Recruiting and Retaining Parents in Behavioral Intervention Trials: Strategies to Consider

Abstract

Objective:

Recruitment and retention are paramount to the success of randomized controlled trials (RCTs); however, strategies and challenges to optimize recruitment and retention are often omitted from outcomes papers. The current manuscript presents strategies used to recruit and retain over 97% parents of young children newly diagnosed with type 1 diabetes for over 15-months post-randomization enrolled in First STEPS, a behavioral, two-site RCT.

Method:

Participants included 157 primary caregivers of young children newly diagnosed with type 1 diabetes. Recruitment and retention strategies are described and include collaboration with medical teams, careful selection and training of study staff, inclusion of a behavioral run-in prior to randomization, financial incentives, creation of a study identity using retention items, obtainment of feedback from community stakeholders, and minimization of participant burden.

Results:

Use of recruitment and retention strategies resulted in enrollment of 58% of eligible and reached families, with retention of the enrolled sample above 97% for over 15 months. Participants reported high acceptability of and satisfaction with specific recruitment and retention strategies.

Conclusions:

The strategies used to recruit and retain caregivers of young children newly diagnosed with a chronic illness were feasible to implement within multidisciplinary diabetes clinics and may apply to other pediatric populations. Future research may benefit from a focus on strategies to engage more diverse samples.

Behavioral randomized controlled trials (RCTs) for caregivers of children with chronic illnesses require a representative sample of participants to evaluate intervention efficacy, yet various barriers often impede successful recruitment and retention. Barriers described in the literature include perceived disconnection of the research team from the medical care team, distrust of research (Barakat et al., 2013; Greenberg et al., 2018; Heinrichs et al., 2005; Lebensburger et al., 2013), burdens of research participation (e.g., travel, cost, wait time), and potentially impersonal contacts with the research team (Greenberg et al., 2018; Heinrichs et al., 2005). Parents also have reported a preference for trial introduction and endorsement by their child’s physician, with whom they have already established a trusting relationship (Greenberg et al., 2018).

To evaluate the efficacy of behavioral interventions, methodologically rigorous RCTs are often considered to be the “gold standard.” RCTs allow for an unbiased analysis of the intervention, as potential confounds are evenly distributed between intervention and control groups, allowing for observed differences in outcomes to be attributed to the intervention (McCoy, 2017). Given the sizeable time and effort often required of participants in behavioral RCTs, attrition during the intervention or follow-up data collection is common. Though some investigators may be inclined to exclude participants who do not receive the full intervention, this practice introduces substantial bias into data analyses. Intent-to-treat analyses, in which all randomized participants’ data are analyzed, are typically recommended for an unbiased test of intervention efficacy (Ten Have et al., 2008). Therefore, maintaining high participant retention is vital to producing generalizable results.

Parent stress is heightened during the diagnosis of a child’s chronic illness, which may interfere with trial recruitment, when conducted during an inpatient hospital stay (Dunn et al., 2012; Landolt et al., 2002). Given the stress and burden of disease management that parents of youth with chronic illnesses commonly experience, successful recruitment and retention of parents can be difficult in pediatric behavioral RCTs (Ghai et al., 2014; Karlson & Rapoff, 2007; Stehl et al., 2009) and rates are often low (Tercyak et al., 2006; Wysocki et al., 2006). For example, only 23% of eligible participants enrolled in a clinical trial of a psychological intervention for parents of children newly diagnosed with cancer (Stehl et al., 2009). In their review of 40 clinical RCTs for parents and youth living with chronic illness, Karlson and Rapoff (2007) documented attrition rates up to 54% across the 40 trials. There are few reports of how researchers recruit and retain participants in pediatric behavioral clinical research (Byrd-Bredbenner et al., 2017).

The current manuscript provides a report of recruitment/retention strategies and associated data from our behavioral RCT, which had a higher retention rate than those of previously reported pediatric type 1 diabetes (T1D) behavioral trials (Herbert et al., 2016). Strategies used in the current trial allowed for successful recruitment of 157 families and retention of over 97% of the sample 15 months after randomization. Compared to those previously reported, discussed strategies are wider in scope (e.g., study design, study launch, throughout trial) and may have facilitated inclusion of a more diverse sample.

Method

Overview of First STEPS Intervention

The First STEPS study was an RCT comparing the efficacy of a behavioral intervention designed to support parents of young children newly diagnosed with T1D compared to usual diabetes care. Randomization occurred on a 3:1 (intervention to usual care) basis. Data were collected at baseline (within 8 weeks of the child’s T1D diagnosis), 3 and 6 months later (during intervention), and following intervention completion (9- and 15- month post-randomization).

The intervention used a stepped care design that provided participants with up to 3 intensity levels, or “steps,” of intervention to provide the least intensive treatment needed for each family. Child glycemic control (Hemoglobin A1c) and parent depressive symptoms (Center for Epidemiologic Studies – Depression Scale (CES-D); Radloff, 1977) were the primary study outcomes and were used to determine intervention step advancement. More specifically, A1c and CES-D were assessed at two clinic visits, corresponding with approximately 3- and 6-months post-randomization; elevated values (i.e., ≥16 CES-D; ≥8.0% A1c) on either measure initiated step movement, while participants with non-elevated values on both measures remained at their current step. If no clinic visit fell within data collection timeframe, CES-D was collected by phone and step advancement was determined by CES-D score only.

Intervention components are detailed in Figure 1 and elsewhere (Tully et al., 2021; Tully et al., 2017). Briefly, all participants in the intervention condition began in Step 1, which included communication with trained parent coaches. Step 2 included five telephone-delivered cognitive behavioral sessions led by a Master’s level counselor. Step 3 included one-time consultations with a licensed psychologist and personalized diabetes data review with a certified diabetes care and education specialist (CDCES). Intervention components were delivered via telephone (Steps 1 and 2) or in-person/telehealth platform (Step 3). In addition, intervention participants also had access to a study website with diabetes and caregiver support resources.

Figure 1.

First STEPS intervention

Overview of Recruitment and Study Participation

Study recruitment occurred at two children’s hospitals in Washington, DC and Houston, TX. IRB approval was obtained at both sites for all procedures. When possible, primary diabetes caregivers for children ages 1–6 were approached while their child was admitted to the hospital for initial T1D diagnosis. In-hospital introduction of the study consisted of research study staff conducting a brief “meet and greet” (i.e., <10 minutes) with parents to establish rapport, describe the study, share a study flier with team contact information, and arrange a time to follow-up by phone after discharge. All potential participants were also mailed a recruitment letter with an additional copy of the study flier.

Study staff called families approximately one week after diagnosis to introduce the study, determine eligibility, and obtain verbal consent for participation. All primary caregivers (including parents, legal guardians, or other caregivers) were eligible to participate. Exclusion criteria included serious mental illness or developmental disability in the caregiver or child; other life-threatening child illness (e.g., cancer, cystic fibrosis); and lack of English fluency (speaking, reading, and/or understanding).

After obtaining verbal consent, participants provided written consent and completed baseline questionnaires through a Research Electronic Data Capture (REDCap) link (Harris et al., 2009); completion was required within eight weeks of diagnosis. At the child’s clinic visit following completion of baseline questionnaires, participants were seen for an in-person orientation session, where they were randomized and informed of group assignment. Details of study participation are provided in Figure 2.

Figure 2.

Timeline of study involvement

Recruitment and Retention Strategies

Strategies for recruitment and retention were identified from literature review (Byrd-Bredbenner et al., 2017; Greenberg et al., 2018; Sullivan-Bolyai et al., 2007) and various RCT recruitment optimization presentations (Kiernan, 2015; Leahey, 2015; Streisand, 2017). We used the following strategies (Table 1).

Table 1.

Recruitment and retention strategies utilized in the First STEPS Project

Guiding Strategy	Specific Recommendations
Collaboration with medical teams	Subcontract clinic staff time Provider assistance with identification of eligible families to recruit Provider endorsement of trial to families Minimize interruptions to clinic workflow/provider burden Provide small gifts of appreciation to clinic providers
Careful selection and training/supervising of study team	Ability to make long-term commitment to trial Experience with clinical research/clinical trials Organized, detail-oriented, flexible, interpersonally effective, warm, friendly Active listening and positive communication with participants
Inclusion of a “behavioral run-in” (Orientation Session)	Remove participants with lower ability or commitment to participate from trial before randomization to minimize attrition
Tiering financial incentives	Increasing incentive schedule across trial
Fostering a study identity and personalizing retention items	Create study name Design study logo Provide customized small gifts/items with study logo at each clinic visit Provision of developmentally appropriate small toys from first contact Mail birthday/holiday/Mother’s Day/Father’s Day cards with study logo
Obtaining stakeholder input	Present materials and obtain feedback from families on pre-existing parent advisory board
Minimizing participant burden	Flexible scheduling and communication time/method Timeliness/minimize wait time Reimburse cost of parking Provision of childcare and child-friendly activities Employ multiple modalities (e.g., texting, phone calls, mailings) and times of day (e.g., evenings, weekends) Friendly, brief reminders about study meetings Confirm contact information multiple times throughout trial and identify secondary and emergency contacts Business cards with single phone number and email address for trial

Collaboration with Medical Teams.

To successfully conduct the trial in the context of large multi-disciplinary clinics, we worked closely with hospital staff, diabetes clinic medical staff, diabetes providers, and clinic administrators. The Principal Investigator (PI) at each site met with the diabetes leadership team, faculty, and staff to present the study aims, procedures, and recruitment strategy prior to trial initiation and 1–2 times yearly throughout the trial to ensure familiarity with the project. At least 2 diabetes health care providers from each site served as grant-funded Co-Investigators, who attended monthly recruitment/retention meetings.

At diagnosis, nursing staff and physicians attend to the child’s daily medical needs and diabetes-team psychologists and social workers provide brief psychosocial support; their support allowed our research staff to identify potential participants and speak with families on the inpatient unit for recruitment purposes (Sullivan-Boylai et al., 2007). Medical teams provided an informational study flier to all newly diagnosed families in the study’s target age range and identified potentially eligible families for research staff to approach on the inpatient unit, allowing research staff to schedule the “meet and greet” to introduce and discuss the trial prior to discharge; during the “meet and greet,” research staff made it clear to families that they were not medical team members.

The study team also met with participants during diabetes outpatient visits. To maintain ongoing, positive relationships with the diabetes care teams, which helped build trust with participants, research staff deferred to medical staff and providers before speaking with a family to avoid disruptions to clinic flow. Deferring to medical staff maintained the medical team’s confidence in the study team’s professionalism and their desire to continue endorsing the trial with potential participants (Swallow et al., 2014). The research team also made intentional efforts to express appreciation of medical staff and their support of the trial (e.g., giving thank-you notes, fruit trays and colorful hand sanitizers gifted around the holidays).

Carefully Selecting and Training/Supervising Study Team.

Recruitment following a child’s T1D diagnosis requires sensitivity and understanding of the significant challenges and stress that families may experience. We assembled an experienced, consistent research team that would make families comfortable during recruitment and facilitate engagement throughout the study. We prioritized a two-year minimum commitment when hiring research staff. We also prioritized hiring staff who had prior experience with pediatric clinical research or clinical trials, detail orientation, flexibility, interpersonal warmth, interest in T1D/chronic illness, ability to travel for data collection (i.e., access to a car, ride-share, etc.), and flexibility to work on weekends/evenings. Research staff were trained and supervised throughout the study period to use active listening and positive communication skills (e.g., using first names, remembering details about the family, validating families’ difficult experiences). Since trial initiation, there have been 7 bachelors’ or masters’ level study coordinators/research assistants across the two sites. The position is advertised as a 1–2 year position, and thus far, we experienced 0% staff turnover prior to the 1–2 year mark. Five have completed their commitments to the trial (3/5 worked on the study for 2–4 years and 2/5 for at least one year).

Inclusion of a “Behavioral Run-In” (Orientation Session).

Participants met with research staff for an in-person orientation session, as a “behavioral run-in” prior to randomization, and to enhance participant engagement with the study (Hilliard et al., 2017; Tsai et al., 2014). An in-person visit allowed us to screen for individuals with the ability and commitment to participate in the study before randomization. In-person orientation sessions commonly occurred in the diabetes outpatient clinic immediately before or after a medical appointment; of the 158 orientation sessions conducted, 93.6% were conducted at the initial scheduled date (148/158); only 6.4% (n=10) were re-scheduled (8 sessions were re-scheduled once, 2 sessions were re-scheduled twice) within the six-month window. In total, sessions lasted 30–45 minutes and were conducted by the same research coordinator, who contacted the family in the future for data collection needs. In the orientation session, participants provided written consent for study participation, as verbal and electronic documentation of consent had previously been obtained. In order to build rapport, parents shared their story of the child’s diagnosis and watched a short, publicly available patient education video about family adjustment after T1D diagnosis. Study staff then reviewed the data collection timeline, incentive structure, and gave participants study retention items. Finally, participants were randomized to the intervention or usual care condition.

Tiering Financial Incentives.

To encourage retention through 15-months post-randomization, the research team provided financial incentives to participants, following completion of questionnaires at each data collection point (Lagor et al., 2013). An increasing incentive schedule was used for completion of the full questionnaire battery, and a flat $10 rate was used for the brief check-in assessments (single questionnaire; see Table 2). Financial incentives were provided via gift card (participant choice between two major retailers) or ClinCard (electronically refillable debit card).

Table 2.

Incentive schedule and retention items

Timepoint	Retention Item
	Financial Incentive	Parent	Child
Recruitment meeting	-	-	Small stuffed animal
After verbal consent, before Orientation session	-	-	Diabetes-themed activity book with study logo
Baseline	$50	-	-
Orientation Session	-	Pen with logo Magnet with logo and intervention-specific information (intervention only)	Goody bag – Child’s backpack with coloring book and small toys
Check-in #1	$10	Magnetic chip clip with logo	Rubber ducky with logo
Check-in #2	$10	Notepad with logo	Small ball with logo
First Follow-Up	$75	Keychain flashlight with logo	Color-changing plastic cup with logo
Second Follow-Up	$100	Thank-you card	Activity coloring book with logo
Child’s Birthday	-	Birthday card with study logo signed by local study team	Diabetes-themed cookbook for young children, coloring poster, animal hat
Clinic visits not in data collection windows	-	-	Coloring sheet, crayons, stickers with study logo
Mother’s & Father’s Day	-	Card with study logo signed by local study team	-
First STEPS Newsletter	-	Mailed periodically with seasonal content designed for youth with T1D and study progress updates	-

Fostering a Study Identity and Personalizing Retention Items.

To promote participant recognition of and affiliation with the study, the research team developed a study identity (Figure 3). A short name that would be memorable and reflect key points about the study, First STEPS, was selected. “First” indicated the time period after diagnosis and “STEPS” was an acronym for Study of Type 1 in Early childhood and Parenting Support. A professional graphic designer developed a study logo, including a graphic icon, text of the study name, and color scheme, which we used on all communications (e.g., letterhead, business cards) and printed on retention items. Specifically, we selected retention items for parents and children that were developmentally appropriate, appealing (e.g., stress balls), and would be frequently used (e.g., cups) to increase logo recognition. Retention items were given to all families at clinic visits through 15-months post-randomization, regardless of group assignment or whether data collection was needed. Further, a small stuffed animal was given to the child at the “meet and greet,” and a backpack with diabetes books, art supplies, and toys was given at the orientation session. Greeting cards with study logo signed by all research team members were mailed on children’s birthdays, Mother’s and Father’s Day, other holidays, and at study completion (Figure 5). Lastly, a study newsletter was sent seasonally with updates about overall study progress. Table 2 summarizes the retention items, and Figures 4 and 5 are examples of branded retention items.

Figure 3.

Example retention item with study logo

Figure 5.

Example Mother’s Day card with study logo

Figure 4.

Study logo

Obtaining Stakeholder Input.

The research teams convened a parent advisory board at each site, who was consulted for input on an as-needed basis. Their feedback guided the selection of recruitment and retention methods including ranking potential retention items, providing feedback on study materials, and suggesting resources to include on the study website.

Minimizing Participant Burden.

Study team members prioritized flexible scheduling and timeliness when meeting with families. To sustain parent engagement, team members brought developmentally appropriate toys to each study visit to entertain the child and had a study volunteer occupy the child for longer meetings (e.g., orientation, Step 3). Parking costs for each visit were reimbursed.

To reduce burden associated with communication, study team members used various HIPAA-compliant methods of communication. Study staff communicated about upcoming study visits through emails and text messages, per participant preference. When phone calls were needed, team members kept communications brief, and longer calls were pre-scheduled to accommodate participant availability (e.g., work schedules). Additionally, study team members maintained a detailed log of all successful and unsuccessful communication attempts, noting the method of communication, date, time, and outcome of each attempt (e.g., participant reached/message left, participant preference for contact method). Of note, successful attempts often occurred on weekends or evenings. Participants confirmed their contact information (i.e., phone numbers, email addresses, mailing addresses) at each study visit. Additionally, when parents moved during the trial, team members attempted data collection however possible (e.g., follow-up questionnaires, obtaining releases of medical information to get A1c values). If participants were unable to be reached by phone or electronic or paper mailing, study team members met families in-person at a clinic visit or reached out to a secondary contact person, whose phone number participants provided at the orientation session. Lastly, to centralize participant communication at each site, one phone number and email address were used per site, which remained consistent throughout the study.

Modification of Strategies

We adjusted some recruitment strategies as needed throughout the trial. First, initial study design included a longer in-person recruitment meeting during the inpatient stay, with goals of obtaining parent consent and possible completion of baseline questionnaires. If unable to complete during the inpatient stay, parents were initially required to complete baseline questionnaires within four weeks of diagnosis. However, given the time-limited nature of the hospitalization at new T1D diagnoses, high time demands of learning diabetes education, and intensity of parent stress, we replaced this longer recruitment meeting with the quicker “meet and greet,” which was reconceptualized as an opportunity to build rapport and arrange for later recruitment. Further, given the difficulty a number of families experienced with completing baseline questionnaires within four weeks of diagnosis, the deadline was extended to eight weeks. Second, our early experiences with recruitment calls were lengthy and, thus, logistically challenging. To reduce this burden on families, giving families a flier during the “meet and greet” and mailing a paper letter with study details allowed parents to learn about the trial before the first recruitment call, which shortened phone calls. Finally, we initially emailed reminders about study visits and found that texts/calls were often more efficient.

Results

Quantitative Recruitment Data

Between May 2016 and July 2019, 398 potential participants were identified (i.e., T1D diagnoses among children aged 1–6 years) across the two study sites. Of these, 34 families were excluded from electronic medical record review due to non-fluency in English (n=18), other significant child medical illness (n=6), plans to receive care elsewhere (n=4), inability to reach caregiver before recruitment window closed (n=2), child no longer within age range (n=2), inability to identify primary caregiver (n=1), and parent younger than 21 (n=1). After preliminary eligibility screening from the medical record, study staff approached 364 families, either in person during inpatient hospital stays for T1D diagnosis with a mailed recruitment letter (n=190), by mailed recruitment letter only (n=169), or in-person only (n=5). Of these contacts, 66 were not enrolled, most often due to inability to reach family (n=23), inability to complete phone screen before recruitment window closed (n=18), ineligibility (n=14), and lack of interest (n=11).

Following these initial contacts, study staff conducted eligibility screenings with 298 families by phone. Of these, an additional 52 families were not enrolled, primarily due to the family’s self-reported inability to commit to study demands (n=27), and an additional 29 families were not enrolled due to ineligibility, most often due to planned care at non-study sites (n=14), or non-English fluency (n=11). After phone eligibility screening, 217 families met inclusion criteria and consented to participate; of whom, 47 families were lost to follow up before randomization, primarily because they failed to complete baseline questionnaires within 8 weeks post diagnosis (n=36), as well as loss of interest (n=8) or change in eligibility (n=3). In total, 170 families completed the required baseline questionnaires. Of these 170 families, 12 families were lost to follow up, before randomization, most often because they did not attend the orientation session (n=10), as well as change in eligibility (n=1) or inability to be reached (n=1).

In total, 158 parents (58% of those reached and eligible) were randomized, 115 to the intervention and 42 to usual care. Of these parents, 54% (n=86) had been approached for a “meet and greet” during their child’s inpatient stay. One participant (and their associated data) was removed mid-study due to a change in eligibility. Therefore, our final sample included 157 parents of young children newly diagnosed with T1D. The study CONSORT table (Figure 6) illustrates the recruitment flow in detail.

Figure 6.

CONSORT Flowchart of First STEPS Recruitment

Participants

Mean age of the 157 participants was 34.88 years, SD=7.01. Most parents were female (n=144, 91.7%) and married (n=117, 76.5%). Primary caregivers self-identified, from a fixed set of categories, as: White, non-Hispanic (n=97; 62.2%), Black/African American, non-Hispanic (n=23; 14.7%), Hispanic/Latino (n=19; 12.2%), Asian/Asian American (n=12, 7.7%), multi-racial (n=4; 2.6%), and American Indian/Alaskan Native (n=1; <1%). Of the sample, 27 (20.0%) had annual household incomes of ≤$40,000, and 44 (28.1%) children had public (n=43) or no insurance (n=1).

About half of the children (M age=4.47 years, SD=1.65) were female (n= 86, 54.8%). Parents identified their child’s race and ethnicity from a fixed set of categories as: White, non-Hispanic (n=93; 60%), Black/African American, non-Hispanic (n=23; 14.8%), Hispanic/Latino (n=23; 14.8%), Asian/Asian American (n=8; 5.2%), and multi-racial (n=8; 5.2%). Children’s mean hemoglobin A1c value at the orientation visit was 8.43% (SD=1.35), which was gathered on average 67 days (SD=37.4 days) from date of diagnosis.

Of 157 participants, 97.5% (n=153) were retained for the duration of the study through 15-months post-randomization (see Table 3). At 3-, 6-, and 9-months post-randomization, 98.7%, 98.1%, and 97.5% of participants completed data collection, respectively. Only 4 (2.5%) participants either withdrew due to time commitment and/or family relocation or were unable to be reached for follow-ups (see Table 4 for a comparison of baseline and 15-month participant demographics). There were no differences in retention between intervention (112/115; 96.5%) and control (41/42; 97.6%) groups.

Table 3.

Data completeness across study time points post-randomization

Data Completion
Group	Orientation	3-mon (1st Check In)	6-mon (2nd Check In)	9-mon (1st Follow-Up)	15-mon (2nd Follow-Up)
SC (n=116)	116	114	113	112	112
UC (n=42)	42	41	41	41	41
Total	157**	155/157	154/157	153/157	153/157
		(98.7%)	(98.1%)	(97.5%)	(97.5%)

SC = Stepped Care; UC = Usual Care

^**1 participant removed due to change in eligibility

Table 4.

Participant Demographics at Baseline compared to 15-months post-randomization

	Baseline n (%)	15-month post-randomization n (%)
Parent Marital Status
__Married	117/153 (76.5)	115/150 (76.7)
__Single	36/153 (23.5)	35/150 (23.3)
Parent Sex
__Female	144/157 (91.7)	141/153 (92.2)
__Male	13/157 (8.3)	12/153 (7.8)
Parent Race and Ethnicity
__White, non-Hispanic	97/156 (62.2)	94/152 (61.8)
__Black, non-Hispanic	23/156 (14.7)	22/152 (14.5)
__Hispanic/Latino	19/156 (12.1)	19/152 (12.5)
__Another	17/156 (10.9)	17/152 (11.2)
Public insurance	43/155 (27.7)	42/153 (27.5)
DKA at diagnosis	57/154 (37.0)	55/150 (36.7)
Annual household income
__<40k	27/135 (20.0)	26/131 (19.9)
__40k–100k	46/135 (34.1)	46/131 (35.1)
__>100k	62/135 (46.0)	59/131 (45.0)

Participant Perspectives

In satisfaction questionnaires, participant perception of the trial experience was assessed, including trial expectations, warmth of study staff, positive aspects, and areas for improvement. Based on post-study satisfaction questionnaires (n=154), about 95% of study participants either ‘strongly agreed’ or ‘agreed’ that recruitment and retention methods were acceptable (i.e., ‘interactions with study staff were positive’ ‘gift cards/items were appropriate’). When participants were asked to provide open-text responses about positive study aspects, their written answers often referenced study staff (e.g., ‘helpful & friendly staff,’ ‘The staff shows great concern for…my child’ ‘[staff member’s name]’), retention items (e.g., ‘We [received] cards for birthday/holidays,’ ‘[child] cuddles every night with [the stuffed elephant provided at the “meet and greet”]’), and low participant burden (e.g., ‘surveys were easy to take,’ ‘parking’).

Some families (n=18) also completed structured, post-study qualitative interviews to gather additional information about trial satisfaction. Purposeful sampling was conducted to balance parent demographic characteristics (e.g., gender, age, race and ethnicity) and intervention response (e.g., site, group assignment, intervention step reached). Broadly, qualitative interviews assessed participants’ impressions of trial participation and the intervention, positive and negative aspects, perceived reasonableness of expectations, barriers to participation, and perceived helpfulness. Research staff who had not worked directly with the participants conducted the interviews to reduce biased responding. Responses often related to recruitment and retention efforts, such as: ‘meeting [staff member] during office visits and doing those questionnaires with her…I enjoyed seeing her;’ ‘The tokens of appreciation are amazing;’ ‘When [child] was…first diagnosed [he would say] do I get to meet with the lady before? He thought that was cool;’ ‘I liked how they were personal… my daughter got a card for her birthday;’ and ‘They would give [child] a little coloring books with stickers’.

Discussion

The current manuscript describes our team’s multi-pronged approach to and implementation of recruitment and retention strategies, which we found to be feasible and useful for recruitment and retention of a relatively large, diverse sample of participants in a behavioral RCT. Of families reached and eligible, 58% were successfully recruited, which is notably higher than recruitment rates of 17% to 40% in similar pediatric behavioral RCTs (Ghai et al., 2014; Kazak et al., 2005; Stehl et al., 2009). Given that recruitment data from the present investigation shows that 54% of enrolled families had been approached for an in-person “meet and greet” during their child’s hospital stay, this suggests that in-person contact is certainly helpful for recruitment, but not imperative, as 46% of enrolled families did not experience in-person contact prior to enrollment.

Over 97% of the sample was retained for 15-months, which is higher than that of prior trials with newly diagnosed and long-term pediatric patients with type 1 diabetes (Herbert et al., 2016; Mackey et al., 2016), and it is also higher than that of trials for parents of children newly diagnosed with a chronic illness generally, which range from 63% to 93% (Hoff et al., 2005; Mullins et al., 2012; Sahler et al., 2013; Stehl et al., 2009). However, a review of 40 intervention studies for parents and youth living with chronic illness reported attrition up to 54% (Karlson & Rapoff, 2007), suggesting that, among parents who decided to enroll in the current study, heightened parent stress at the time of diagnosis combined with the strategies discussed herein may have contributed to parents’ continued engagement with clinical research team members. For example, our retention strategies, such as providing a stuffed animal and toys to children and training staff to use positive communication skills with parents (like validating families’ difficult experiences) may have had a positive emotional impact on parents during an otherwise stressful and isolating time, leading parents to continue engagement with us. Further, retention did not differ between our intervention and usual care groups (96.5% vs. 97.6%), suggesting that receiving an active intervention was not the primary reason participants stayed in the study and further points to the effectiveness of our retention strategies.

Lessons learned from this trial’s experiences with recruitment and retention may have implications for investigators designing and conducting behavioral RCTs. We made purposeful protocol decisions to reduce participant burden for busy parents of young children, including evening and weekend availability for study contacts and intervention activities, text messaging, and providing child-friendly activities and/or childcare at study visits. Investigators may consider incorporating funds to support these strategies into grant budgets. For example, budgeting for graphic design costs to develop a professional, identifiable study logo and to purchase retention items printed with the study logo can help participants maintain study recognition over time. Inclusion of funding for effort for medical Co-Investigators ensures all collaborators are able to dedicate time to supporting the study’s success. Budgeting adequate funds for study coordinators and research assistants, who embody and implement many of these recruitment and retention strategies (i.e., personalizing relationships with participants, reimbursement for their travel to meet families at satellite clinics, tracking data collection) is essential. Funds were also intentionally allocated to an increasing financial incentive structure over time to encourage trial retention. Finally, investigators conducting RCTs that will require extensive participant activities should consider using a behavioral run-in prior to randomization. These procedures convey respect for participant time and interest and offer additional opportunities to ask questions about trial procedures, increasing the likelihood that participants will complete the full trial (Ulmer et al., 2008).

The recruitment and retention approaches described herein resulted in a study sample that was relatively more racially, ethnically, and socioeconomically diverse than most other trials with young children with T1D (Marker et al., 2020; Pierce et al., 2017). However, some of our strategies and exclusion criteria may have systematically excluded a subset of the pediatric T1D population, which limits generalizability. As with all RCTs, participation required considerable parental effort, and it is possible that the requirement of baseline questionnaires to be completed within eight weeks of the child’s diagnosis and research procedures made it challenging for families with fewer resources and less time to enroll. In addition, parents who were not fluent in English were excluded from participation, and relatively few fathers enrolled. While the research team employed in-person recruitment methods coupled with mailings, phone calls, and text messages, a subset of participants could not be reached. Repeated failed contacts could be due to several factors, such as disinterest in study participation, limited resources for participation, or another reason. Given the lack of IRB approval to follow-up with families who did not consent to study participation, interpretations are limited.

Future research should explore barriers to trial participation for families who were unable to be reached in an effort to develop effective recruitment strategies for this subset of families with children with pediatric chronic illness. This might include brief qualitative interviews or survey completion with participants, who were previously unreachable, while they are present in clinic. More work is also needed to understand if similar recruitment/retention strategies are effective with more diverse populations (e.g., Spanish speaking). Given the growing racial and ethnic diversity of the United States, it is essential to identify strategies that allow for increased representation within clinical trials (Mather & Lee, 2020). Additionally, future investigations should strive to assess whether recruitment strategies are effective in recruiting fathers, who could offer a different perspective on pediatric chronic illness management.

In conclusion, the presented strategies appear to be useful, feasible to implement in the context of a multidisciplinary clinic, and highly acceptable for both recruitment and retention of caregivers of young children newly diagnosed with T1D. These strategies may be translatable for use in other RCTs for behavioral interventions for other pediatric chronic condition populations.