Abstract
Stakeholder-developed interventions are needed to support pediatric intensive care unit (PICU) communication and decision-making. Few publications delineate methods and outcomes of stakeholder engagement in research. We describe the process and impact of stakeholder engagement on developing a PICU communication and decision-making support intervention. We also describe the resultant intervention. Stakeholders included parents of PICU patients, healthcare team members (HTMs), and research experts. Through a year-long iterative process, we involved 96 stakeholders in 25 meetings and 26 focus groups or interviews. Stakeholders adapted an adult navigator model by identifying core intervention elements and then determining how to operationalize those core elements in pediatrics. The stakeholder input led to PICU-specific refinements, such as supporting transitions after PICU discharge and including ancillary tools. The resultant intervention includes navigator involvement with parents and HTMs and navigator-guided use of ancillary tools. Subsequent research will test the feasibility and efficacy of our intervention.
Keywords
Introduction
Clinicians and families make complicated medical decisions for pediatric intensive care unit (PICU) patients (1,2). In some cases, decisions have life-altering implications, such as whether to place a tracheostomy tube, participate in research, or withdraw life-sustaining therapies (3 -5). Research demonstrates that high-quality communication positively impacts decision-making for families (6,7).
Despite the importance of high-quality clinician–family communication, few formally developed and tested interventions exist to support parental decision-making in the PICU (5,8). Moreover, there is growing emphasis on the importance of engaging key stakeholders in research, including intervention development (9,10). However, few data describe methods for and outcomes of stakeholder engagement in research (11 -14). Thus, there is a need for a PICU-specific communication and decision-making intervention. There is also a need for literature describing stakeholder engagement in developing interventions.
In this article, we describe the process and impact of engaging stakeholders to develop a PICU-specific communication and decision-making intervention. We also describe the resultant intervention. Subsequently, we will test our intervention in the clinical setting.
Methods
Intervention Development Process Overview
With few models for engaging stakeholders in intervention development (15), we developed a process (Figure 1) combining qualitative methodology and community-based participatory research principles. Qualitative research uses experiential data to describe a phenomenon (16). Community-based participatory research engages those affected by an issue to effect change and has been useful for intervention development (17). For us, the “community” equates to stakeholders: parents of PICU patients, healthcare team members (HTMs) caring for PICU patients, and clinical research experts. We involved stakeholders as research team members (research stakeholders) and as participants (community stakeholders) in focus groups (FGs) and interviews (Is).
Intervention development process. PICU indicates pediatric intensive care unit
Composition of Research Stakeholders
Our research stakeholders included 3 groups: (1) parents, (2) HTMs, and (3) research experts (researchers). We included 7 parents of 6 children (cared for at 3 different ICUs) with diverse medical problems including cancer, complex chronic illnesses, gastrointestinal disorders, and neurologic/neuromuscular disorders. Four parents had a child who died prior to this project. The 11-member HTM group included PICU and subspecialty physicians, as well as a PICU bedside nurse, advanced practice nurse (APN), social worker, chaplain, and hospital bereavement program representative. Healthcare team members worked at the same institution. The 9-member researcher group included experts in palliative care, decision-making, communication, statistics, clinical research, adult intensive care, and a parent. The researchers represented 3 institutions. Acting as a liaison, 1 parent participated in the parent group and researcher group and 1 nurse participated in the HTM group and researcher group. The research stakeholders were 83% (24/29) women and 83% white.
Define Intervention Core Elements
We started with an adult intensive care unit navigator-based model called Four Supports (18). We chose this model because other settings have successfully used navigators (19 -21) and because our work identified potential benefits to having a point person akin to a navigator (3).
The 3 research stakeholder groups met separately to identify desirable core intervention elements. We then brought the 3 research stakeholder groups together to discuss each element, obtaining additional suggestions via e-mail. This process involved 25 meetings over 12 months. Two authors (K.N.M. and L.C.C.) summarized the meeting input based on the review of the meeting notes and e-mails. Because we did not record meetings, results about meeting content are presented qualitatively.
Community Stakeholder Input via FGs/Is
The hospital institutional review board approved the FG/I component. After obtaining written consent, we conducted FGs/Is with additional parents and HTMs, the community stakeholders. We organized FGs into categories: bereaved parents, nonbereaved parents, PICU physicians, subspecialty attending physicians, APNs, PICU bedside nurses, social workers, chaplains, and other team members (eg, case managers). We avoided interdisciplinary FGs because we felt that mixing groups might inhibit conversations. For example, a nurse might be reluctant to comment on physician practices with physicians present.
We identified English-speaking parent FG/I participants through medical record review, identifying English-speaking parents of PICU patients admitted within the previous 6 months to 3 years and with a PICU admission of >3 days. Two study team members (the principal investigator and hospital bereavement program director) reviewed the list of eligible bereaved parents to exclude parents experiencing complicated grief. We sent letters to 90 parents, then made phone calls to nonresponders 2 weeks later. Research stakeholders identified additional parent participants by distributing brochures to parents they knew who met our eligibility criteria. English-speaking parents unable to participate in a FG due to scheduling or transportation barriers participated in an interview (conducted at the hospital or by phone). To recruit Spanish-speaking parent participants, we identified and requested study participation from Spanish-speaking parents of current PICU patients. We chose this approach to maximize Spanish-speaking parent recruitment, which in our experience is more difficult after a child’s PICU discharge. Spanish-speaking parent interviews were conducted in Spanish at the hospital.
We invited HTMs by e-mail or at staff meetings to participate in FGs conducted at the hospital. The FG/I moderators (a healthcare communication researcher, a sociologist, and a Spanish-speaking social worker with qualitative research experience) used a semi-structured discussion guide. No moderators provided clinical care to the children of participating parents or worked with participating HTMs. The guide prompted moderators to ask about PICU communication, mechanisms to improve PICU communication, comments (positive or negative) about having a navigator to support PICU communication, and characteristics of an ideal navigator (see Appendix A). Following FGs/Is, participants completed a survey requesting demographic information.
Focus Group/Interview Data Analysis
We audio recorded FGs/Is and transcribed recordings verbatim. Focus groups/interviews conducted in Spanish were transcribed, translated, and reviewed by the Spanish-speaking moderator for accuracy. We used a qualitative data analysis program, Dedoose (Dedoose version 5.2.0, 2014; SocioCultural Research Consultants, LLC [www.dedoose.com], Los Angeles, California). We analyzed qualitative data using directed content (22). We identified excerpts (blocks of text with similar content) and assigned excerpts to 3 content areas: (1) PICU communication patterns, barriers, and facilitators (not discussed in this report), (2) input on intervention design, and (3) navigator characteristics. Two people read 2 conference transcripts and created unique lists of codes and then narrowed the code list by eliminating redundancy and combining similar codes. The revised code list was then used to code 2 other transcripts, creating new codes as necessary. This process was repeated until no new codes were identified, creating a final coding dictionary. Finally, the 2 coders independently assigned codes to each excerpt. Disagreements were resolved by consensus until achieving 100% agreement. We assigned codes to broad categories based on group consensus. We used descriptive statistics for all quantitative data using Excel (Microsoft Office 2010, version 14.0.7116.5000)
Research Stakeholders Finalized the Intervention Design
The principal investigator preliminarily described the intervention by combining input from the community stakeholders and research stakeholders. The research stakeholders reviewed, amended, and finalized the intervention.
Results
Core Elements Identified by Research Stakeholders
The core intervention elements of navigator activities identified by the research stakeholders included initial meetings with HTMs and parents, daily meetings (“check-ins”) with HTMs and parents, family meetings, and end-of-stay activities.
Everyone indicated that the intervention should be individualized to parents’ needs and support communication among HTMs, not just communication between parents and HTMs. The groups felt the intervention should focus on parents but include relevant family members. Thus, hereafter, we use the term parents to describe 1 parent, both parents, and/or other family members.
Some HTMs, particularly social workers and chaplains, expressed concern about overlap with services by existing HTMs. To avoid duplicating or compromising existing support systems, people suggested that the navigator meets with the patient’s HTMs prior to meetings with parents. After each meeting with the parents, the navigator should report relevant components of their discussion with the HTMs.
The research stakeholders felt regular family meetings could improve communication, if parents desire such meetings. Some stakeholders felt that having regular family meetings could normalize the experience, as opposed to relegating family meetings to emotionally charged discussions occurring during acute crises and/or when clinicians deliver bad news or discuss difficult decisions.
Research stakeholders, particularly parents, identified transitions out of the PICU as challenging, specifically transfers to a non-PICU hospital bed, discharges from the PICU to another facility or home, or when the child dies. Stakeholders felt the navigator could support such transitions by preparing parents for their next situation, ensuring parental understanding about subsequent medical care and having postdischarge contact with parents. Developing a checklist for end-of-life care and providing bereavement support were discussed.
Finally, parent stakeholders suggested that parents receive written information about people’s roles (eg, what a resident does), difficult medical terms, and therapies encountered in the PICU (eg, ventilators). Some stakeholders suggested that a list of questions might help parents consider what to ask HTMs.
Input From Community Stakeholder FGs/Is
We conducted 2 parent FGs, 1 included parents of living children (n = 3) and 1 included parents of deceased children (n = 4). Sixteen parents participated in interviews. The parents’ children had neurologic disorders (n = 9), cancer/tumors (n = 4), respiratory disorders (n = 4), musculoskeletal disorders (n = 1), cardiovascular anomalies (n = 2), sepsis (n = 2), and liver disease (n = 1). Fifty-two HTMs participated in 8 FGs (Table 1). Three parents and 3 HTMs did not complete surveys. The FGs included a social worker and chaplain who were also part of the research stakeholders because they provided unique PICU expertise and experience. Comments about our proposed intervention fell into 2 categories: intervention design (Table 2) and navigator characteristics (Table 3).
Demographics of the Community Stakeholders (Focus Group/Interview Participants).
Abbreviations: APN, advanced practice nurse; HTM, healthcare team member; NA, not applicable; PICU, pediatric intensive care unit.
aMissing data from 3 parents.
bMissing data from 3 HTMs.
cReported as vocational.
dIncludes case managers, speech and music therapists, and child-life specialists.
Input on Intervention Design and Exemplar Quotes From Focus Groups and Interviews.
Abbreviations: APN, advanced practice nurse; HTM, healthcare team member; ParentE, English-speaking parent; ParentS, Spanish-speaking parent; PICU, pediatric intensive care unit; Psysoc, psychosocial; RN, bedside nurse.
aIndicates the number of excerpts (block of text with similar content) based on the stakeholder group. The numbers in the shaded rows represent the number of excerpts assigned to each category. The numbers in rows without shading represent the number of excerpts assigned to each code. Percentages indicates the percent of excerpts assigned to a code within each category. The sum of percentages for a particular category may be >100 because some excerpts were associated with more than one code.
Input on Navigator Characteristics and Exemplar Quotes From Focus Groups and Interviews.
Abbreviations: APN, advanced practice nurse; ParentE, English-speaking parent; ParentS, Spanish-speaking parent; PICU, pediatric intensive care unit; Psysoc, psychosocial; RN, bedside nurse.
aIndicates the number of excerpts (block of text with similar content) based on the stakeholder group. The numbers in the shaded rows represent the number of excerpts assigned to each category. The numbers in rows without shading represent the number of excerpts assigned to each code. Percentages indicates the percent of excerpts assigned to a code within each category. The sum of percentages for a particular category may be >100 because some excerpts were associated with more than one code.
Input on Intervention Design
All respondents, except Spanish-speaking parents, described benefits to assigning families a support person when discussing the navigator role. English- and Spanish-speaking parents focused on how the navigator could serve as a liaison between HTMs and parents, support understanding, and advocate for parents. Healthcare team member comments focused not only on how the navigator could serve as a liaison and advocate but also on how the navigator could organize and facilitate discussions. Parents proposed having regular family meetings.
In addition to the navigator, stakeholders suggested other resources or tools. Parents (particularly English-speaking parents) and HTMs (except physicians) felt a handbook with information about the PICU could help parents. Some parents, a physician, and a nurse suggested that a “sign-in log” might help parents track who comes and goes from their child’s room.
Most HTM-voiced concerns about having a navigator indicated worry that the navigator’s role could overlap with that of existing HTMs. Parents did not express this concern. Another concern expressed by HTMs and parents was that this position could be too much work for a single person.
Navigator Characteristics
All stakeholders talked about personality traits that would enhance the navigator’s ability to provide emotional support. Those characteristics included: supportive, empathetic, calm, objective, assertive, organized, and respected. Everyone indicated that she/he should be able to communicate medical terminology in lay language. All groups except Spanish-speaking parents described that she/he should be able to mediate conflict. Having the ability to engage people from diverse cultural backgrounds was a major focus of conversation related to navigator skills. Spanish-speaking parents recommended having Spanish-speaking navigators and/or high-quality, available interpreters. Input on the navigator’s preferred background varied with 43.8% of comments, indicating he/she have a medical background (like a nurse), 21.9% a psychosocial background (like a social worker or chaplain), and 34.4% both a medical and psychosocial background.
Final Intervention Design
Research stakeholders named the intervention “PICU Supports.” Reflecting stakeholder input, PICU Supports seeks to provide emotional, communication, decision-making, information, and transition (ie, transitions out of the PICU or death) support. PICU Supports includes navigator activities (Table 4) and ancillary tools (Table 5), delivered based on individual needs/desires. PICU Supports is meant to augment existing HTMs or processes as needed (Figure 2).
The PICU Supports design. Navigator activities are noted in blue. PICU indicates pediatric intensive care unit.
PICU Supports Navigator Activities.
Abbreviations: HTMs, healthcare team members; PICU, pediatric intensive care unit
PICU Supports Ancillary Tools.
Abbreviations: APN, advanced practice nurse; HTMs, healthcare team members; PICU, pediatric intensive care unit.
Discussion
We describe a process for engaging stakeholders in intervention development. Stakeholder input added elements to our intervention (PICU Supports) including attention to transitions out of the PICU and tools such as the PICU handbook, a question prompt list, a place for journaling, and a mechanism for tracking HTMs involved in patient care. These elements might not have been included in PICU Supports without stakeholder input.
The need to address transitions out of the PICU was only partially addressed by the adult model which focused on end-of-life care. While PICU Supports addresses the challenges faced by parents of dying children, it also addresses the discharge needs of surviving children. The need for this expanded focus may reflect differences in mortality rates between adult ICUs (as high as 33%) versus PICUs (typically less than 7%; 23,24). It could also reflect the need for healthcare systems to address post-ICU syndrome, a documented problem for which organized treatment options are lacking (25). Regardless, stakeholder input directed its inclusion in PICU Supports.
The inclusion of ancillary tools is another stakeholder-driven distinguishing feature of PICU Supports. Variations in these tools have been studied in other settings. For example, researchers have shown the benefits of educational brochures and diaries for surrogates in the ICU (26,27). We included these elements because stakeholders identified them as potentially beneficial.
Stakeholder concerns also influenced the design of PICU Supports. As a result of worries about overlap with the roles of existing HTMs, stakeholders designed PICU Supports to function as an adjunct to ongoing resources with the navigator integrated into the medical team. To enhance navigator-medical team integration, PICU Supports has many contact points between the navigator and the medical team including before and after daily family discussions and organized family meetings and through the use of the provider information sheet (Table 5).
Challenges we encountered highlight some limitations. Scheduling parent FGs was difficult. Resultant smaller FGs leave the possibility that the conversations might have been more robust had we organized larger groups. We also had difficulty recruiting culturally diverse stakeholders and fathers; most stakeholders were white and most parents women. Thus, our intervention may not reflect needs of underrepresented groups. Also, finding resolution is difficult when stakeholders disagree. For example, there was no consensus about the navigator’s background. Finally, stakeholder input resulted in a complex intervention, raising questions about implementation feasibility.
This work, therefore, represents a first step. A subsequent pilot implementation study is needed to address unanswered questions such as what would be the ideal navigator background; whether or not to include all intervention components described here; which families should receive PICU Supports; how much time effort/cost PICU Supports requires; if a navigator should be available on weekends and evenings; and how to provide ancillary materials (eg, via the Internet).
Other limitations are noteworthy. We did not record research stakeholder meetings and thus only present qualitative information about those discussions. Focus group HTMs came from the same institution, although research stakeholders and community stakeholders included HTMs from other institutions and parents of children admitted to other PICUs. We cannot account for recall or recruiting bias. We did not analyze information conveyed in FGs/Is via body language or intonation. Finally, stakeholder input reflects a limited “pool” of knowledge/experience. We may have missed key stakeholders (eg, hospital administrators) or neglected important approaches to improve communication and decision-making.
Conclusion
Stakeholder input impacted the design of PICU Supports. Engaging stakeholders with varied backgrounds was challenging, identifying the need for innovation in recruiting stakeholders from diverse populations. Future work is needed to determine the implementation feasibility and impact of PICU Supports.
Footnotes
Appendix A
Authors’ Note
Navigate Study Team members are Terra Caswell, Sabrina Derrington, Reggie Duerst, Michael Goldberg, Anna R. Goldberger, Natalie Haber-Barker, Elora Hussain, Kristin James, Amy Johnston, Alexandra Kathman, Katie Leander, Elizabeth Martinez, Marcelo Malakooti, Lea Redd, Jennifer Reichek, Enrique Rojas, Erin Rowell, Karen Rychlik, Rei Schenk, Craig Smith, Pamela Spadino, Gretchen Steffenson, Elizabeth Terhorst, Michael Terhorst, Stacey Turner.
Acknowledgments
The authors would like to express their appreciation and gratitude to all the stakeholders who contributed to this work. The author would also acknowledge The Navigate Study Investigators study team for their support in the study.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research and/or authorship of this article: This work was supported through a Patient-Centered Outcomes Research Institute (PCORI) Program Award (CD-1304-6449).