obesity trials registered in the Clinical Trials Registration database identified little information on recruitment of low-income minority families into long-term obesity prevention or treatment trials [7]. One reason may be that parents may not perceive their child to be at risk of obesity [15]. Health-care providers are also unlikely to diagnose obesity in children [16, 17]. Even when diagnosed, parents, especially low-income parents, may not prioritize obesity as an important health problem for their children, given a host of more immediate stressors and challenges they may face [18, 19].

The Childhood Obesity Prevention and Treatment Research (COPTR) consortium is funded by the National Institutes of Health (NIH) to develop and evaluate four distinct multi-level (individual, family, school, clinic, and community) and multi-component community-based child obesity prevention and treatment trials, including individualized recruitment protocols. The COPTR consortium collaborated to share, refine, and document recruitment strategies systematically and successfully enrolled 1745 low-income racially/ethnically diverse parent–child dyads. A retrospective reflection of the COPTR recruitment experience offers a unique opportunity to identify strategies that resulted in successful recruitment in this hard-to-enroll population.

This paper describes the strategies used to recruit and enroll predominately low-income racial/ethnic minority parents and children into the four community-based obesity prevention and treatment trials run by the COPTR consortium. It may provide insight for future investigators seeking to enroll similar populations.

The COPTR consortium ran for four independent community-based randomized controlled trials and a research coordinating unit (RCU) funded by the NIH [20]. The study designs have been previously described [20‐24]. The two prevention trials targeted preschool children and were conducted by the University of Minnesota [21] and Vanderbilt University [22], respectively. The two treatment trials targeted pre-adolescents and were conducted by Case Western Reserve University (CWRU) [23] and Stanford University [24], respectively. All four trials were 3-year multi-level interventions and enrolled low-income racial/ethnic minority populations. The four trials had their own recruitment methods, interventions, and measurement protocols and agreed a priori to collect an extensive array of common measures [20]. To reach a low-income population, one trial (Minnesota) used annual family income below $65,000 [21] as an eligibility criterion, while the other trials recruited participants in low-income neighborhoods and did not apply an income criterion [22‐24]. By design and based on the power calculations for each trial, recruitment goals varied for the prevention (n = 500 and 600) and the treatment (n = 240 and 360) trials [21‐24].

The RCU was located at the University of North Carolina at Chapel Hill. The RCU supported the development of data collection protocols for common measures and coordinated the collaborative working committees, subcommittees, and working groups, and interactions with the consortium’s NIH-appointed independent data and safety monitoring board. The consortium-wide Recruitment, Consent, Retention and Adverse Events (RCRAE) subcommittee consisted of a staff member responsible for recruitment and an investigator from each trial. The RCRAE subcommittee was responsible for the development of a cross-trial standardized method to report recruitment progress over the consortium’s 18-month recruitment phase (24 months for Vanderbilt). Each trial reported monthly recruitment and enrollment progress to the RCRAE subcommittee using four predefined phases: screening, eligibility, consent, and randomization. Reasons for discontinuation during the recruitment process were documented. Recruitment progress was reviewed monthly by the RCRAE subcommittee. The monthly RCRAE meetings provided opportunities for the committee members to discuss perceived effective strategies, lessons learned, and solutions to recruitment barriers. The site representatives further discussed those solutions within their sites and decided whether and how to implement those potential solutions. The RCU prepared and submitted quarterly progress reports to the NIH and biannual progress reports to the data and safety monitoring board for review.

Two and a half years after the randomization of participants across all four trials was completed, the RCU created a list of the key recruitment strategies frequently documented in the literature based on our recent systematic review [7] and categorized them into seven stages primarily based on temporal sequence: (1) planning, (2) recruitment staffing, (3) community outreach and participant identification, (4) eligibility screening, (5) consent and assent, (6) measurement, incentives, and eligibility, and (7) enrollment and randomization (Fig. 1). Based on these seven stages of recruitment, the RCU developed a survey with open-ended questions that asked each trial group what recruitment strategies they had used (Additional file 1: Table S1). This retrospective survey also included open-ended questions to query trial groups about their perceived top three recruitment strategies and barriers. In addition, each trial group identified barriers they had experienced from a list developed by the RCU based on a literature review. The survey questions were sent to the staff member in each trial who was responsible for recruitment and who had served on the RCRAE subcommittee. Each site independently completed the survey questions with input from the recruitment staff member and the investigator. One completed survey from each trial was sent to the RCU and the RCU manually coded and summarized the responses.

The four COPTR trials successfully recruited their targeted number of children within their projected recruitment phases. All four trials retrospectively reported using similar and different recruitment strategies to achieve their recruitment goal (Table 1).

The COPTR consortium successfully recruited 1745 low-income parent-child dyads into two prevention and two treatment trials for child obesity. Recruitment strategies from the four trials were categorized into seven stages based on a comprehensive review of the recruitment literature. Although recruitment strategies used in trials and cohort studies have been previously published, those work has not defined and classified the recruitment process into specific stages. These stages provide a step-by-step framework that may help investigators better understand the demands on recruitment resources and allow them to plan and implement strategies for successful recruitment more efficiently.

Incentives are an important, yet less discussed, element of successful participant recruitment and may be particularly relevant when recruiting lower-income study participants [7]. Incentives that are offered for data collection in child obesity prevention and treatment studies include a variety of forms such as gifts, food, recipe books, exercise equipment, grocery gift cards, and cash [7]. Among these, a monetary incentive is one of the most commonly used. Monetary incentives can enhance data collection response rates, but must be balanced with the cost per participant enrolled [25] and the potential for coercion. Of the 43 studies from our systematic review [7], we were able to identify only one child obesity intervention study conducted in an underserved population that reported the use of monetary incentives. In a trial to prevent weight gain in Hispanic children aged 2–6 years, a total value of approximately $30 in incentives (e.g., a cutting board, a kitchen timer, and a gift card for a local supermarket) were distributed to participants after collecting all the baseline data [26]. In the present four COPTR consortium trials, gift cards totaling $40, $50, or $110 were perceived to be effective incentives to compensate families for the time and effort required to complete baseline recruitment and to produce the observed high measurement completion rates among participants in geographically and racially/ethnically diverse low-income families. Whilst, different incentive schedules could have different effects on participant recruitment, the four trials in our study did apply different incentive schedules and all trials successfully recruited participants.

A unique feature of the COPTR consortium is that it runs both prevention and treatment trials. This enabled a rough comparison of the recruitment of participants into childhood obesity prevention versus treatment trials. The number of screened children per randomized child in the two prevention trials was larger than that in the two treatment trials. This is consistent with our systematic review [7]. There are several speculative explanations. One is that prevention trials included children in the normal BMI percentile range; therefore, their parents may not be as concerned about childhood obesity as the parents of overweight or obese children. Furthermore, children with a BMI lower than the 50th percentile were not eligible for the COPTR prevention trials, narrowing the eligible pool. It can be difficult for recruitment staff to exclude those children accurately without measuring them, potentially increasing the ratio of children screened to randomized.

Collecting all baseline data before randomization has been widely used in the implementation of randomized controlled trials to enhance data completion and retention rates [27‐29]. The COPTR consortium required a minimum set of baseline data to be collected for a family before it was eligible for enrollment and randomization. Collecting baseline data helps participants understand the burden of data collection and serves as the entry point into an intervention study. Those who do not provide the minimum data at baseline may be more likely to provide insufficient data or drop out during the follow-up. Having many missing values may introduce bias and erode the statistical power. Although the minimum baseline data requirement could reduce the generalizability by excluding participants who showed less adherence to the recruitment process, the COPTR consortium investigators decided that balancing these potential threats favored requiring a minimum set of baseline data. In an effort to reduce loss to follow-up, recruitment staff clearly communicated study demands and researchers’ expectations of participation at the screening and consent stages to ensure that participants accurately understood the benefits, potential risks, and time commitments of the interventions and assessments. Some investigators have suggested that before potential participants are enrolled into a study, their willingness to participate should be evaluated. Those who seem hesitant to participate or cooperate should not be enrolled [30].

Successful recruitment into the four trials of the COPTR consortium may also be partially attributed to its organization and structure. The regular conference calls within the RCRAE subcommittee and the biannual review of progress with the data and safety monitoring board served as a platform for each trial group to share their progress and the challenges of recruitment. These not only allowed each trial group to have access to outside expertise on recruitment but also created a timely monitoring and feedback system. However, these interactions may have reduced the independence of each trial group’s response to the perceived effective strategies and barriers.

Several key strategies might have contributed to our success in recruiting a relatively large number of low-income racial/ethnical minority families from various settings, including clinics, community settings, and schools. We learned that careful planning throughout the recruitment process and a sufficient budget for recruitment are critical for recruitment success. Having a large enough budget provides a financial basis for providing incentives, such as monetary incentives, transportation reimbursement, and childcare to compensate for the time and effort of participating in assessments. Personnel-related strategies include using trusting community partnerships and hiring recruitment staff from the community who are personality appropriate and willing to work flexible hours. Investigators should develop clear and detailed protocols for contacting participants, actively approach potential participants, recruit at times and locations that are convenient for participants, evaluate whether participants understand the activities and expectations of the study, assess participation motivation, and enroll only those who plan to continue participating over the study period and who have provided sufficient baseline data. It also proved helpful to use a tracking database and to promise to return clinical assessment results to participants.

The recruitment of low-income racial/ethnic minority families from wide-ranging geographic areas and diverse settings into obesity prevention and treatment trials can be accomplished using comprehensive strategies that include: (1) community partnership support, (2) hiring recruitment staff who are culturally sensitive, personality appropriate, and willing to work flexible hours, and (3) repeated contacts with potential study participants.