Step 1: Needs assessment

We started the needs assessment process by creating two relevant stakeholder groups to advise intervention development, literature review, and data collected by our team. One of these groups was our expert panel, which consisted of individuals across the US with expertise in bioethics, neonatal clinical trials, stakeholder engagement, adult education, health disparities, qualitative methods, and intervention development. The other group was our parent panel of 10 former NICU parents from the Pacific Northwest Seattle region; more than half were from minoritized populations and more than half had experience being recruited for neonatal research. Parent panel members were paid for their participation. These two groups provided guidance through all steps of the project’s development.

The needs assessment consisted of a literature review supported by a recent systematic review of barriers and facilitators to participation in pediatric research [16] (see Table 1) and a review of findings from previous work conducted by our team members, including surveys [8, 9] and interviews of parents approached for NICU research [17, 19], interviews [21] and Delphi surveys [29] around issues of respect in research, and interviews with pediatric research coordinators [15, 22].

Table 1 Literature review supporting development of BRIEF intervention

We chose to structure the BRIEF intervention to follow our previously developed 4-stage conceptual framework of factors impacting relationship building in the researcher-participant-surrogate recruitment interaction within pediatric research [15, 22]. That framework developed out of interviews with pediatric research team members guided by a model of trust-building within community academic research partnerships [59]. Our prior conceptual work identified the relational level (i.e., interactions between research team members and parents) as central to relationship-building between pediatric research team members and parents of potential participants. Our framework delineated four temporal stages of relationship building: pre-approach (deciding whether, when, and how to approach a family), initial connection (starting the interaction with parents), building connection (ongoing interaction, including discussing the study), and follow-up (closing the interaction, preparing for next steps, and working towards positive longer-term relationships) [14]. For BRIEF, we prioritized considerations of equity (present at all levels of the framework) [15, 22], in order to support increased inclusion of and improved experience for minoritized and marginalized populations. We did this by designing our contributing studies, in particular our interviews of parents approached for neonatal research (#2 in the following paragraph), so that minoritized and marginalized populations were over-sampled to assure their views were included.

We were informed by four recent empirical studies performed by our research teams. These include (1) surveys of parents approached for a single neonatal clinical trial, (2) interviews of parents approached for several neonatal research studies, (3) interviews of adult participants in genomic research, and (4) a Delphi survey process with patients in community clinics. We will briefly describe these four lines of inquiry and their contribution to the development of BRIEF.

First, we were informed by recent surveys of parents who were asked to participate in a neonatal clinical trial [8, 9]. We had conducted surveys at 12 US NICUs with parents of infants who enrolled in the High-dose Erythropoietin for Asphyxia and encephaLopathy (HEAL) trial [60] or who were eligible but declined enrollment. In these surveys, most parents, but particularly those who chose not to participate in HEAL, wanted greater involvement of the clinical team when learning about research. Important items included the central role of a positive relationship with members of the research team, altruism as a major motivator for participation, and challenges around the right timing for approaching a family for a neonatal clinical trial.

Second, we recently completed a project interviewing parents approached for neonatal research, including both those who participated in and those who declined participation in neonatal research. Key findings from this work included parents’ desire for greater involvement of the clinical team in learning about research and the essential role of emotions and relationships in enrollment decision-making [17]. Motivations for participation included altruism and a hope for benefit to their infant; reasons against participation included burdens of participation to the family and perceived risk to infants [19].

Third, we were informed by recent interviews with participants in genomics research about respect [21]. Findings revealed domains of respect that illustrate the importance of attending to how researchers demonstrate respect across the entire research process. Specific key findings included connecting with families, showing empathy during interpersonal interactions, and responding to the family’s needs. Additionally, presenting the option to participate or not in a neutral way was seen as a way researchers may illustrate respect.

Finally, we were guided by findings from a modified Delphi survey with patients in community clinics about priorities for feeling respected in research [29]. These asked respondents to identify the most important factors for demonstrating respect in this setting. The highest ranked were as follows: receiving information to make a decision and positive interpersonal interactions, reiterating the importance of improving the recruitment and consent process, including the interpersonal relationships that take place during this process.

The IM framework differentiates between the “personal level” at which an individual performs a particular behavioral outcome and the “environmental level,” the social and physical environment where the individual exists. Applying the IM framework to the current project, the research team members exist at the “environmental level” and serve to impact change through their social interactions and providing a welcoming physical space to parents (see Fig. 1). Parents then are the individuals at the “personal level” making the decision that leads to the behavioral outcome of interest: whether or not they choose to participate in neonatal research. Importantly, our team agreed that the parent recruitment experience was vital and valuable independent of their ultimate enrollment decision. While an increased enrollment rate is an important goal to increase the representativeness of the research population, each family must be free to make the enrollment decision best aligned with their values. Therefore, we determined that we must measure outcomes related to the parental recruitment experience.

Fig. 1

Logic model of change: targeting “environmental level” of research team members. For BRIEF, our team decided to focus on the environmental level (e.g., directly impacting behavior of research team members) rather than personal level (e.g., directly impacting behavior of parents)

Informed by these data as well as a literature review on the topic, we created a logic model of change for the development of the BRIEF intervention (see Fig. 1). With our expert panel, parent advisory panel, and partner clinical trialists, we set out to determine the appropriate agent of change for our intervention. We identified that the intervention could directly target: (a) parents deciding whether to participate in neonatal research, (b) research team members, or (c) both. We decided to focus the BRIEF intervention on research team members. This choice reflects the importance of the relationship developed between researchers and potential parents approached for consent, including the researchers’ responsibility to improve parents’ experience and address disparities in research participation.

For neonatal clinical trials in the US, there is wide variability by site regarding who approaches families for participation [61]. This may include the following: the physician primary investigator, physician co-investigator, research coordinator, research assistant, and research nurse. These individuals may have anywhere from extensive to no knowledge about clinical medicine, may or may not be part of the infant’s medical team, and may or may not have deep knowledge of the research. To be broadly applicable across sites, we designed BRIEF to be useful for all individuals approaching families for neonatal research regardless of their role, training, and experience.

Step 2: Creation of matrices of behavioral determinants and performance objectives

We drew on the previous qualitative research done with research participants and research staff (as described above) to identify key features of respectful, equitable recruitment interactions. We conceptualized our environmental outcomes within our previously created framework, in which we delineated four stages of relationship building in pediatric clinical research (pre-approach, initial connection, building connection, and follow-up) [14]. The identified key features represent the environmental outcomes we hoped to improve with our intervention.

The integration of determinants from step 1 led to several major findings that align with the constructs of the social cognitive theory (SCT). First, parents wanted to learn about research including participants’ risks, benefits, and research burden through a closer interaction with the research team (behavioral capability and outcome expectations). Second, parents believed that arming themselves with information about research would increase their confidence to make an informed decision (self-efficacy). Third, parents needed to be emotionally ready to enroll in research and valued emotional empathy from the research team (emotional coping responses, expectancies, determinism). Fourth, parents were motivated to participate in research when they understood the benefits for their infants and/or were guided by altruism (behavioral capability and outcome expectations).

SCT proposes that human actions and behaviors are influenced by the dynamic interplay between three groups of determinants: personal, behavioral, and environmental factors [62]. This theory focuses on the interaction between internal personal factors such as attention, memory, and motivation and external factors such as the opportunity for a desired outcome. A key component of SCT is self-efficacy, defined as an individual’s belief that they are capable of performing a behavior [63]. Behavior change therefore is determined by the interaction and alignment between outcome expectations (e.g., an individual’s belief that their behavior will result in a particular outcome) and self-efficacy expectations (e.g., an individual’s belief that they can produce that outcome). Research team members responsible for obtaining consent for neonatal clinical trials are highly motivated to improve their ability to connect with families and meet them at the stage of the families’ readiness for research but often feel limited in their skills to create a positive environment for parents. We selected SCT because its constructs align with the environmental needs of parents deciding to enroll in neonatal trials, and by addressing the determinants, we will increase the capacity and skills of the research team to create an environment that is conducive to positive research recruitment for parents.

For BRIEF, we selected four SCT determinants of behavior change: behavior capability (knowledge and skills to perform task), self-efficacy, outcome expectation, and subjective norms. These were prioritized based on relevancy to the problem and changeability with the potential intervention. We then began identifying and prioritizing performance objectives through multiple rounds of discussions and meetings with our expert panel and our parent advisory panel. We met with each group separately multiple times over a 9-month period to present our initial list of 16 items and then received feedback both in the group meeting and through email correspondence with individual members. Over this period, the items were iteratively refined to a final set of 10 performance objectives with extensive input from our parent and expert advisory groups. Figure 2 shows the selected outcomes, performance objectives, and determinants.

Fig. 2

Logic model of change: selection of environmental determinants, performance objectives, and outcomes

We next began linking determinants to performance objectives (see Table 2). For example, performance objective 2 asks the research team member to meet with the bedside nurse before approaching the family and use that interaction with nursing to support the early steps of relationship-building. We identified four determinants to support research team members doing this effectively: demonstrate the ability to identify and engage with the bedside nurse and share information gained with the family (behavioral capability), express confidence in learning about the patient and family situation from nursing and verbalize knowledge gained to family (self-efficacy), expect that nursing may have important insights into how the family would prefer to be approached about research (outcome expectation), and expect that nursing will respond positively to engagement with the research team (subjective norms).

Table 2 Environmental outcome matrixStep 3: Choosing theoretical methods and practical strategies

After identifying SCT-based determinants, the research team moved to step 3 in which we selected theory-informed methods and practical applications through which the determinants may be influenced. We chose methods from three theories: SCT, theory of information processing [64, 65], and the trans-theoretical model [66]. These theories were chosen largely based on our review of prior literature of similar interventions to improve recruitment processes in other settings [25,26,27] as well as input from our expert panel members with experience and expertise in IM and in intervention development.

As the behavioral determinants captured personal as well as environmental factors, we were guided by the construct of SCT to help identify the performance objectives that align under each of the four stages of relationship building.

We chose to inform the development of the BRIEF materials using the information processing theory as the research teams play the role of information receivers (receiving information on how to best present neonatal research to families) and givers (giving information to families in an effective manner). This enabled our team to think about how information from BRIEF would be processed by the research team to help research teams anticipate how research information will be processed by the families. The theory of information processing suggest that individuals do not simply respond to stimuli but process information based on structural features of the incoming information [64, 65]. The act of processing information is impacted by such things as where the information is stored, cognitive processes to transfer stored information, and awareness of how information is processed [67, 68]. We chose this theory to optimize our presentation of critical information as part of the pre-work (e.g., chunking key items into a four step model) how we formatted our synchronous session (e.g., guided discussion to support learning of material), as well as provide opportunities for role play to practice these skills.

The findings from step 1 showed that families have different amounts and aspects of information about research as well as emotional readiness to engage in research. Research teams approaching these families need to be aware of their stage of readiness for research participation and provide information that will meet the needs of families to motivate behavior change. The trans-theoretical model suggests that behavior change is a conscious and intentional process that occurs over six stages: precontemplation, contemplation, preparation, action, maintenance, and termination [66]. Progress through stages occurs through constructs of motivational readiness, processes of change, self-efficacy, and decisional balance [69]. These constructs describe why and how a behavior change occurs and can be utilized to support desired behaviors [70, 71]. From this theory, we were able to create a process whereby we could identify and respond to participants’ varied stages of readiness to change (e.g., from pre-work responses) in order to support the synchronous session.

We then identified practical applications suitable for the population (research team members) and the context (busy NICU research teams). Table 3 shows the methods chosen and relevant practical applications, which in turn supported the components of the intervention. Components to support these were chosen: pre-work including didactics, personal stories from parents, example scenarios, and self-reflection and a single synchronous session including didactics, sharing of personal experiences, and role-play practice scenarios. All components were directly linked to one of the ten specific performance objectives and a location within our conceptual framework.

Table 3 Theoretical methods and practical applicationsStep 4: Designing the intervention

We developed and refined the intervention components, including self-guided pre-work and a single in-person synchronous session. Educational videos were developed in collaboration with the Seattle Children’s Marketing and Communications team. Videos included NICU parents and research coordinators sharing their experiences and perspectives and scripted skits depicting hypothetical research recruitment approaches between research coordinators and NICU parents. The pre-work also included self-reflection questions to increase engagement and prepare learners for the in-person session. Thinkific, a software platform for the creation and delivery of online courses [72], was used to develop the pre-work where participants would engage with the intervention materials by learning about the performance objectives, viewing educational videos, and answering a series of questions inquiring about their experience with recruitment and the performance objectives. Because the target audience includes anyone on a research team that would recruit participants (i.e., investigators and coordinators), we aimed to use a platform for the pre-work that would maximize accessibility (e.g., allow for access at any time and from any location). The online pre-work platform was completed within three weeks prior to the in-person session.

Each research team member participated in a single 2-h group in-person session. This format was selected to maximize the efficacy of the role-play exercise, optimize engagement, and support team building and was felt to be feasible with a single-site neonatal clinical trial. The intervention occurred in a small group format (target 4 or 6 participants per session) to enable 2 or 3 partner break-out groups as well as ample opportunity for discussion in the larger group. Sessions occurred in a conference room at the medical center with audiovisual capabilities. Components were created to be easily adaptable to a remote synchronous setting (e.g., videoconferencing) which would likely be needed for expansion to other settings. Our team developed four role-play scenarios (two scenarios, each with two distinct characters being approached for a different neonatal research study). Each of the four scenarios was created to facilitate practice with a sub-set of the performance objectives (see Fig. S1). They were designed around two neonatal clinical trials to allow the practice to apply across different trials.

The synchronous session was divided into four sections: introduction, an overview of the conceptual model, and didactics; team member discussion of reflection questions from pre-work; role plays; and closing discussion with a feedback survey. The intervention was designed to be piloted in the NICU, but applicable to multiple recruitment settings, and to achieve two goals: (1) improve the recruitment experience for families and (2) increase enrollment rates (and decrease disparities in enrollment). All components were created iteratively in collaboration with our parent panel and expert panel members. The synchronous session was led by the project lead and lead author of the current manuscript.

Improving the research experience for parents from marginalized and minoritized populations was a central goal of this project. Because of this, as we have described, we prioritized information from these groups in the identification of our performance objectives and the development of our content. However, we must recognize that parents from marginalized and minoritized populations are highly heterogenous with variability in whether and which additional challenges they face. Because of this, we did not recommend researchers change their approach solely based on parental demographics. Our team and our parent and expert advisors felt such broad generalizations would be inappropriate.

Step 5: Creation of an adoption and implementation plan

Next, we identified a partner neonatal clinical trial to pilot the BRIEF intervention. For partnership, we identified the single site Darbepoetin plus slow-release intravenous iron (DIVI) [73]. DIVI is a phase II trial to demonstrate Darbepoetin’s feasibility, safety, and potential benefit with different iron formulations and its impact on transfusion requirements and on neurodevelopmental outcomes for infants born before 32 weeks gestational age. Logistical and methodological advantages included broad inclusion criteria, the need for early enrollment (within the first three days of life), and a team highly committed to supporting this pilot-phase collaboration, which was essential to get feedback to be used for the creation of a version to be used across a multi-site clinical trial.

In future iterations, we will test the BRIEF intervention at multiple sites and for multiple neonatal clinical trials. To maximize the applicability across settings, we will endeavor to include studies with variety across key components of neonatal research, including the following: inclusion criteria, enrollment window (including whether prenatal consent is possible), intervention characteristics, and follow-up duration. When this is done, we will need to create an adoption and implementation plan specific to the settings chosen.

Step 6: Creation of an evaluation plan

To evaluate the intervention, we developed a multi-stage evaluation plan that included five distinct levels of outcomes.

Data at the infant level (P1) included the dichotomous outcome of whether the infant ultimately participated in the DIVI study. We also designed a targeted data extraction for all infants eligible for the DIVI study at birth in order to collect demographic and medical information to evaluate for any differences between DIVI participants and DIVI non-participants. Selected demographics include race, ethnicity, insurance status, gestational age, birthweight, and sex. Selected medical information include key prenatal factors (maternal prenatal antibiotics, prenatal steroids), key factors at birth (Apgar scores, resuscitation needed), and disposition to discharge.

Two outcomes were specific to the parent experience level. We invited all parents who were approached by the DIVI team for a recruitment discussion to participate in a one-time REDCap-based parent survey (P2). The rationale of this outcome was to evaluate parents’ experience of being recruited to participate in the DIVI study with a focus on areas addressed by the BRIEF intervention, to gather demographic information, and to inquire about parents’ willingness to participate in a follow-up interview. As part of the IM process, the survey development was centered on assessing parental experience around each of the ten performance objectives present in the BRIEF intervention through an iterative process including input from our parent advisory panel and expert panel as well as cognitive interviews with NICU parents to assess interpretation of our questions. We developed two novel 5-point Likert scale questions for each performance objective (total 20 questions), in addition to novel questions about related higher level constructs (e.g., respect, trustworthiness, equity), a validated scale for trust in medical researchers [74], and respondent demographics.

Parents who participated in our survey were invited to participate in in-depth parent interviews (P3). The rationale of this outcome was to evaluate the parent experience of being recruited to participate in the DIVI study with a focus on areas addressed by the BRIEF intervention and to establish qualitative data on the parent experience. The interview guide was also structured around the ten performance objectives and related constructs with a similar process as for the survey questions described above.

The next outcome level focused on the interaction between parents and research team members. For a subset of parents approached, we gathered data from the parent-research team member recruitment discussions (P4). The data included two components: audio-recordings of the recruitment discussions between the DIVI team members and the parents and research team member self-assessment of the sample sub-set of DIVI recruitment discussions using the BRIEF Assessment Tool (Fig. S2), which was created as part of the IM process. The BRIEF Assessment Tool included 10 Likert scale questions, each linked to a specific performance objective. The rationale for these outcomes was twofold: first was to compare DIVI team members’ self-assessment of their achievement of each performance objective before and after BRIEF intervention training, and second was to compare DIVI team members’ self-assessment with external assessment of audio-recorded recruitment discussions using the same tool.

Our final outcomes were specific to the research team member level (P5). These included a multi-phase data collection process consisting of surveys after the pre-work, surveys after the synchronous session, and interviews at the project’s close. The rationale of these outcomes was to evaluate the DIVI team members’ experience of the BRIEF intervention so we may refine it for future iterations.

As of the writing of the current manuscript, data collection for all five levels of outcomes for this single-center pilot assessment of BRIEF within the partner DIVI trial has been completed. Analysis is ongoing, and our findings are currently being drafted, after which they will be submitted as manuscripts to be considered within peer review medical journals.