Discharge communication practices in pediatric emergency care: a systematic review and narrative synthesis

We conducted a systematic review with narrative synthesis following methods outlined by Popay et al. [16]. This type of review explores relationships within and between studies [17]. Techniques are employed to expose the context and characteristics of the included studies to facilitate comparison of similarities and differences across studies [18]. Further details of the review protocol have been published elsewhere [19]. The protocol was registered with PROSPERO (CRD42014007106).

We embedded an integrated knowledge translation (iKT) approach in our review, whereby we met with key knowledge users (e.g., ED clinicians, administrators, parents, and researchers) during each stage of the review to strengthen the relevancy of our research questions and tailor our recommendations. We employed a number of communication strategies to maximize engagement including email and face-to-face and teleconference meetings. Critical reflection of engagement was conducted through detailed documentation of team discussions at all meetings, outlining input from the different stakeholder groups, and underlying rationale for decisions made at each stage of the synthesis process. We also integrated a knowledge user check-in strategy, where at regular intervals during each meeting, we paused to summarize the discussion and solicit questions. Three key meetings included (1) a teleconference to refine the research question, (2) a face-to-face meeting to discuss the results of the preliminary analysis, and (3) a teleconference meeting to discuss findings and draft recommendations. Finally, authors from three of the included studies (all ED physicians from academic teaching hospitals) were contacted by email and invited to provide feedback on the preliminary findings. Expert feedback was provided through a brief telephone interview lasting 15–30 min.

We used the Behaviour Change Wheel (BCW) [20], a theory-based method for characterizing and designing behavior change interventions, as a guiding framework to examine the discharge communication interventions in the included studies. Pairs of two reviewers (KM, MB, AG, or JAC) independently coded the narrative descriptions of all interventions. A directed content analysis approach was used to classify intervention description according to the nine intervention function types (i.e., the mechanism by which the intervention is proposed to function) of the BCW [20]. Reviewers met to compare consistency after coding the first and third study and then after every five studies. Differences in coding were discussed to achieve consensus.

We also coded the 75 included studies to identify barriers and enablers to intervention implementation and effectiveness as reported by the study author. Full text articles were loaded into NVivo (QSR International) and were coded by two reviewers (JAC and AG) [21]. Reviewers met throughout the process to ensure consistent coding of barriers and enablers and to discuss sections that were challenging to code.

The search strategy was developed in consultation with a library scientist and peer reviewed by a second library scientist from the Alberta Research Centre for Child Health Evidence (ARCHE; University of Alberta) prior to being implemented in three databases from their date of inception to July 7th, 2017: CINAHL, Medline, and Embase. Please see Additional file 1 for the list of all search strategy terms. We also hand-searched relevant emergency pediatric journals for articles published between January 1st, 2009 and October 31st, 2018: Annals of Emergency Medicine, Academic Emergency Medicine, Pediatric Emergency Care, Journal of Emergency Medicine, and Journal of Emergency Nursing. Additionally, a further search was conducted by emailing the ED administrators of the 15 Pediatric Emergency Research Canada (PERC) sites and the Chairs of the Pediatric Research in Emergency Departments International Collaborative (PREDICT), Pediatric Emergency Medicine Collaborative Research Committee (PEM-CRC), Pediatric Emergency Care Applied Research Network (PECARN), and the Paediatric Emergency Research in the United Kingdom & Ireland (PERUKI) and reviewing relevant national and international websites to identify existing discharge communication policies and procedures that might be relevant in a pediatric ED context.

Studies were included if they described or evaluated changes in the structure or process of care in the ED to enhance discharge communication as their primary objective. Quantitative, qualitative, and mixed methods study designs were eligible for inclusion. Studies were excluded if interventions took place outside of the ED or primary outcomes were not relevant to discharge communication. The published protocol paper for this review provides further details on the process of selecting studies [19]. Detailed inclusion and exclusion criteria for studies and abstracts can be found in Additional file 2.

Data were extracted by teams of two independent reviewers using a customized data abstraction form and discrepancies were assessed by JAC. Full data extraction included study design details (i.e., year of publication, country, type of ED), detailed description of the intervention (i.e., description of individual components, reported use of theory or assumptions about causal mechanisms supporting the different components, duration, and timing of the intervention), implementation strategies (i.e., training or instructions provided to participants, timing), participant details (i.e., age of the child, illness presentation, parent characteristics, health care provider characteristics), and author reported factors and/or processes identified as impacting implementation. Our primary outcome of interest was any change in process, parent, healthcare provider, or patient outcome related to the discharge communication intervention. We extracted details about authored reported primary outcomes (i.e., timing, measures, target of the intervention). Data was managed using excel and intervention descriptions were exported to NVivo 10 (QSR International) for analysis [19].

Study quality was assessed at the study level by two independent reviewers using critical appraisal checklists from Joanna Briggs Institute (JBI) [22]. Each study was appraised using the appropriate checklist for the study design. JBI checklists are designed to assess the quality of a study’s methodologies and outcomes, and to identify potential sources of bias or confounding variables within the study [22]. Options for checklist responses included “yes,” “no,” “unclear,” or “not applicable.” Each reviewer used the JBI definitions and checklists and provided evidence from the article to support the reasoning for their scoring. At two points in the appraisal process, the reviewers met to meet consensus and resolve discrepancies. A third reviewer was consulted when discrepancies could not be resolved.

The primary aim of this review was to understand how and why discharge communication interventions and processes work in a pediatric ED context. To address this aim, we employed a number of data synthesis strategies proposed by Popay et al. [16] to examine the individual study findings. We developed summary tables, which outlined study design, clinical context, intervention components, intervention target, quality appraisal, outcome measures, and direction of effect. Findings of included studies did not lend themselves to quantitative analysis due to the wide range of intervention descriptions and outcome measures. We also developed structured textual summaries of all included studies that outlined primary objectives, details regarding context and setting, descriptions of any interventions and implementation strategies, relevant outcomes, and barriers reported by study authors. Content analysis was carried out on all extracted data to assist with identifying importing themes and gaps in the existing evidence.

Study quality was assessed using critical appraisal checklists from the Joanna Briggs Institute (JBI) [22]. The majority of the cross-sectional studies were of moderate quality (6/8 appraisal criteria present). The two appraisal questions that had mixed results were related to identifying and addressing potential confounding. Only a third of the studies (n = 9; 32%) identified possible confounders and even fewer studies noted how these confounders were adjusted for during data analysis (n = 8; 29%).

The 24 RCT were appraised using the Checklist for Randomized Controlled Trials and 46% of the studies (n = 11) did not clearly identify if a true randomization method was used. Additionally, only four RCT clearly stated how allocation concealment was used during the study [60‐63].

The Checklist for Quasi-Experimental Studies was used for 11 studies in the review. These quasi-experimental studies (n = 11) scored consistently well across all of the appraisal checklist criteria, with the exception of concerns about composition of the comparison groups. We found that just over half of studies (n = 6; 55%) had at least one or more key variables with a 10% difference or more between the study groups. These differences raise some concerns about potential selection bias.

Finally, the three included qualitative studies were assessed using the Checklist for Qualitative Research. Deficits were identified in all studies regarding reporting of the philosophical perspective guiding the study methods and clearly describing the role of the researchers within the studies.

We identified 44 discharge communication interventions among the 75 included studies in the review (Table 2). Interventions were comprised of one to four intervention function types, according to the BCW (Table 3) [20]. The heterogeneity of intervention descriptions and outcome measures limited our ability to carry out meta-analysis.

The majority of interventions were delivered to children and parents by clinicians [27, 32, 42, 43, 49, 54, 60, 61, 67, 68, 71, 74, 76, 77, 81, 83, 86, 87, 89, 90, 92]. Various clinicians were involved in the implementation of interventions including ED clinicians [27, 42, 49, 54, 60, 61, 65, 68, 75, 80, 81, 83, 86, 87, 92], medical students/residents [67, 74], primary care providers (PCP) [82, 85], nurses [32, 43, 71, 84], pharmacists [61], and respiratory therapists [77]. Additionally, one study utilized a discharge facilitator to provide specific discharge communication to parents for gastroenteritis [76].

Most studies provided few details about the timing of intervention implementation, typically referencing that it was provided during the ED stay [60, 61, 63, 64, 66‐68, 73, 74, 76, 80‐92]. Thus, it was difficult to determine the exact point during the care process the intervention was delivered and the duration of the intervention. When specific timing of the interventions was indicated, it was either described as at the time of discharge [63], upon ED admission [77], directly following triage [65], or after the child was evaluated by an ED physician [69].

Conceptual mapping is a useful technique for exploring relationships in the data of narrative synthesis [98]. Using a consensus approach, three members of the research team reviewed the narrative summaries of all included studies during two half-day meetings and identified several key concepts and relevant terms to describe how discharge communication worked or did not work for children and their parents in ED settings (Fig. 2). The final schematic depicts the key barriers and enablers for discharge communication across all included studies. Key concepts include ED context, knowledge, attitudes and beliefs, ED healthcare providers, child and parents, intervention, education, and outcomes. Important linking terms include health literacy, timing, illness severity, readiness to learn, rapport, and training.

Barriers and enablers to successful discharge communication were identified at a number of levels, including the intervention, healthcare providers, child/parent, and the context in which care is delivered. The majority of interventions targeted parents and included education as a core strategy. Although not explicitly stated, the educational interventions in the studies appear to operate on the assumption that imparting information will improve health care provider or parent knowledge and subsequently change behaviors. Few studies used theory to guide design of the intervention [71, 80, 81, 83, 84, 89, 90, 95], and there was no formal assessment of potential barriers to changing parent, child, or health care provider behaviors. This made it difficult to understand how the interventions were expected to improve outcomes. Barriers related to intervention design (language, literacy level, content, readability) and delivery (timing during the visit, duration, fidelity, mode, extent of interactivity) were identified as important factors across a number of studies [26, 55, 57, 60, 65‐67, 74, 78, 84, 86, 89, 99]. While there was variation in who delivered the interventions, it is important to highlight that in a number of the experimental trials, the interventions were delivered by individuals who were not regular members of the ED team [33, 51, 59, 63, 64, 66, 69, 70, 72, 73, 79, 82, 85, 88]. A collaborative approach to co-designing interventions was proposed by some authors as a possible strategy to enhance implementation [33, 40, 49, 70].

A number of barriers and enablers related to the patient and parent were reported including complexity of the patient presentation [53, 57, 69, 75, 84], first language and health literacy [35, 42, 46, 52], expectations of care [25], and past experience [66, 71]. Parents’ health literacy was considered an important factor in both the design [54, 63, 67] and the mode of delivery [46, 50, 57, 65] of the intervention. Parental stress was also considered an important influence on parents’ readiness to engage in discharge communication [26, 53, 69]. Actively engaging parents in the discharge process throughout their ED visit [27, 78, 81] and presenting information using multiple delivery modes [33, 51, 68, 75, 80, 83, 96] were thought to be important enablers. Parent decision making about the nuances of follow-up care post discharge can be complex [94]. It was also recognized that parents leverage a number of external knowledge sources including family and non-healthcare workers when making decisions about care for their child at home following an ED visit [52, 64, 65, 97].

Several studies suggested that ED healthcare providers recognize discharge communication as an important aspect of their everyday professional practice for both clinical and medicolegal reasons [23, 32], but there appears to be a fuzzy boundary between patient education and the provision of discharge information [70]. Discharge communication practice variation was common across a number of studies [30, 31, 100]. Healthcare provider training [23, 27, 28, 55, 74, 77, 86], beliefs and attitudes about discharge communication [40, 41, 57], and limited use of medical jargon [53, 96] were identified as important factors influencing success. Rapport between healthcare providers and parents was also considered an important factor contributing to successful discharge communication [69, 81, 95]. Although we did not identify any interventions that focused specifically on building rapport between providers and parents, the quality of the interaction between providers and patients/parents was highly valued as contributing to successful discharge communication [34, 46, 60, 69, 76, 81, 84, 88, 97].

The context of the ED including factors related to managing multiple patients in a compressed period of time and emphasis on patient flow can pose unique barriers to discharge communication [30, 55, 92]. Approaches to address some of these contextual barriers include a dedicated space for teaching, easily accessible resources, and the inclusion of strategies that reduce the time burden on ED providers and can easily be incorporated into the busy workflow of the ED [55, 79, 80, 86, 96]. However, Macy et al. (2011) noted that even the use of video education can be hindered during times of high patient volumes [65]. Ideal structures support discharge communication that is consistent and comprehensive yet succinct and relevant [23, 44, 97]. ED team practice norms regarding discharge communication [42, 78] and roles in the ED specifically dedicated to discharge communication may assist with consistent and reliable delivery [27, 39, 76].

Although there was wide variation in primary outcomes across studies, parent knowledge, comprehension, and recall were deemed important outcomes in a number of studies [33, 51, 59, 63, 65, 67, 68, 70‐72]. Unfortunately, there was little consistency in how knowledge, comprehension, and recall were measured across these studies.

We received a total of 36 responses to our email surveys (Canada = 17; Australia = 13, USA = 6) with no relevant pediatric ED discharge communication policies or guidelines identified. Following our search of the gray literature, three documents were discovered that included either policies or guidelines regarding the provision of discharge communication [101]. A joint policy developed by the American Academy of Pediatrics and the Committee on Pediatric Emergency Medicine recommended the provision of discharge communication within a family-centered care approach, [101]. Generally, this guideline contained very little direction in the way of discharge content and processes, and it is unclear if it translates into active policy development [101]. The Scottish Intercollegiate Guidelines Network (SIGN) has produced a generic discharge document tool that can be implemented to guide the discharge process [102]. The discharge document provides guidance on important content that should be communicated at discharge, including primary discharge diagnosis, presenting complaint, significant procedures, new medications, and follow-up instructions [102]. However, no specific recommendations are made about how the SIGN discharge document should be generated, updated, and used in ED clinical practice [102]. Finally, the National Guidelines Clearinghouse contained one active policy that is currently utilized by affiliated children’s hospitals in the USA. This policy outlines standards for discharge communication as the primary responsibility of the ED nurse and that all patients are to be discharged home with explicit plans for follow-up care, a summary of the visit, appropriate instructions for care, and prescription medications [103]. Prior to discharge, it is the responsibility of the nurse to verify that all health care concerns have been addressed, parents are aware of when to return to the ED, documentation is complete, and medications have been reviewed with the patient [103].

This review followed established review methodologies including a comprehensive search of the published and gray literature, assessment of risk of bias, and established narrative synthesis strategies to explore relationships across a range of study designs. We incorporated an iKT approach to ensure our findings were relevant to stakeholder needs.

Due to the heterogeneity of the interventions and outcomes, we were unable to perform a meta-analysis. While we used an established behavioral framework to guide our examination of discharge communication interventions, our classification of the included interventions was limited to the details provided in the study reports. It is possible that the interventions contained additional functions that were not captured in our analysis because they were not reported by the study author.