Research Approach and Framework
To answer our research questions, we will use a systematic review of the literature with meta-analysis if appropriate. We have followed the guideline detailed in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) checklist to develop this protocol [23] (Additional File 1). The protocol has been registered on PROSPERO (CRD42022331536). The PRISMA checklist [24] will inform reporting of the final review and the PRISMA-Search (PRISMA-S) checklist has informed reporting of the database and supplementary search methods [25]. A plain language summary of the protocol is presented as Additional
File 2.
Eligibility criteria
We will include all studies that report the effect of trauma-informed care interventions in family caregivers of children with life-limiting conditions on one or more of our selected outcome measures as per Table 1 and Additional File 3. We will include research using all study designs but will evaluate the level of evidence of each study and will stratify our reporting of results by prioritizing the most robust evidence. Quantitative studies without a comparator will be included in a narrative synthesis of results. Qualitative studies, including those without a comparison group, will be included in a thematic synthesis to understand caregivers’ lived experience with trauma-informed care services.
We will include all primary literature meeting our specified inclusion criteria. We will exclude commentaries, study protocols, discussion papers, and book chapters. The authors of related conference abstracts will be contacted to request further information. If no further information is available, the abstract will be excluded. We will not limit to those studies published in English and, should studies in other languages be identified, we will seek the support from collaborators fluent in those languages to translate study findings. Eligibility criteria will be tested through piloting by two team members (PP and IZ) and agreement will be established among all team members prior to commencing the study selection process.
Information sources
A search strategy has been developed by our team of pediatric care researchers and persons with lived experience (e.g., family caregivers of children with life-limiting illness) in conjunction with a health sciences librarian (MM). Comprehensive database searches were conducted in the following databases on May 16, 2022 by MM: Ovid MEDLINE (Epub Ahead of Print, In-Process & Other Non-Indexed Citations, Ovid MEDLINE® Daily and Ovid MEDLINE® 1946-Present), Ovid Embase + Embase Classic (1947 to 2022 May 13), EBSCO CINAHL Plus with Full Text (1981 to present), Ovid APA PsycInfo (1806 to May Week 2 2022), Wiley Cochrane CENTRAL, ProQuest Social Services Abstracts, and ProQuest Applied Social Sciences Index and Abstracts [link to appendix]. A combination of subject headings and text words specific to each database were used to search for the concepts of traumatic stress and family caregivers as well as relevant synonyms. No limits or filters were used. The Ovid Medline search strategy was peer reviewed by a qualified health sciences librarian prior to search translation using the Peer Review of Electronic Search Strategies (PRESS) guidelines [26]. We will also search the reference lists of included studies and identified systematic literature reviews and topical reviews.
Search strategy and study record management
Our search strategy is presented in Additional File 4. The strategy was piloted by: (a) ensuring it captured relevant studies known to our team; (b) initially screening in duplicate the first 200 articles in each database for general article relevance; and (c) screening 500 articles retrieved from a search of a term not included in our strategy to ensure no relevant articles were found in this discrepant search case. We will conduct new searches to capture studies that may have been missed by our original strategy should we uncover other relevant databases or search terms not included in our strategy during extraction and reference list searching.
Data management
Search results were deduplicated in EndNote using an optimized method [27] and were then uploaded to Covidence where additional duplicates were removed. All screening (title/abstract and full-text) will occur in Covidence.
Selection process
Titles and abstracts were independently screened in duplicate by three team members (IZ, PP, and KW). Disagreements regarding study inclusion were resolved through discussion with a third team member with lived experience (BP or CC). Full-text reports appearing to meet eligibility criteria will be located. If a full-text report cannot be located online (including through our access to affiliated libraries), we will attempt to identify the email addresses of abstract authors and contact them for access to the full report. If there is no reply from the authors after two weeks, we will send a follow-up email. If there is still no reply after a subsequent two-week period, we will consider the abstract unavailable. Full-text reports will also be screened in duplicate with discrepancies resolved by a third team member with lived experience. We will record the rationale for excluding each full-text study. Only those studies with a comparator group will be included in a meta-analysis.
Data collection process
We will develop a standardized REDCap-based electronic form with associated coding manual for the abstraction of data from each included study. This form will be developed by our team, piloted in duplicate with 4-5 randomly selected studies, and revised where needed. The form will also include key elements recommended for reporting in qualitative studies [28]—specifically, methodological orientation and theory, analytical process, as well as the presented themes and quotations that relate to our objectives. Two team members will conduct data abstraction and coding of information independently. Disagreements will be resolved through discussion with a third team member with lived experience. Corresponding study authors will be contacted should any uncertainties arise.
Data extraction items
In addition to items listed above, the following data items will be coded for each included study: (a) data source, (b) study country, (c) language, (e) study design, (f) method, (g) source of funding, and (h) sample size. The following characteristics of family caregivers will also be coded: (a) age, (b) sex, (c) gender, (d) race, (e) ethnicity, (f) income, (g) type of caregiver, (h) relationship status, and (i) family makeup (i.e., details of other children). The following characteristics of the caregiver’s associated child will be coded: (a) age, (b) sex, (c) gender, (d) race, (e) ethnicity, (f) diagnosis, (g) treatment regime, and (h) alive or passed.
As per the Template for Intervention Description and Replication (TIDieR) checklist [29], the following characteristics of the trauma- and resiliency-informed care intervention will be collected: (a) type, (b) goal, (c) associated materials, (d) procedure, (e) intervention provider profession, expertise and specific training, (f) location, (g) session frequency, schedule, and duration, (h) any described tailoring of the intervention to individuals, (i) modifications to the intervention during the study, (j) assessments of intervention adherence by caregivers, and (k) assessments of intervention fidelity.
For all quantitative studies we will code the outcome measure used and reported mean and standard deviation values for each outcome (primary and secondary) measured pre- and post-intervention. In the event means and standard deviations are not reported we will collect effect estimates, confidence intervals, test statistics, P-values and individual participant data where reported.
Risk of bias in individual studies
The checklist developed by Hawker et al [30] will be used to appraise the quality of studies in this review. This 9-item checklist was designed to accommodate the appraisal of studies using disparate research methodologies. The tool has been recommended for use in systematic reviews that include studies from varying epistemological and methodologic perspectives [31] and has been successfully applied in systematic reviews such as this one [32]. Two reviewers will independently use the tool to appraise the quality of eligible studies. Any disagreements will be resolved by discussion involving a third reviewer.
Data synthesis
Descriptive statistics (i.e., frequencies and percentages) will be used to summarize and present quantitative data across all included studies. For further analyses, we will plan to make a judgement on the most appropriate data synthesis process depending on the amount and type of available data after data abstraction. Our decision-making process will be transparent and clearly documented in our result dissemination materials.
In the instance that sufficient data are available for quantitative syntheses, we will use random-effects models to combine data across studies using open-source R software. Outcomes will be analyzed using weighted mean differences (with 95% CI) or standardized mean differences (95% CI) if different measurement scales are used. The I2 statistic will be used to quantify heterogeneity. We will use the criteria set out by Higgins et al [33] to interpret the statistic; namely, 0–40% may be considered unimportant, 30%– 60% as moderate, 50%–90% substantial and 75%–100% considerable. If heterogeneity is substantial, we will not perform a meta-analysis and instead use only narrative synthesis.
We will examine potentially important effect size modifiers using meta-regression analysis. Modifiers are based on previous meta-analytical work into the effect of trauma- and resiliency-informed care practices [34] and are (a) treatment modality (individual v. group); (b) intervention focus (prevention v. treatment); (c) length of treatment (in hours); (d) control group type (no treatment v. waitlist v. attention control); (e) outcome measure type (primary v. secondary); (f) quality assessment level of study; and (g) gender (male v. female). We will conduct subgroup analyses if a meta-regression analysis is statistically significant.
In the instance that sufficient data are available for a qualitative synthesis, thematic synthesis methods that have been previously described [35] used successfully in syntheses of qualitative pediatric health research [36] will be used. Copies of identified studies will be uploaded to NVivo 12 software and two reviewers will independently conduct line-by-line coding of the results section of each study, specifically coding direct quotations and descriptions of participants’ comments and behaviors using an inductive approach. As data are coded, a bank of codes will be developed and applied across all studies. Coding choices will be frequently discussed between will all team members, and discrepancies will be resolved using group discussion. Codes will be collapsed, if appropriate, and grouped into themes. Themes produced from the data and relationships among themes will be discussed as a group.
Given our plan to include a broad group of study designs, we will plan to use a narrative synthesis method to bring together all results [37]. Following the methods of previous systematic reviews of pediatric family caregiver health [32], we will use review of tabled text-based data to identify perceived areas of correspondence and relationships between studies and health outcomes.
Meta-bias (publication bias, selective reporting)
We will assess the impact of risk of bias using current and appropriate methods per the Cochrane Collaboration.
Confidence in cumulative evidence
We will use the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach to determine the degree of certainty to place in our quantitative findings [38]. We will use the Grading of Recommendations Assessment, Development, and Evaluation Confidence in the Evidence from Reviews of Qualitative research (GRADE-CERQual) approach to determine the degree of certainty to place in our quantitative findings [39].
Person with lived experience-research leadership collaboration plan
Study leads (BP and LJ) have and will continue to be intricately involved in every aspect of this review. We have collaborated on all decisions related to the review objectives, design, and methods, will co-lead the data collection and analyses phases, and collaborate on all aspects of dissemination. We will continue this collaboration through the following mechanisms: (a) standing bi-weekly meetings (30-90 minutes in length depending on stage of review) to review progress and results, identify and troubleshoot issues, and plan for dissemination and next working steps; (b) ad hoc meetings at any time between scheduled meetings; (c) short conversations via email or text to address emergent issues expediently.
Results dissemination
Our dissemination activities will target three stakeholder audiences: (a) family caregivers, (b) pediatric clinicians, and (c) health care decision and policy makers. We have designed our strategy to accomplish the following dissemination objectives: (a) increase awareness of the impact of medical and associated traumas on the psychosocial and behavioural health of family caregivers, including associated risk and protective factors; and (b) mobilize immediate clinical practice, research, and health care policy solutions to improve trauma-related outcomes for family caregivers. Achieving these objectives will directly and rapidly inform care. The exact nature of hospital practice and policy changes will be informed by our study data and our routine consultations with our family caregiver partners. As a team, we already envision several low-cost solutions that could be readily implemented in Canadian pediatric hospitals and other care settings. Specifically, in alignment with the Marsac et al [9] position paper on implementing a trauma-informed approach in pediatric heath care networks, the following could be implemented: routine risk screening family caregivers to identify those for targeted delivery of evidence-based interventions and training for clinicians in trauma-informed practices.
Through an integrated knowledge translation strategy, the family caregivers and interdisciplinary clinicians embedded in our team will ensure effective and expedient uptake of this research into practice. These stakeholders have expertise in research co-design and have been engaged in designing all aspects of this review. They will be intricately involved in supporting data collection, results interpretation, and dissemination, and will take lead roles in advocacy efforts to support care institution policy change related to the trauma-informed care of family caregivers. To further mobilize our findings we will use the following end-of-study knowledge translation mechanisms - with all co-designed product available in English and French: (a) care institution policy papers for the care sites and government (papers will cover study findings and policy options identified through family caregiver consultation); (b) media releases; (c) 1-page plain language summaries for newsletters and our affiliated organization websites; (d) a social media campaign using bespoke content disseminated through our social networks; (e) presentations at major pediatric conferences and clinician- and family caregiver-focused educational meetings (to review identified results); and (f) open access publications in leading journals to target researchers and practicing clinicians. Further, specific implications pertaining to family caregiver subgroups (i.e., across age, sex, gender, race, ethnicity, etc.) will be highlighted in dissemination efforts to enhance impact.