Key Points

Admission to the hospital can be a challenging time for children and families because of the physical effects of serious illness and the psychological effects of illness and hospitalization.1 There are also economic strains on families because of work absences and other indirect costs of hospitalization (eg, transportation).² In Canada and the US, most children cared for in hospitals have medical conditions managed in general pediatric inpatient units (GPIUs).^3,4 Children admitted to GPIUs may have been previously healthy and are hospitalized for an acute, common illness (eg, gastroenteritis) or have chronic conditions (eg, asthma) and are hospitalized for an acute, common illness or exacerbation of the chronic disease. The specialty dedicated to caring for these hospitalized children in the GPIU is pediatric hospital medicine. Over the past decade, seminal research has contributed to improvements in the care of hospitalized children in select areas, such as bronchiolitis, and handoff bundles (ie, medical handover).^5–8 Yet, compared with other hospital-based clinical areas, such as oncology⁹ and critical care,^10,11 there remains a relative lack of high-quality research in pediatric hospital medicine, particularly robust randomized clinical trials.^12,13

Although there have been health systems level analyses identifying high-priority conditions for research in pediatric hospital medicine based on volume, costs, and variation in care in the US and Canada,4^,14–16 little is known about which topics should be prioritized for research from the perspective of patients, families, and clinicians. Prioritization processes can help minimize avoidable research waste¹⁷ by mitigating the mismatch between research that is conducted and research that is relevant to patients and clinicians.¹⁸ National research bodies, such as the Patient-Centered Outcomes Research Institute in the US¹⁹ and the Strategy for Patient-Oriented Research in Canada,²⁰ suggest engaging patients, caregivers, and stakeholders in defining research priorities. James Lind Alliance (JLA) Priority Setting Partnerships (PSPs) bring together patients, caregivers, and clinicians to identify important unanswered questions in clinical areas. Priority Setting Partnerships have been used for a range of conditions and care settings,²¹ such as emergency care²² and intensive care.²³ The JLA method facilitates high levels of patient engagement, with patients involved in all steps of the research process.^21,24 Engaging patients and families as coresearchers has been reported to have several benefits, including increased response rates,^25–29 credibility of results,^26,27 credibility of dissemination,²⁹ and quality of results and rigor of research.^30,31

To establish meaningful relationships and maximize research results, patients and caregivers should be engaged early in research design, particularly in the research question conceptualization phase.30 To our knowledge, no previous studies have identified the most important research questions in pediatric hospital medicine from the perspective of patients, caregivers, and clinicians. Patient and family involvement in pediatric research prioritization studies have focused on preventive care,³² children with neurodevelopmental problems,³³ emergency care,³⁴ and patient safety in the hospital.³⁵ Therefore, the primary aim of our project was to conduct a research priority setting study using a JLA PSP to identify the most important unanswered clinical management questions in pediatric hospital medicine that are meaningful to patients, caregivers, and clinicians.

We conducted a JLA PSP36 using a modified Delphi technique along with a nominal group technique from March 1, 2020, to September 28, 2021.³⁷ Our protocol is available in eMethods 1 in Supplement 1. The study followed the Reporting Guideline for Health Research Priority Setting with Stakeholders (REPRISE). The project was reviewed by the Hospital for Sick Children’s Research Ethics Board and deemed to not require approval because its activity fell in the domain of preresearch/information gathering. The entire study, including each survey and the workshop, was explicitly about identifying research questions to be used for future research projects.

The research priority setting study focused on clinical management questions relevant to the care of children (age 0-18 years) hospitalized in a GPIU in Canada. We used the American Academy of Pediatrics definition of pediatric hospital medicine: “patients with acute and/or serious complications of common problems, multiple comorbidities and/or injuries, complex chronic diseases, acute mental health problems, special health care needs, technology-dependent conditions, and those needing palliative care.”38 Clinical management was broadly defined as the diagnosis and treatment of conditions, including processes of care (eg, how to promote better sleep in the hospital) as well as specific conditions (eg, interventions for asthma). The PSP focused on conditions in which clinical management was typically led by general pediatricians and did not distinguish between GPIUs located in a children’s hospital or community hospital.

Multiple recruitment approaches were used to ensure the diversity of respondents, including targeting vulnerable groups, specifically Indigenous communities and high social risk groups. We developed a video explaining the priority setting process,39 ensured that the graphics in communication materials captured a diverse range of patients and parents/caregivers (eg, graphics included individuals using a wheelchair and those with different ethnicities), used colors that met accessibility standards, and provided a paper version of the survey to facilitate participation of those who did not have access to online technology.

Anonymous study data were collected and managed using Research Electronic Data Capture (REDCap),40 hosted at The Hospital for Sick Children. The survey (eMethods 3 in Supplement 1) asked participants the following question: What concerns, comments or questions do you have about the care of children in the hospital on the general pediatric ward that you would like answered by research? There were also optional demographic questions on occupation, age, gender, ethnicity, place of residence (province/territory), and setting of residence (urban/rural) to track respondent types and target dissemination efforts toward underrepresented groups. Descriptive statistics were used to summarize participant demographic characteristics and survey responses.

Phase 1 survey data were analyzed in Microsoft Excel, version 15 (Microsoft Corp). Following removal of out-of-scope questions, the raw questions were reviewed iteratively by the PSP coordinator (D.A.) with regular discussions with co-chairs of the steering group (P.J.G. and K.C.). Each response was coded into broad categories, developed iteratively after reviewing the responses several times. Similar or duplicate questions were combined where appropriate. To assist, steering group members were paired into dyads comprising clinicians and youths and/or parents/caregivers; each dyad was responsible for reviewing up to 5 categories of responses and tasked with forming summary questions. The steering group provided oversight to ensure appropriate interpretation of the raw data through open discussion in group meetings to ensure that questions were understandable to all audiences. Examples of raw data for all final prioritized questions can be found on the JLA website.41

Highly focused and targeted literature searches were conducted to determine whether the summary questions had already been answered. Given the broad scope, questions were considered unanswered if there was no systematic review, if a recent systematic review indicated insufficient evidence, or if there was insufficient evidence outlined in position statements from the Canadian Paediatric Society, American Academy of Pediatrics, or National Institute for Health and Care Excellence. The searches were conducted by an information specialist (M.R.A.) who worked with an experienced librarian (Q.M.); a detailed description of the search strategy is available online (eMethods 2 in Supplement 1).⁴² To ensure retrieval of the most recent literature, searches were restricted to 2010 onward, were limited to material published in English, and were conducted from December 13, 2020, to April 4, 2021. Questions deemed to be answered by earlier research were categorized as answered. The remaining unanswered questions were brought forward to phase 2. Initial categorization of the evidence was completed by one of us (M.R.A.) and was then reviewed (P.J.G.); final decisions were made by consensus, with input of the steering group as required.

Two half-day (August 18-19, 2021) virtual (Zoom Technologies Inc) workshop sessions were held to generate consensus on the top 10 unanswered questions in pediatric hospital care. In-person meetings were not possible because of the COVID-19 pandemic; therefore, the JLA adapted its workshop delivery method to be conducted online, and learnings from other JLA PSPs were applied.43 The steering group, with guidance from the JLA, recruited 24 youths, parents/caregivers, and clinicians from across Canada for the workshop. Workshop documentation and guidance was sent to participants 3 weeks before the workshop. Childcare reimbursement was offered, and compensation was provided to patients and/or caregivers according to published guidelines.²⁰

The workshop was chaired by one of us (K.C.) and used a modified nominal group technique,44^,45 which is a well-established structured, multistep facilitated group meeting technique. Three additional JLA facilitators managed the small-group discussions to ensure balanced contributions from youths, parents/caregivers, and clinicians. We evaluated the extent of patient engagement in the final workshop via a survey.

The phase 1 survey was completed by 188 participants (Table 1), with 54 responses (29%) from those with lived experience and 134 responses (71%) from clinicians. Most respondents were female (148 of 167 [89%] vs 17 of 167 [10%] male]; mean [SD] age, 39.5 [12.4] years). Self-reported ethnicities were African (n = 3), Asian (eg, Middle East and South, East, and Southeast Asian; n = 28), Caribbean (n = 2), European (n = 5), Indigenous (First Nations, Metis, Inuit; n = 2); Latin, Central, or South American (n = 3); White (North American or European; n = 136); and prefer not to say (n = 5) or other (n = 2). There was representation from most regions of Canada, primarily Ontario (86 of 167 [52%]).

In total, the 188 participants provided 495 questions/statements. Of the 495 responses, 58 were deemed out of scope, for example, because they focused on surgery, intensive care, or resource allocation (Figure). The remaining 437 questions/statements were sorted into 22 categories (eTable 1 in Supplement 1); the 4 categories with the largest number of responses were treatment (n = 73), communication (n = 56), management (n = 37), and family experience and support (n = 37). With steering group oversight, 75 summary questions were drafted.

Of the 75 questions, 4 were deemed to have sufficient evidence to mark as answered (Table 2). The steering group reviewed the remaining 71 questions and identified 21 that were supported by a single respondent (eTable 2 in Supplement 1); these 21 questions were discussed by the steering group to ensure that they did not represent issues of concern to groups who were underrepresented in the survey and then were pragmatically removed to make the phase 2 survey more manageable for respondents. Of the original 75 questions, 50 went forward to phase 2; all 71 questions are listed in eTable 3 in Supplement 1, with correspondence evidence identified in eTable 4 in Supplement 1.

In total, 201 participants completed the phase 2 survey (Table 1), with 44 responses (22%) from those with lived experience and 157 responses (78%) from clinicians. Most respondents were women (165 of 186 [89%] vs 19 of 186 [10%] men; mean [SD] age, 40.0 [11.0] years). Self-reported ethnicities were African (n = 4), Asian (eg, Middle East and South, East, and Southeast Asian; n = 21), Caribbean (n = 5), European (n = 6), Indigenous (First Nations, Metis, Inuit; n = 4); Latin, Central, or South American (n = 1); White (North American or European; n = 165); and prefer not to say (n = 5) or other (n = 2). Most participants were from Ontario (92 of 183 [50%]). European was included as a separate category from White (North American or European) for those who identified as European but not White.

Survey responses were sorted by overall rank, lived experience responses, and clinician responses (Table 3). The steering group reviewed the overall ranking and unanimously decided to include all questions that fell into either group’s (lived experience and clinician) top 10 items for the workshop. As such, a total of 16 questions were brought forward for discussion in phase 3 (Figure).

The workshop involved 24 participants: 12 clinicians (4 pediatricians, 5 nurses, 2 child life specialists, and 1 occupational therapist), 10 parents/caregivers, and 2 youth patients from across Canada. In addition, 4 experienced JLA PSP facilitators (including K.C.) and 3 observers were present. The workshop participants engaged in thought-provoking discussions over 2 half-day sessions before reaching consensus on the top 10 unanswered research questions (Table 4).

Using the approach recommended by national research bodies,19^,20 this PSP brought together youths, patients, parents/caregivers, and clinicians to identify the top 10 research priorities for pediatric hospital medicine. To our knowledge, this project is the first to examine major knowledge gaps related to the in-hospital care of children by integrating the perspectives of those with lived experience and clinicians. By involving patients and parents/caregivers in the process, we provided a voice to important stakeholders, including those who may only have a brief stay in the hospital or who do not fall within a specific disease entity for research. The top 10 research priorities lay the foundation for researchers and research networks to focus on improving outcomes in pediatric hospital medicine. The long list of 71 questions (eTable 3 in Supplement 1) identifies several important clinical areas in need of further research.⁴¹

Using the Krumholz framework of outcomes research,46 most of the top questions focus on the patient perspective, rather than comparative effectiveness or health systems improvement. Although there were specific research questions identified in the survey (eg, effective use of heated high-flow nasal canula; eTable 3 in Supplement 1), the top research questions were quite broad compared with other PSPs, particularly those addressing a specific condition or clinical area,^33,47–49 which reflects the comprehensive clinical area covered. Our scope encompassed clinical management, which comprised diagnosis and treatment, including processes of care (eg, insertion of an intravenous cannula) and specific conditions (eg, bronchiolitis). Other PSPs that have focused on broad areas, such as occupational therapy,⁴³ intensive care,²³ or emergency medicine,²² had similarly broad questions yet were able to generate focused research questions that were funded.^50,51 For example, a top unanswered question in intensive care was on supportive families after discharge, leading to a multicenter longitudinal study exploring the outcomes of children and families in the first year after discharge from intensive care.⁵¹ The themes identified in the top 10 questions are a salient reminder that families are less concerned about minor differences regarding which test is ordered or which treatment is given compared with whether shared decision-making was conducted⁵² or how supported they felt during hospitalization.⁵³ These sentiments were not only highlighted by those with lived experience but were also ranked highly by clinicians given the significant overlap in phase 2 survey rankings (Table 3). Although many questions are broad, this implies that subsequent research can potentially have a large influence on clinical care. For example, research to increase communication between families and clinicians or on integrated shared decision-making can be broadly applicable to multiple conditions in the hospital.

The top research priority focused on best practices and care models for hospitalized children with medical complexity. These children often have functional limitations, substantial family-identified needs, and a reliance on technologies for activities of daily living.54 Diskin et al⁵⁵ conducted a 3-stage modified Delphi study, including clinicians and family caregivers, to identify research priorities for children with neurological impairment and medical complexity. Most top-ranked questions focused on the management of specific clinical issues, such as feeding and irritability, and none focused on in-hospital care. There are knowledge gaps on the optimal models of care (eg, primary inpatient care, consultative or episode-based care model)⁵⁶ and on how best to support parents/caregivers⁵⁷ addressing, for example, whether in-hospital care should be led by specialized complex care teams, with or without trainees; the role of the outpatient clinicians in the inpatient setting; whether a different care model should be implemented at children’s vs community hospitals; and how family perspectives on patient care models should be evaluated.

Specific populations were explicitly addressed in 2 of the other top 10 questions. The first focused on how to meet the needs of children and youths with developmental disabilities on the GPIU. Although there is an overlap with children with medical complexity, final workshop participants felt strongly that children and youths with developmental disabilities face unique issues in the hospital. One earlier Canadian JLA PSP identified research gaps for children with neurodevelopmental disorders, but primarily focused on outpatient care.58 There were concerns about the lack of hospital-friendly initiatives or resources for these children (eg, strategies to support a child with autism spectrum disorder admitted with an asthma exacerbation while using inhalers). The second focused on Indigenous parents, families, and children and youths—a community whose children have disproportionately higher hospitalization rates—compared with children of other ethnic groups.⁵⁹ Despite this increased burden, few studies have examined the care and outcomes of Indigenous children on the GPIU or evaluated how to ensure meaningful participation of Indigenous communities in health research. Given the failure of past research endeavors, including causing harm,⁶⁰ initiatives to improve care in this population will require building reciprocal, trusting relationships with Indigenous stakeholders.

The outcome associated with hospitalization was another important theme. Prioritizing alternatives to shorten the length of hospital stay reflects that 65% of all hospital encounters in children are at community and general hospitals.3^,4 Although there has been extensive research in adult care about alternatives to inpatient care, such as hospital-at-home interventions,⁶¹ there are limited examples in pediatrics, such as home phototherapy for neonatal jaundice⁶² or home oxygen therapy for bronchiolitis.⁶³ There was also recognition of the need to mitigate the outcomes of prolonged inpatient hospitalization, such as excessive screen time for children and separation from families.⁶⁴ The mental health outcome of hospitalization was also a prominent discussion point in the final workshop because the COVID-19 pandemic has exacerbated mental health concerns among children and youths as well as parents/caregivers. A priority was family support, such as counseling, recognizing the influence of parental mental health on children.

Strengths of our project include applying the JLA PSP approach, which is a well-used, rigorous, validated methodological approach to identify research priorities. We also had sufficient representation from a range of health care professionals from across Canada and had 100% attendance at the final workshop. The number of respondents to both surveys was consistent with several other JLA PSPs.47^,49 The top priorities for pediatric hospital medicine are also likely relevant for hospitalized children in other care settings, such as in oncology or surgical inpatient units. Furthermore, the priority setting method can inform other research groups who want to identify priorities for similarly broad care settings, such as adult internal medicine.

The project has limitations. First, the scope of the PSP was broad, which introduced challenges in communicating relevance to families in the survey. The project also required that the evidence-checking progress was restricted to systematic reviews and guidelines that may have missed important, high-quality research related to the summary questions. Second, there was limited representation from allied health care workers from community hospitals in both surveys, and from individuals who work in rural and remote regions. Third, although demographic information on survey participants was optional, most of the participants identified as White (North American or European), with few participants identifying as Indigenous. Despite this limitation, the third ranked question was focused on Indigenous parents, families, and children and youths. There were also few respondents who identified as Latin, Caribbean, and African. Fourth, the top 10 questions reflect broad areas of clinical care and may prove challenging to address with research. However, early work with our research network65 has already identified multiple specific questions to be pursued.

Published: April 26, 2022. doi:10.1001/jamanetworkopen.2022.9085

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2022 Gill PJ et al. JAMA Network Open.

Corresponding Author: Peter J. Gill, MD, DPhil, Peter Gilgan Centre for Research and Learning, 686 Bay St, Toronto, ON M5G 0A4, Canada (peter.gill@sickkids.ca).

Author Contributions: Dr Gill and Ms Cowan had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Gill, Bayliss, Sozer, Breen-Reid, De Castris-Garcia, Green, Frappier, Cowan, Arafeh, Parkin, Cohen, Mahant.

Acquisition, analysis, or interpretation of data: Gill, Buchanan, Green, Quinlan, Wong, Frappier, Chan, Arafeh, Anwar, Macarthur, Parkin, Mahant.

Drafting of the manuscript: Gill, De Castris-Garcia, Quinlan, Chan, Mahant.

Critical revision of the manuscript for important intellectual content: Gill, Bayliss, Sozer, Buchanan, Breen-Reid, De Castris-Garcia, Green, Wong, Frappier, Cowan, Arafeh, Anwar, Macarthur, Parkin, Cohen, Mahant.

Statistical analysis: Gill, Mahant.

Obtained funding: Gill, Mahant.

Administrative, technical, or material support: Gill, Bayliss, Sozer, Breen-Reid, De Castris-Garcia, Green, Quinlan, Frappier, Cowan, Chan, Arafeh, Anwar, Macarthur, Mahant.

Supervision: Breen-Reid, Mahant.

Conflict of Interest Disclosures: Dr Gill reported receiving grants from Physicians’ Services Incorporated Foundation and Hospital for Sick Children, expenses reimbursed to attend conferences as a member of the steering committee from EBMLive, and expenses reimbursed to attend meetings as a member of the institute advisory board from the Canadian Institute of Health Research (CIHR) Institute of Human Development, Child and Youth Health outside the submitted work. Dr Parkin reported receiving grants from the Hospital for Sick Children Foundation and the Canadian Institutes of Health, and nonfinancial support from Mead Johnson Nutrition for a trial on oral iron supplement outside the submitted work. Dr Cohen reported receiving advisory fees from the Ontario Ministry of Health’s Committee to Evaluate Drugs on drug-related policy issues outside the submitted work. Dr Mahant reported receiving grants from Physician Services Incorporated Foundation during the conduct of the study; grants from the CIHR, and fees as a member of the advisory board from the Journal of Hospital Medicine outside the submitted work. No other disclosures were reported.

Funding/Support: The study was supported by the CIHR Catalyst Grant in Patient-Oriented Research (grant PAO-169399).

Role of the Funder/Sponsor: The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: The members of the Canadian Pediatric Inpatient Research Network are listed in Supplement 2.

Disclaimer: The opinions, results, and conclusions reported in this article are those of the authors and are independent from the funding sources.

Additional Contributions: We acknowledge the participants of the surveys and the final workshop participants for their contribution and time. Lily Ren, MI (Stanford University) and Quenby Mahood, MI (Canadian Agency for Drugs and Technologies in Health) assisted with the evidence checking; they did not receive compensation. Suzannah Kinsella, Maryrose Tarpey, and Tamara Rader were JLA facilitators and managed small-group discussions. They received financial compensation for their role. The following individuals helped acquire grant funding and disseminate study results: Kate Langrish, RN (The Hospital for Sick Children); Judy Van Clieaf, RN, MN (The Hospital for Sick Children); Andreas Laupacis, MD (Li Ka Shing Knowledge Institute, Unity Health Toronto); David Johnson, MD (University of Calgary); Andrea Tricco, PhD (Li Ka Shing Knowledge Institute, Unity Health Toronto); and Frank Gavin, MA (CHILD-BRIGHT National SPOR Network). None of these individuals was compensated for their work.