Transforming Professional Identity in Simulation Debriefing

Summary Statement: There continues to be a lack of detailed understanding of how debriefing works and how it enables learning. To further our understanding and simultaneously illuminate current knowledge, a metaethnographic qualitative synthesis was undertaken to address the research question: how are interactions in simulation debriefing related to participant learning? Ten databases were searched (up to November 2020) and 17 articles were selected for inclusion. Initial interpretive synthesis generated 37 new concepts that were further synthesized to produce a new theoretical framework. At the heart of the framework is a concept of reflective work, where participants and faculty recontextualize the simulation experience bidirectionally with clinical reality: a process that facilitates sensemaking. This occurs in a learning milieu where activities such as storytelling, performance evaluation, perspective sharing, agenda setting, and video use are undertaken. The outcome is conceptualization of new future roles, clinical competence, and professional language development—a process of transforming professional identity.


INTRODUCTION
Debriefing is a critical element of simulation-based learning with good evidence that this is related to measurable outcomes in participants. 1The description of various debrief approaches or models 2 dominates the simulation debriefing literature, with a corresponding emergence in methods to assess the quality of debriefing. 3The structure of these models is often a multiphase approach, 2 with a focus on participant reflection during an analysis phase.This approach is grounded in experiential learning theory 4,5 and the often implicit epistemological underpinning that this process of reflection, where perspectives are transformed, results in new knowledge. 6ebriefing remains almost universal in use yet relatively underresearched, with a lack of detailed understanding of how debriefing works and how it enables learning. 7[10][11] Although these factors are products of learner-faculty interactions, there is limited empirical research on the nature of these interactions: what are these interactions, how do these interactions occur, and what relation do they have to participant learning?By understanding more about how debriefing works and seeking answers to these questions, we can refine debriefing practice, benefitting learners, and improving patient care.We can also lay the groundwork for the next generation of research on simulation debriefing.
Why is a synthesis appropriate for developing our understanding of simulation debriefing?
The existing research that has examined interaction in simulation debriefs is diverse in nature and published across different research disciplines.This makes holistic interpretations of debrief interaction challenging.A qualitative synthesis is designed to further develop understanding in a field and to simultaneously illuminate current knowledge, making it ideal for simulation debriefing.This synthesis aims to: a. Identify and share what is currently known about interactional factors and their relationship to knowledge; b. generate new knowledge regarding debriefing practice; c. understand the relationship between debriefing evidence and practice; and d. identify knowledge gaps where future research could be beneficial.
Although numerous qualitative synthesis approaches are available (see 12 for an overview of approaches), a metaethnographic approach is particularly suited to address the aforementioned points: the process is deeply interpretive and theory-generating and will provide a comprehensive collective understanding.The data of interest in metaethnography are the interpretations of authors of selected studies (defined as second-order constructs); these undergo interpretative synthesis in a process known as reciprocal translation.This synthesis is therefore conducted in accordance with Noblit and Hare's 7-phase metaethnographic approach, 13 with specific adaptations to ensure maximum rigor and research process transparency.

METHOD
The 7 phases of metaethnography are divided into related parts as detailed in Table 1.Established frameworks, including adaptations where necessary, are used throughout the phases to report the metaethnographic process we have undertaken.
Phase 1 is primarily focused on research initiation, study registration, and research question definition.The specific research question is to understand: how are interactions in simulation debriefing related to participant learning?The question has been deconstructed in the population, intervention, comparison, outcome, and setting (PICOS) framework (Table 2).Although other frameworks, such as SPIDER (sample, phenomenon of interest, design, evaluation, research type) are also appropriate for a qualitative synthesis, the PICOS framework was chosen because it facilitates a focused yet compressive search strategy. 14The original protocol for this synthesis has been registered on the PROSPERO portal (CRD42018115207) and, along with further details about the methodological approach, has been previously published elsewhere. 15he remaining phases are detailed further, and the process is outlined in the subsequent sections of this article.
Phase 2 is concerned with deciding what is relevant to the initial interest.In this synthesis, we are seeking studies examining debrief interaction and their relationship to learning.This is achieved through the collection and interpretation of naturalistic data-that is, data that would occur naturally, without being elicited by a researcher.This ensures the data synthesized originate from similar methodologies because there is an inherent risk of weak conclusions being drawn from data that are generated from different epistemological positions. 16,17earching was based on the PICOS framework (see Table, Supplement Digital Content 1 http://links.lww.com/SIH/A937, for search terms and search string) with predefined inclusion and exclusion criteria (see Tables, Supplement Digital Content 2 http://links.lww.com/SIH/A938, which list the inclusion and exclusion criteria).Included articles focused on facilitator-guided postevent debriefing because research suggests this is the most common format of debriefing in the health care simulation community. 2Full details of the search strategy are presented in Table 3 using the sampling strategy, approaches, range of years, limits, inclusion and exclusion criteria, terms used, electronic sources reporting framework.Details of study selection, data extraction, and analysis (phases 4 through 6) are presented here in the Data Analysis and Results section.
Phase 3 is concerned with data familiarity and extraction, achieved through multiple readings of the studies.As recommended by the Cochrane Group, 18 a custom data extraction process was constructed, consisting of 2 main spreadsheets.The first spreadsheet focused on study details and is based on the PICOS framework. 19This spreadsheet also included additional key elements related specifically to debriefing based on reporting guidelines by Cheng et al 20 (Table 4).The second spreadsheet was used to collect data for quality assessment of the included studies.This was based on a bespoke framework incorporating the CASP (Critical Appraisal Skills Programme) 18,21 and JBI-QAC (JBI Qualitative Appraisal Checklist) tools, 22 with additional debrief-specific fields.The function of quality assessment is to understand research conduct and author choices because these are woven with the interpretations authors make and what they choose to represent.Understanding author choices enables a more complete reading of the metaethnography.
Phase 3, therefore, consisting of organizing and summarizing the raw data, is itself an unrecognized form of interpretation that informs the analytic process. 23It then enables the  metaethnographer to group studies based on similarities, which is the hallmark of phase 4. 24 Importantly, the process of extracting data is not a neutral action-the metaethnographer has interpreted and then extracted material they feel is worthy of further interpretation.The "raw" data of interest, and subsequent groupings, are thus deeply tied to the researcher(s), their perspectives, and interests.The groups generated in phase 4 form the basis for phase 5.In Noblit and Hare's original work, there is a choice in phase 5 between 3 analytical approaches: reciprocal translation (synthesis based on similarities), refutational translation (when studies or authors refute each other), and line of argument formation (which begins with reciprocal translation and if appropriate, progresses to new theory generation in phase 6).In this synthesis, we decided that reciprocal translation was the most appropriate first step in analysis, seeing no evidence of refutations in the data.
Reciprocal translation involves discrete stages: (i) further data (metaphors) are extracted in a table or matrix format based on groupings and primarily includes direct author quotes; (ii) metaphors are compared with one another (translation); and (iii) new concepts are formed.Stages (ii) and (iii) are concerned with how the interpretations of 1 author can be seen in the interpretations of other authors.This is the central activity in doing reciprocal translation and is achieved by examining interpretations using direct quotes from authors (see Fig. 1 for an outline of this analytical process).For example, the following authors' direct quotes are taken from 3 different manuscripts and treated as metaphors: • Fryer-Edwards et al 25 state that in its simplest form, faculty work to specifically "elicit or help name take-home point" and go further and "ask learners for a commitment to try something new from the session" (p.639); • Kihlgren et al 26  These metaphors, alongside other similar metaphors, are iteratively compared with one another in a process of reciprocal translation.The outcome of this process is the formation of a new concept.In this example, the newly formed concept is labelled as "faculty-driven future action plans: concrete commitments to change" and will have a corresponding description that  also considers the context of the contributing articles.Metaphors are placed in a large spreadsheet to facilitate this process and may contribute to more than 1 newly formed concept.Phase 6 requires further interpretative work by analyzing how the concepts formed in phase 5 are related, linked, and intricately meshed.The product is creation of a new theory or "an explanatory framework that can account for all the synthesis findings" (p.11). 28hase 7 is presentation of the metaethnography, and this includes relation to existing literature, subjective interpretation, and reflexivity (presented here in the Discussion section).

Study selection (phase 2)
Searches were completed at the end of November 2020 and 128 articles were selected for full text appraisal.Seventeen studies were selected for inclusion in the synthesis (see

Data extraction and grouping (phases 3 and 4)
Selected studies were read multiple times and appropriate data were extracted (see   Supplement Digital Content 9 http://links.lww.com/SIH/A945, which shows the full matrix) in preparation for reciprocal translation.

Reciprocal translation and concept formation (phase 5)
The process of reciprocal translation was then undertaken (see Document, Supplement Digital Content 10 http://links.lww.com/SIH/A946, which details the process of reciprocal translation and concept formation).The translations were crystalized with a concept title that captures the essence of the translation work and satisfies the 5 criteria as stipulated by Noblit and Hare 13 : economy, cogency or conceptual clarity, range, apparency, and credibility.Each concept is accompanied by a description and contributing study details, providing some contextualization for the reader of this synthesis.In total, 31 new concepts were generated, grouped into 8 categories (see Table, Supplement Digital Content 11 http://links.lww.com/SIH/A947, which lists the concepts with descriptions) and is presented in Figure 3.
Careful reading of the translations at this stage is essential to get a sense of how the included studies contributed to the formation of new concepts; this is the core philosophy of metaethnography that is sometimes overlooked.It will also allow the reader of this synthesis a sequential process map: from metaphor extraction, continuing through the process of reciprocal translation and concept formation, which then in turn inform the new theoretical framework.

Consideration of study demographics and context on data interpretation
At this juncture, it is helpful to review study demographics and contextual factors (see Table, Supplement Digital Content 11 http://links.lww.com/SIH/A947, which includes details of learner groups, country where the research was conducted, and debrief approach used).This enables a degree of judgement or confidence in the newly formed concepts.This is relevant because it is these concepts that are further synthesized and contribute to new theory generation in phase 6.

Geography and culture
All 17 studies originated from North America (n = 3), Australasia (n = 3), and Europe (n = 11), of which 9 were from Scandinavian countries (with Sweden accounting for 5).Although there will be differing debriefing practices within the same country, there is evidence to suggest an influence of macroculture on how debriefing may unfold.Ulmer et al 29 suggested that in countries with a lower power distance index (PDI)-representing a reduced hierarchy gradient and more evenly distributed power in organizations-debriefs showed greater spontaneous participant discourse, increased intragroup interaction that was less dependent on faculty directives, and a greater emphasis on human factors discussions.There is also evidence demonstrating the influence of power distance on interprofessional relations 30 and the ability to speak openly about mistakes among peers, 31 which influences debriefing interactions.Because all included studies were conducted in countries with low PDIs, this is likely an influencing factor on the debriefing practices and interactions observed by study authors.This therefore influences the concepts and overarching theory generated in this synthesis.We identified no studies from higher PDI countries for this review, and thus understanding and interpreting the influence of culture on debriefing in these settings is an area for further empirical study.

Learner characteristics
Six studies included exclusively postgraduate (PG) participants and 10 studies included only undergraduate (UG) participants; 1 study included both groups.It is notable that in nearly all studies that included UG participants (n = 11/12), nurses were included as learners.Similarly, the 7 studies focused on PG participants all included medical professionals, of which 3 were exclusively medical professionals.Across all studies, participant groups tended to be limited to these 2 professions.
The influence of this on the data is evident; for example, as seen in T3 and T4, where emotions and topics of discussion may be influenced by UG and PG status, and T7, which exclusively contains studies with undergraduate participants.As a result, the data presented here can be seen through the lens of UG nursing and PG medical learners.Similarly, there is a trend of underresearch on simulation debriefing in other health care professions (eg, dentistry and allied health).

Simulation modality
The simulations reported in the included studies were typically focused on acute care scenarios using high-fidelity human simulators or human-manikin simulators.Only 1 scenario used simulated patients and another incorporated part-task simulators.Debrief discussions, progress, and group interactions may be influenced if simulated patients are included in either the simulation or the debrief, and this review shows a lack of contributing studies in this area.Equally underrepresented are part-task simulations or procedure-based simulations.This may be significant because these may have entirely different debriefing approaches centered around deliberate practice and mastery learning.

Debrief approaches
The most cited debrief model was a variation of a DAA (description, analysis, application) model (n = 4, all of which were conducted in Scandinavian countries), although there was no one translation heavily influenced solely by this group of studies.Debriefs were typically 15 to 45 minutes long, and the use of video (n = 6) was evident in different debrief approaches.Overall, the concepts generated in phase 5 are not restricted to a particular debriefing approach.
Researcher prevalence In the same way that demographics may influence interpretations, so may factors originating from prolific authors or research departments.One author, P. Dieckmann, was cited in 3 different studies (P2, P4, and P5) that feature prominently in the category of "interaction patterns" (T25, T26, and T27), more than any other category.In these studies, the construct of "reflection" occurs, with authors using a priori frameworks for data capture.This explicit positionality in relation to research approach and interpretative frameworks contributes to the metaethnographic analysis and interpretation.
Further learner and faculty demographic details It was not possible to examine other faculty factors (eg, experience, age, professional specialty, sex) because the data were not reported in the included manuscripts.Similarly, participant factors including clinical experience (grade), simulation experience, age, and sex were not reported.These factors may not be particularly helpful or meaningful because they are not standardized across various settings or professions, and aspects such as experience will be contextualized to the individual.

Reconstruction, recontextualization, and conceptualization: the sensemaking function of debrief discourse producing professional identity transformation
The 31 newly formed concepts were further synthesized through an interpretative process (phase 6) to produce a new visual explanatory framework of what the surveyed literature says about how learning occurs in simulation debriefings.In the following sections, the overall theoretical model (Fig. 4) is presented first; it is then broken down into smaller components and more thoroughly articulated while referencing back to a more detailed visual representation of the framework (Fig. 5).
The surveyed literature situates the learning that occurs in debriefing, in a process of what we have labelled reflective work.Reflective work consists of purposeful and structured discursive activities that recontextualize simulation experience bidirectionally with clinical reality.The space in which these interactions unfold is what we have called a socially constructed learning milieu, where interactions between faculty and participants set the content of debrief discourse.This space is naturally influenced by the events of the simulation experience, which may be relived and reconstructed in debrief discourse, and often occur before reflective work commences.The learning milieu and trajectory of discussion is further influenced by faculty primarily through their underlying agenda, their experience, and the debrief approach that is enacted.The outcome of doing reflective work is conceptualization of new ways of working, specifically linked to participant roles, clinical competence, and professionalized language: a process of transforming professional identity.
The process of sensemaking is deeply entwined with debrief discourse and interactions, which is presented in more detail in Figure 5.The narrative description of the theory that follows, combined with the visual representations in both figures, helps to show the relationships between different activities as well as the key actors and the influencing factors.
Relived and reconstructed simulation experience The surveyed literature indicates that a core component throughout the debrief is the emergence of narratives and emotions, evidencing a cathartic role.Participants may narrate their experiences spontaneously if the environment allows: for instance, this may be encouraged in a laissez-faire debriefing approach that is enacted by faculty with an open questioning approach.If emotions (fostered via a debrief model) are not surfaced early during the debrief, they may appear later in debrief discourse.Emotions may then impact progressivity, which demonstrates the inherent motivation of learners to share and attend to their emotions.Furthermore, the surveyed literature suggests that stories are not limited to active participants and may also represent past lived experiences of other observing participants and faculty.
Emotions are typically representative of poor self-perceived performance or competence by participants who have taken part in the simulation activity.This is managed by the observing participants and faculty by reconceptualizing performance away from negative connotations and instead keeping talk positive and focused on positive observed actions.Participants themselves also diffuse tension around negative self-perceptions by initiating laughter, which is then mirrored by participants and faculty.Reconceptualizing performance is also undertaken as an activity with the use of video, which functions as a third person, etic perspective.As such, it is used simultaneously to reconceptualize performance away from negative discourse and to contribute to a reconstructed simulation experience.
In addition to narratives, emotions, and video playback, participants and faculty share their observations.Individually, all these actions provide a fragmented contribution to reality, each illuminating a different aspect of the simulation experience.To fully reconstruct the simulation experience, beyond that experienced by individual participants and observers, these jigsaw pieces combine.In doing so, the shared, factlike, reconstructed experience is imbued with a sense of depth, consisting of emotions and perspectives: a shared "web" of experience.Thus, the relived and reconstructed simulation experience is foregrounded and available for further analysis from multiple approaches.Although this phase is not essential in all debriefings, it functions to enable deeper explorative discussion in a learning milieu, where discourse centered around the simulation activity is transposed with clinical reality: a process of recontextualization.
The learning milieu: the site of recontextualization of simulation experience and clinical reality A substantial component of debriefing practice is the simultaneous recontextualization of simulation experience to clinical practice, and of clinical practice to the experience.Recontextualization means considering the full range of experience, encompassing aspects such as actions, events, emotions, probing underlying rationale, behaviors, and perspectives in relation to clinical reality.Clinical reality embodies clinical practice, which itself is integral to an individual's professional identity.The discourse here constructs the core learning milieu, where learning as an interactional process flourishes and continues through to transforming professional identity.Similarly, events from clinical practice can be relocated to the debrief, where the simulation experience is used to make sense of participants' clinical experience.
Authenticity, or the simulation-reality gap, is thus paramount for successful recontextualization.Authenticity is dependent on the simulation and clinical experience of participants, the role enacted within scenarios (collectively, these can be argued as psychological fidelity), and further defined by social and physical fidelity.Faculty work to ensure that this simulation-clinical gap, which is dynamic in nature, is minimized.The surveyed literature suggests that if the gap is too large, participants may seize on the lack of authenticity to rationalize their experience and thus, a learning milieu may fail to establish.
The learning milieu consists of simulation experiences, emotions, and clinical uncertainty, combined with a realization for change.This enables participants to be receptive to different perspectives that arise from fellow participants, including observers, faculty, simulated patients, and video.Thus, topics that may have been seen as disorienting dilemmas or as having caused cognitive dissonance can be resolved; this, in turn, facilitates perspective transformation and conceptualizes new ways of working.
The learning milieu is diffuse yet definable.Topics of discussion can be set by participants and faculty by sharing narratives, by making observations that are directly linked with perspectives, and by evaluating participants' performance.Performance evaluation entails both positive aspects of observed performance and analysis of multiple participants acting as a team, labelled "team performance evaluation".More critical analysis also takes place, although only in the context of a psychologically safe environment.The learning milieu may be set using video playback with faculty constructing, through an ensemble of multimodal resources, a particular version of reality.Participants align to this reality in the proceeding discourse, thus satisfying the underlying faculty agenda.Faculty agendas may also include reconceptualizing participants' self-perceived performance, directing talk relating to positive performance, and ensuring course objectives are integrated and surfaced in discussion.Faculty also enable discussions around human factors and professional language to take place through specific discursive strategies.
Progressing learning talk: the leading role of faculty in progressivity For talk to remain located in the learning milieu on specific topics and move beyond description to deeper analytical discourse, it must progress in detailed sequences.Progressivity is the label given to talk as it continues between interlocutors (individuals in a conversation), and which enables the current and further sequences of talk to continue. 32In debrief discourse, faculty play a leading role in progressivity-their primary role is to ensure that continuing sequences of talk focused on the specific task that participants are discursively involved in.Faculty control this progressivity through their underlying agenda, the interactionally constructed debrief approach, and their own experiences.
Novice or inexperienced faculty may lack both subject-specific knowledge and educational strategies to enable progressivity.Similarly, participant contribution is stifled by faculty in exchange for close adherence to a debrief approach.Faculty adhering closely to a debrief approach materializes as a protocol-inuse.Here, success is attributed to discussion of course objectives.This contrasts with a more relaxed, laissez-faire approach, where facilitators take the role of chairperson, enabling participants to "self-select" in conversation, volunteer perspectives, and participate in storytelling: here, talk is progressed by participants themselves.In this relaxed approach, the learning agenda and feedback are learner-centered, with a more developmental dialogue centered around participant needs in relation to the scenario and clinical practice.We agree with Fryer-Edwards 25 that "learning edge" is an appropriate label characterizing this type of participant-centered talk.
Much of the discourse in the learning milieu emerges in a nurturing environment, which constructs and is constructed by psychological safety.Aspects such as laughter, investing in relationship building, and fostering and maintaining a positive affective milieu, enable participants to engage in risk-taking behavior (such as self-selection), engage in risky conversations (such as providing perspectives), and facilitate the process of critical reflection (defined here as the ability to examine one's own perspectives considering additional information).Such an environment also fosters dialogue on topics that may otherwise be avoided, such as negative feedback on performance.To create this environment, all group members interact in shared intentional activities, such as reconceptualizing performance, focusing on positive performance analysis, and constructing a debrief approach by permitting and actively encouraging self-selection and perspective sharing.Therefore, breaches in psychological safety may negatively impact the learning milieu and impair debrief progressivity.
Transformative learning: conceptualizing a new reality Individuals' lived simulation reality is disrupted by the process of reliving and reconstructing the simulation experience, demonstrating that there are differing, sometimes contrasting, conflicting, or reaffirming perspectives.This, in conjunction with examining their own professional identity, achieved in part through recontextualization, creates space for transformative learning.This transformative space may be amplified by feelings of disorientation and cognitive dissonance, manifesting as a realization for change in oneself.This motivation for change is further fostered by faculty by enabling a learner-centered approach, in part by tailoring discourse located in the learning milieu to the participant's needs.
At the core of this is the concept of risk-taking behavior, which in turn is linked to the sharing of perspectives and progressivity to allow the process of critical reflection to unfold.The emergent outcome of recontextualization and sensemaking is conceptualizing a new professional identity.We define professional identity here as encapsulating a participant's role within a system, individual clinical competence, and professionalization.A new professional identity is achieved in part through constructing complex professional language, and is solidified by faculty, who invite concrete commitments to change practice.It can thus be appreciated how challenging one's professional identity is risky business and requires a high degree of psychological safety.
Faculty also define new learning edges for participants, and these become part of learners' professional identities as areas of developmental need.Space for such discourse occurs toward the end of the debrief, signified by the application or summary phases in debriefing models.This contrasts with the reconstruction of simulation events that is typically situated toward the beginning of debriefing models often as a descriptive phase.However, whereas the entire sensemaking process does have a trajectory, the components need not be organized in a linear fashion for the entire debriefing session; the focus may shift depending on contextual factors.

Learning as process
The surveyed literature, through this metaethnographic qualitative synthesis, depicts how learning, as an interactivedriven process, occurs in the debrief.The explanatory framework visually materializes the central processes of examining simulation experience, recontextualizing experience bidirectionally with clinical reality, and conceptualizing and transforming one's professional identity within a health care system.This nonlinear process is the product of dynamic social interaction between participants, observers, and faculty who socially construct the microactivities themselves, and in doing so, are active sculptors in the learning process.

Considerations of study quality and collective rigor
The theoretical framework that we have generated from the included studies is necessarily a product of the original included studies: this metaethnography is, after all, an analysis of the results of all the included studies.As such, the rigorincluding the research design, analytical lenses, and reporting choices-of the original studies has a direct impact on the theoretical framework generated by this metaethnography.Indicators of rigor become more influencing if they are pervasive throughout multiple studies, and therefore they must be examined collectively.
There were consistent limitations regarding research conduct evident across the studies (see Table, Supplement Digital Content 12 http://links.lww.com/SIH/A948, which presents the main quality appraisal findings, including author self-declared issues).Across all included studies, the approach to ethics application and consent procedures varied the most.Only 12 manuscripts documented details of the ethics approval process, with nearly all of these citing they sought informed consent from course learners.Only 5 of the 17 manuscripts documented seeking informed consent from faculty.This raises the prospect of faculty either being unaware of the research being conducted or not fully aware of how data were being captured, analyzed, and presented.Although unlikely to influence data integrity, this may raise concerns around research conduct.
Authors acted as faculty in 4 studies, and this status was unknown in a further 5 studies.This could be a potential threat to data integrity.Data may, for instance, be selected by these authors to portray faculty positively.Interpretations of debrief interaction by researchers may be viewed through their "debriefing as faculty" lens, and interpretations that may seem negative may not be surfaced by other coauthors.Author, group, or departmental perspectives may dominate the analytical discourse when there may have otherwise been contrasting and valid arguments.However, there is clear evidence of analysis being grounded in the data across studies, revealing the analytical choices of study authors.
Furthermore, there is an argument that to fully understand the data, one must be familiar with the culture in which it takes place.This is influenced in part by faculty experience, training, and description of other characteristics such as base specialty and was not reported in 11 manuscripts.In contrast, description of course participants was provided either partially or fully in all manuscripts, although justification of participant sampling was not always provided.Providing a compressive description of study demographics, as advocated in the simulation-based reporting guidelines by Cheng et al, 20 would aid in both interpretation of the original studies and when synthesizing studies.
The included manuscripts were more consistent in nearly always stating the aim of the study.Those that stated clear theoretical or conceptual frameworks also stated the methodology and were congruent between the stated data collection methods, analysis, and interpretation.In nearly all cases, the data analysis method was clearly cited, with example data provided, although often more granular details such as the type of transcription convention used were not reported.Only 10 manuscripts acknowledged potential research limitations or gave evidence of reflexivity, with little evidence of authors challenging their own interpretations or considering potential risks to data integrity.Explicit discussion of reflexivity was absent in 9 manuscripts, and fewer than half (n = 7) showed evidence of challenging their own interpretations.
Collectively, authors tended to aspire for larger sample sizes across multiple research sites and often excluded certain discourse segments during analysis, citing this as a study limitation.There were additional limitations identified in this analysis but not reported by authors; for example, grouping data streams and data coding conduct (see Table, Supplement Digital Content File 5 http://links.lww.com/SIH/A941,column "AN", which details the additional identified limitations).
Holistically, many of the limitations, including issues with ethics and funding declarations, are likely to stem from reporting requirements and restrictions: they are artefacts of the publishing process.Reporting anomalies highlighted in this study and detailed in the supplement files do not necessarily suggest that the included studies' research conduct itself was problematic or substandard. 17,33Indeed, most of the identified issues do not impact data analysis and interpretation.More surprising, however, are issues like poor reporting of faculty background, researchers acting as faculty, and poor consent procedures.

DISCUSSION
In addition to presenting the synthesis findings, phase 7 of a metaethnography entails (i) subjective interpretations and (ii) field relations or relating the findings to the literature.A section on reflexivity is important for readers of this synthesis to be able to contextualize subjective interpretations and is an aspect of best practice in metaethnographic work. 28,34

Reflexivity and subjective interpretation
The perspective of the first author, who was primarily responsible for interpreting the studies included in this review, is an important consideration when interpreting the resulting framework (Table 5).For instance, the researcher's background may mean that greater emphasis was given to some domains than others, and there is a possibility that different readings of the included manuscripts may result in different metaphors of interest being drawn out.The transparent nature of data extraction and analysis is intended to reassure readers of the rigor of the overall framework while still situating the emergent discourse in the context of the researcher's positionality.The coauthor provided reasoned perspectives on the data, which contributed to the creation of the framework, adding further rigor to the work.
The background of the first author (R.K.) matches the demographics of included studies, specifically: the experience in debriefing models, debriefing in low-PDI countries, and simulation focused predominately on nursing and medical professionals in acute care scenarios.The core philosophy of R.K. as

TABLE 5. Reflexivity and Researcher Professional Identity
• Clinician-medical doctor for more than 15 y in the United Kingdom • Simulation educator (faculty) and researcher for more than 10 y • Postgraduate-level lecturer for more than 5 y • Previous research in health care debriefing using the approach of discursive psychology and similar methods • Specific areas of interest include psychological safety, debriefing, peer observation, feedback processes, interaction analysis • Undertaken formal Train-The-Trainer courses in specific debriefing models and approaches: ○ Debrief Diamond (Guy's and St Thomas' NHS Foundation Trust); ○ Description, Analysis, and Application (EuSim); ○ Debriefing With Good Judgement -Advocacy With Inquiry (Centre for Medical Simulation, CMS); ○ Learning Conversation (Advanced Life Support Group, ALSG).
a debriefer is centered around psychological safety and creating a nurturing environment that enables difficult conversations to take place, with strategies aligning with those in the framework.The background of the second author (G.R.) is similar.Similarly, the framework echoes perspectives on the role of the experienced debriefer in moving discussions forward to a learning edge: what participants want to talk about, identified through emotions, comments on simulation and clinical practice, and narratives introduced in the debrief.Using diverse perspectives, particularly those of other participants, facilitates what we hope to be high-yield learning.

Relation to existing literature Psychological Safety
Multiple elements present in the framework echo some of the strategies articulated in the simulation literature to foster psychological safety during debriefing.6][37] Psychological safety can thus be seen as an emergent construct, based on these dynamic interactions in the debrief, and is visible in the framework.This is particularly relevant because, as concluded from a recent synthesis examining psychological safety in health care, low psychological safety limits individual and organizational learning. 38By creating a trusting and nurturing space, coupled with interactionally constructed recontextualization of experience, discourse can move to what Edmondson would define as the "learning zone", 39 meshed within the learning milieu.It is here that critical reflection can thrive 40 and can provide further benefits of increased confidence and team familiarity when returning to clinical workspaces. 41arning conversation Through the framework, the surveyed literature suggests novice debriefers prefer staying close to the debrief model they are following, and as such, can be rigid in their discourse.This may be a reflection of an increased cognitive load for the debriefer: Fraser et al 42 argued that faculty must attend to aspects such as scenario recall, formulating questions, setting discussion topics, attending to emotions (their own and participants'), and addressing learning objectives and learner needs.Consequently, novice debriefers may have less working memory available to engage in more complex tasks, such as listening for elements of cognitive dissonance and aligning participants' learning edge with course objectives and the faculty agenda.This position is further supported by authors of different debrief approaches who argue that as debriefers become more experienced, they may move away from more scripted and protocolized structures. 43,44Importantly, however, Kolbe 45 warns about straying too far from a structure, arguing that "ill-structured" debriefs risk failure.The analysis of Morse 46 suggests that failure arises from faculty primarily focusing on "actions not on learner thinking" (p.175), providing further evidence that tailoring and deepening talk to participant needs is cognitively demanding for debriefers.
The framework suggests that progressing the debrief to the stage of being a learning conversation requires experienced faculty to be able to draw on a combination of educational strategies, their own perspectives, and subject knowledge.This aligns with Tavares et al, 47 who argue that such "learning conversations" rely on faculty having a "sophisticated repertoire of conversational strategies" (p.1024), which attend to the needs of the learner while maintaining psychological safety and integrating performance observations.Debriefs as learning conversations, therefore, tend to move away from strict adherence to scripts, models, and mechanicalistic feedback processes, and instead toward identifying and managing performance improvement. 47This is in keeping with earlier descriptions of "learning conversations" used in accredited life support courses, 48 which attempt to desegregate feedback and debriefing, to create a more learner-centered conversation.
Reconstructing clinical events A central component of the framework that facilitates debrief structuring is the reconstruction of events-the creation of a shared web of experience.As participants share stories, they are presenting their reality of the simulation experience to their peers and facilitators.Gordon et al 49 aptly states the role of narratives: "…when a story is told, the narrator constructs and presents identities, events, and realities in interaction with others" (p.2).Although the literature advocates for the sharing of emotions during this phase, possibly to surface cognitive dissonance, the framework suggests there is a risk of self-critical performance.This may be one reason why Jaye et al 43 argue that such emotive analysis, including performance analysis, should be discouraged during the "descriptive" phase, and instead undertaken in the "analysis" phase, while accepting that there may be cultural influences in how participants respond to emotional venting.Faculty thus strive for a balance: a narrative should serve to construct a shared understanding of the experience, which can go on to influence the learning milieu, while avoiding negative analysis that could impact psychological safety.It seems here that a central function of the reconstruction of events is to provide a contrast to clinical reality and to set the ground for further analytical work, as posited by Secheresse and Nonglaton. 50Such debrief progressivity requires confidence in facilitation skills, particularly being able to decipher participant discourse at moments of dissonance, and identify these as developmental opportunities. 51

Cognitive dissonance and its role in reflection
The core phenomenon of cognitive dissonance and the subsequently associated discursive activities are pervasive in the learning milieu; this is where "reflective work" is done.Parker and Myrick 40 argue that the process of reflection involves content analysis and process reflection (or strategies for problem solving) and can be seen to be an integral component of recontextualization.They further argue that personal transformation occurs when individuals go beyond this, examining their own underlying premises for action (premise reflection), and engaging in a process of what Mezirow calls perspective transformation: "a critique of the presuppositions on which our beliefs have been built" (p. 1). 6Parker and Myrick 40 warn against faculty dominating the discourse because this may result in discussion remaining at the level of content reflection.In the framework, this appears as novice faculty who are unable to progress the debrief.
The work of Hernandez et al 52 looking at the resolution of cognitive dissonance, which was centered around implicit bias, demonstrated the central role of other participants in a group setting to provide alternative perspectives.These alternate perspectives function to promote ongoing self-examination of beliefs, particularly against normative or contextual standards.Parker and Myrick 40 further suggest that "the community in which individuals interact is influential in providing powerful norms and cultural influence" (p.329), consistent with Kolb's  (1984)  4 perspective that learning involves transactions between the person and the environment.The literature therefore supports the role of all participants and faculty providing these perspectives of the culture (clinical reality), while engaging in recontextualization of experience.
A recent interview study exploring learning by trainee pharmacists provides further evidence, from a different methodological lens, for the transformative nature of such interactions in a simulation debrief.Tallentire et al 53 highlighted the features of a disorienting dilemma: a sense of responsibility (akin to "realization for change") followed by an "exploration of options for new roles, relationships, and actions" for transformative learning (p.3).This assertion that transformative learning does occur in simulation debriefs, particularly centered around reconceptualization of professional roles, is supported by other researchers. 54,55Other literature thus supports the trajectory identified in this framework: topics and specific frames of reference, originating from dissonance, are identified and discussed against cultural norms (clinical reality), progressing to transformation of professional identity.
Viewing the framework through existing theoretical lenses Learning theories, such as those of Kolb, 4 Vygotsky, 56 Dweck, 57 and Knowles (in 58 ), join Mezirow's 6 and others in the framework.This is encouraging and validating: rather than proposing an entirely new theory of learning, the framework thus surfaces some of the complex ways in which participants learn using existing theoretical perspectives.
Kolb's 4 experiential cycle, which itself leans on arguments from Lewin, Dewey and Piaget, is perhaps the most frequently cited learning theory in simulation literature.Kolb argues that a key requirement is to surface learners' beliefs and theories (or premises), examining and testing them, and integrating more "refined ideas".Kolb further argues that a learning experience arises only when the "fabric of expectation" of daily routine is violated (p.28). 4 This is conceptually similar to Mezirow's definition of a disorienting dilemma and is reflected in the framework as agenda setting, identifying a learning edge, and constructing the learning milieu.The learning milieu, which is constructed through activities such as a shared web of experience, performance evaluation, perspective sharing, and storytelling, is thus aimed at focusing attention on the violations of the reality of learners.
It is through the process of reflective observation, where participants "reflect on and observe their experience from many perspectives" (p.30), where new understandings or theories are accessed. 4Thus, the interplay between participants' underlying premises and the new theories they are faced with is a central component of reflective work.It is this interplay that leads toward transformative learning; it is a process that can generate internal (or intrapsychic) conflict.Central to the idea of abstract conceptualization, and indeed transformative learning, is resolution of conflicts.Thus, the role of faculty, as depicted in the framework, is not only to set the learning milieu based on initially perplexing events, but also to relate this to clinical reality (recontextualization), while simultaneously inviting alternative perspectives and with the goal of resolving conflict.This process is, according to the framework, the essence of reflective work.
Engaging in reflective activities is, by its nature, risky business: participants are sharing their learning edge, which they may associate with clinical experience and competence, and may share their uncertainty about their own performance and therefore their identity.They must then engage in revealing their underlying premises and accepting alternative perspectives.This process requires "openness, willingness, and engagement" (p.50). 59Although Rogers 59 argues that "autonomy, effective feedback, connection to peers, access to faculty, and appropriate challenges" (p.51) can enable engagement in the learning milieu, the framework (and the broader literature) goes further, advocating for a psychologically safe space.Boud and Walker argue that a sense of trust, titrated to the level of disclosure, is required for reflection.They emphasize the importance of creating a climate to allow expression of emotions: "conditions of trust and security" (p.194). 60Similarly, Jones 61 argues that a "safe and inclusive learning space" (p. 6) is needed to engage in critical conversations as part of transformative learning.
The role of other theories in the framework remains open to further interpretation.For instance, scholars of Vygotsky 56 may argue that narratives, performance analysis, and ultimately defining a learning edge are comparable with the work of identifying an individual's zone of proximal development.Furthermore, faculty progressivity and the presence of plural perspectives are representations of the "more knowledgeable other" that helps to enable knowledge construction.However, this perspective does not allow for the benefits of other parties in the debrief, who can both observe and cocreate group-level knowledge once a common conceptual understanding is achieved. 62n this cocreation, concepts of learning can be seen less as individualized, cyclical processes and more as social, group ones.
A further consideration of applying Kolb's cycle to the current framework is that there is no space for active experimentation.Although experimenting in the debrief has been cited in the literature as either part of the debrief or as a concluding element, 63,64 this is less typical of debrief practice and is not characterized in the framework.Although some may argue that active experimentation occurs in clinical practice or in repeated simulation, a more fruitful line of enquiry may be to ponder the question: 'what are learners actively experimenting'?From an assimilation perspective, there may be new knowledge which can be practiced.However, this assumes that the individual will encounter a near identical situation subsequent to the simulation.If it is to occur in clinical practice, as some postulate, it is plausible that other, intervening transforming experiences will occur and indeed, Kolb 4 himself states that "no two thoughts can ever be the same, since experience intervenes" (p.26).Instead, and in line with Kolb's assertion, we can view learning as a continued developmental process: within simulation debriefing, this consists of professional identity transformation, even if these are small, incremental adjustments.Thus, active experimentation is perhaps more accurately described as ongoing experiences-worldly interaction-and ongoing development.

Research limitations
Almost certainly, additional research has already been undertaken in the intervening period between this search concluding and the framework being generated.Furthermore, it is plausible, and somewhat expected, that new research would serve to reconsider previous translations.Best practice advocates for synthesis to be continually updated: the provision of supplement files with this manuscript will, we hope, aid in this endeavor.
No search strategy will provide complete coverage of the literature.Equally, an unfocused approach is likely to render the selection of studies unmanageable.One further study by Pelletier and Kneebone 65 was identified while the analysis and write-up were being undertaken.This was not captured in the 3008 articles identified through initial searching.Reassuringly, some of the key metaphors in this manuscript resonate with existing translations, although full interpretative work and reexamination against other metaphors would be required to fully synthesize interpretations (see Document, Supplement Digital Content File 13 http://links.lww.com/SIH/A949, which outlines key metaphors from Pelletier and Kneebone's manuscript and how these relate to existing translations).

Future research
This work lays the foundation for new research avenues in simulation debriefing.Specific research questions can be generated from the metaethnography, such as: • What are the different ways in which disorienting dilemmas or cognitive dissonance are discursively constructed in interaction?• What is the structure and function of participant narratives?
• What genres exist within participant narratives?
• What are the discursive functions of laughter in relation to fostering psychological safety?• How does group discourse change over time?
• How are questions constructed, and how do they function in debrief discourse?• What are the pedagogical strategies that enable progressivity?• How do different modes, or an ensemble of multimodal communications, contribute to the interaction patterns and sensemaking for participants?
Alongside the framework, these examples of new research endeavors will further enrich our understanding of debriefing as an educational activity.
Although the framework has multiple interconnecting components, it is nevertheless restrictive in its ability to portray the dynamic nature of debriefing.There are always dynamic aspects such as psychological safety, moving between reconstruction and recontextualization, intrinsic tensions of time, and the setting of a negotiated agenda.Relationships between and among participants and faculty also morph over a relatively narrow time span.Furthermore, the framework has limited applicability in contexts outside those of the original studies.As such, this identifies areas for future empirical work: for example, debriefs in countries with high PDI; debriefs with different health care professional groups; debriefing after using different simulation modalities; and researching different debrief models.In parallel, the metaethnography also identifies an opportunity to undertake more fine-grained analysis of the conversations to understand how these interaction patterns of talk are constructed.For example, there is scope to understand the methods that participants undertake to achieve actions such as producing narratives, evaluating performance, recontextualizing performance, managing sensitive topics to maintain psychological safety, and doing reflective work.Such future research can be seen as testing interpretations of metaethnography, which Britten et al 24 argue should be seen as an appropriate endeavor after concluding a metaethnography.

CONCLUSIONS
The function of this work has been to synthesize existing research on debrief interactions to understand how interactions in simulation debriefing are related to learning.In the simplest form, the synthesis findings suggest that learning occurs through facilitated recontextualization of simulation experience with clinical reality.This occurs as a process, through talk, in a socially constructed "learning milieu".The framework provides direction on some of the specific debriefing elements that may feature in faculty training such as psychological safety, inviting multiple perspectives, tailoring discourse to participants' learning edge, and how to progress debrief talk in a learning milieu.The approach of interaction analysis can also be used as a strategy to review, analyze, and refine debrief practice as individual or group-reflective exercises.Faculty development and refinement of practice optimizes participant learning and ultimately leads to simulation-based education having a greater impact on patient care.This synthesis can provide the initial footings on this path.
8 http:// links.lww.com/SIH/A944, which outlines the process of arriving at the groupings [tab a] along with frequently occurring themes [tab b]).These groupings formed the basis of a further data extraction episode, where metaphors from each manuscript were directly extracted into a matrix (see Table,

FIGURE 1 .
FIGURE 1. Analytical flow from phase 3 to phase 6. Phases 3 and 4 function to familiarize the researcher with the data, extract study details, and to form groups.In preparation for phase 5, metaphors are extracted in a spreadsheet and an interpretative process of reciprocal translation is undertaken.Phase 6 is a further analytical step where newly formed concepts from phase 5 are synthesized, and a new theory or explanatory framework is proposed to account for the data.

FIGURE 2 .
FIGURE 2. PRISMA 2020 flow chart.The search was initially conducted over a 6-month period (June-November 2018) and then updated 2 years later (November 2020) because the review was interrupted and there was new published and relevant material (see document, Supplement Digital Content 3 http://links.lww.com/SIH/A939 for details of the individual searches).The 128 articles that were assessed for eligibility were then appraised.At least 1 reason was cited if articles were deemed illegible, although often, there were multiple reasons (see spreadsheet, Supplement Digital Content 4 http://links.lww.com/SIH/A940 for details regarding article appraisal).

FIGURE 3 .
FIGURE 3. Newly formed concepts from phase 5.In total, 31 concepts or translations were generated during this phase.See document, Supplement Digital Content 10 http://links.lww.com/SIH/A946 and table, Supplement Digital Content 11 http://links.lww.com/SIH/A947 for greater description of each of these concepts, and specifics about how they were generated from metaphor translation.

FIGURE 4 .
FIGURE 4. Overall theoretical model of how learning occurs in simulation debriefs.The relational process elements related to participant learning constructed by debrief interactions.Interactions are themselves subject to influencing factors (bottom right).

FIGURE 5 .
FIGURE 5. Transforming professional identity framework: detailed description of the learning process.Learning materializes through nonlinear processes of reliving and reconstruction of simulated events, recontextualization with clinical reality, and conceptualization of a changed professional identity.These are interactionally constructed debrief activities, subject to influences, and driven by agents.

TABLE 1 .
Key Practice Points in Undertaking a Metaethnography Route of analysis: decide if the metaethnography is to be a reciprocal translation synthesis, refutational synthesis, or line of argument synthesis or combination of all.† Not essential if number of included studies is small and no disparate groups of translations.
*‡Not essential if translations do not yield new information.

TABLE 4 .
Data Extraction Fields of Study Details

Table ,
Supplement Digital Content 5 http://links.lww.com/SIH/A941, which outlines full study details data set;Table, Supplement Digital Content 6 http:// links.lww.com/SIH/A942, for simplified study details data set; and Table, Supplement Digital Content 7 http://links.lww.com/SIH/A943 for quality appraisal data for included studies).Following this process, 7 main groups were formed showing how the studies were related: faculty factors; participant centeredness; faculty/participant actions; debrief interactions; video use; simulation scenario content; and psychological safety (see Table, Supplement Digital Content