How does guided group reflection work to support professional identity formation in postgraduate medical education: A scoping review

Abstract

Purpose

In postgraduate medical education, guided group reflection is often applied to support professional identity formation. However, little is known about how guided group reflection is shaped and how it works. Our scoping review synthesizes existing evidence about various approaches for guided group reflection, their aims, components and potential working mechanisms.

Methods

We conducted a scoping review using JBI (Joanna Briggs Institute) guidelines for conducting scoping reviews. We searched PubMed, PsycINFO, EMBASE and ERIC databases for all research articles published in English or Dutch in an iterative team approach. The articles were extracted and summarized quantitatively and qualitatively.

Results

We included 71 papers (45 primary research papers and 26 non-empirical papers including program descriptions, theoretical concepts and personal experiences). We identified a diversity of approaches for guided group reflection (e.g. Balint groups, supervised collaborative reflection and exchange of experiences), applied in a variety of didactic formats and aims. We distilled potential working mechanisms relating to engagement in reflection, group learning and the supervisor’s role.

Conclusions

There are significant knowledge gaps about the aims and underlying mechanisms of guided group reflection. Future systematic research on these topics is needed to understand the effectiveness of educational methods, that can help facilitate learning conditions to best shape professional identity formation (PIF) in educational curricula.

Keywords:

• Guided group reflection
• medical specialty training
• postgraduate medical education
• professional identity formation
• scoping review

Practice points

• Several approaches for guided group reflection (GGR) can be used to promote PIF.

• When designing GGR, it is important to carefully consider the role of assessment and mandatory nature, given their potential facilitating or hindering role towards engagement in reflection.

• GGR groups and their supervisors should be consciously composed to balance learning needs.

Introduction

In postgraduate medical training, professionalism is seen as an essential attribute for future doctors to provide high-quality care (Frank et al. 2015; General Medical Council 2017; Edgar et al. 2020). Professionalism has been described as ‘the necessity for physicians to adhere to high ethical and moral standards to gain the trust of their patients’ (Royal Dutch Medical Association 2007; van Luijk et al. 2009; Mak 2019). No universally accepted definition of professionalism exists, as it is constructed by the community of practice (van Luijk et al. 2009; Goldie 2012; Birden et al. 2014; Berkhout et al. 2018; Veen et al. 2020). Education fostering professionalism should take into account the context of the profession itself. Postgraduate medical education (PGME) has often been identified as a key stage in professional identity formation (PIF) (Goldie 2012; Cruess et al. 2016; Sarraf-Yazdi et al. 2021), as with the transformation from student to practitioner, the focus shifts from doing like a physician to being a physician (Pratt et al. 2006; Barnhoorn et al. 2022).

In the medical context, ongoing reflection is considered a core activity in PIF and can be done during (reflection-in-action) or after an experience (reflection-on-action) (Schön 1983). Firstly, reflection is crucial in gaining new knowledge and skills, by analysing an experience, evaluating learning in light of current knowledge and drawing conclusions for future learning activities (Kolb 1984; Goldie 2012). Secondly, learners use reflection as a meta-cognitive skill to select, monitor and evaluate their approaches towards a learning task (Zimmerman 1995, 2000); Mann et al. 2009; Sandars 2009; Birden et al. 2014; Winkel et al. 2017). Thirdly, reflection can help recognize and understand the influence of personal beliefs and value systems on the effectiveness of outcomes in clinical settings, which is an essential part of PIF (Sandars 2009). While reflection is considered an intrapsychic process, it can be fostered during collaborative learning (Mann et al. 2009; Aronson 2011). Guided group reflection (GGR) is trained in multiple group sessions, guided by a supervisor, in which each new session builds on previous ones, allowing participants to potentially put their acquired knowledge into practice and reflect on their experiences in the next session (Schön 1983; Kolb 1984; Sandars 2009; Aronson 2011; Goldie 2012; Uygur et al. 2019).

While in PGME many approaches for GGR exist to stimulate PIF (Winkel et al. 2017; Yazdankhahfard et al. 2019), an overview of how GGR is shaped (i.e. aims and components) is lacking, and so is the evidence about its working mechanisms. Reviews often focus on healthcare professionals in general instead of on PGME, the phase in which PIF actually takes place (Mann et al. 2009; Sandars 2009). Reflection as a concept is addressed and not GGR.

In this review, we will provide a literature overview by synthesizing (1) published approaches for GGR aiming to foster PIF and their aims, and (2) their components and potential working mechanisms.

Methods

Given the exploratory nature of this research field and breadth of potentially relevant literature, we conducted a scoping review in accordance with the Joanna Briggs Institute (JBI) five phase methodology (Peters et al. 2020): (1) development of a scoping review protocol (van Oorschot et al. 2022); (2) development of a search strategy; (3) selecting articles; (4) extracting results (charting the results); and (5) collating, summarizing, and reporting the results.

Research questions

We aim to answer these research questions:

• What educational approaches for GGR in PGME are described?

• What are the components of approaches for GGR and what are their potential working mechanisms?

Eligibility criteria

We developed inclusion criteria using the population-concept-context (PCC) format (Table 1) (Peters et al. 2020). The population of interest included postgraduate medical residents. We included studies that cover the concept of GGR, in which reflection was focusing on personal and professional functioning within a future medical profession to foster PIF. The context of interest was PGME in both primary and secondary medical care settings all over the world, in order to take different cultural values and customs into consideration.

Table 1. Inclusion and exclusion criteria.

Criterion	Inclusion	Exclusion
Population	Postgraduate medical residents	Undergraduate students Other health care professionals in training (e.g. nursery and dentistry) Health care professionals
Concept	Guided group reflection (>2 participants, guided by a supervisor) Longitudinal (≥2 sessions) Reflection focusing on personal and professional functioning	Stand-alone sessions One-time workshops Reflection focusing on clinical tasks only (e.g. supervision by an experienced physician in terms of medical skills), medical errors or logistics, unless they also focused on aspects that concern PIF
Context	Postgraduate medical education in primary and secondary care
Type of article	Empirical paper: Original research paper based on primary research data Non-empirical paper: Descriptive paper (e.g. perspectives, conceptual papers, program descriptions)	Commentaries Conference papers, if no follow-up journal article available
Language	English or Dutch	All other languages

We included empirical papers, based on primary research data, and non-empirical papers, such as perspectives, (theoretical) concepts and educational program descriptions for inclusion. We included quantitative, qualitative as well as mixed methods studies. There was no date restriction, to prevent the exclusion of potentially relevant papers and we chose to include full text articles in English or Dutch, due to present language skills of the research team and the large number of papers available in those languages.

Search strategy

We searched the databases PubMed, Embase, ERIC and Psycinfo, assisted by a librarian on 23rd of December 2021 and included terms related to (1) reflection, (2) topics that may concern PIF, determined by the research team based on literature and expert opinion, and (3) postgraduate medical education. The search string was adjusted for each database included (Supplementary Appendix A).

We hand-searched reference lists of all included articles for additional studies. Additionally, we consulted topic experts to help identify any further relevant articles and hand-searched Dutch peer-reviewed medical journals.

Screening

Following the search, all identified references were uploaded into Endnote 20 and duplicates removed. After pilot screening of the first 20 references by the research team (FvO, MB, MV, AT, SvD), we independently screened titles and abstracts in pairs (FvO, MB, MV) for developing inclusion criteria, using screening management software Rayyan (rayyan.ai). Thereafter, three independent reviewers (FvO, MB, MV) reviewed the full text of potentially relevant studies in detail, which enabled tightening the inclusion criteria. We resolved discrepancies between reviewers by consensus and if necessary, by consultation of the research team.

Data extraction and charting

We developed two data extraction forms using Microsoft Excel; for primary research papers and for non-empirical papers (Supplementary Appendix B). The data extracted included information about study characteristics, information about the PCC format and key findings relevant to our review questions. We iteratively identified and specified variables as we progressed. Two independent reviewers (FO, MB) piloted data extraction with 10 primary and 10 non-empirical papers (28% of all included papers) and revised the extraction forms. As the extraction of the first 20 papers proved consistent in content between the two reviewers, one reviewer (FO) extracted the data from the remaining papers. If the first reviewer was uncertain about the extracted data, it was cross-checked by the second reviewer (MB). The cross-checking process involved validating the extracted data in the original papers, followed by a discussion session between the two reviewers to address any discrepancies.

Data analysis and presentation

Using Microsoft Excel, we performed quantitative descriptive analyses of study characteristics. To answer our research questions we used qualitative analysis, informed by the learning activity template proposed by Keck et al. that distinguishes the ‘what’, ‘why’, ‘who’, ‘how’, ‘where’ and ‘when’ of a learning activity (Keck et al. 2021). Within the extracted data we searched for components of GGR. We considered ‘components’ as all characteristics (e.g. group composition, supervisor) of an approach for GGR. Finally, we distilled potential working mechanisms from the extracted data, informed by the realist methodology that assumes that complex interventions (in our case GGR) have an underlying intent and logic about how they are believed to work (how, for whom and under what circumstances) (Wong et al. 2012; Pawson 2013). In this approach, potential working mechanisms help to understand what it is about the intervention that makes it work by explaining how components contribute to outcomes. It consists of the interaction between components of a learning activity and the participant’s response or reasoning upon that (see glossary terms) (Greenhalgh et al. 2017).

Reflexivity

The team consisted of researchers with different backgrounds (two psychologists, three medical doctors and one philosopher). Two team members have experience in supervising GGR. The team frequently met during all stages of the review to ensure that the data was being analyzed consistently. This approach led to depth and richness of the team discussions and a more comprehensive understanding of the data.

Results

First we will report on general study characteristics and then we will describe our results on educational approaches for GGR and their aims (what and why) and their components (how, when, who). In the final section, we will interpret these results and share some identified potential working mechanisms of GGR.

Search results

Figure 1 shows the numbers of articles at each stage of the inclusion process in a PRISMA diagram, resulting in 71 included full-text papers.

Figure 1. PRISMA flow diagram.

Study characteristics

All included articles are shown in Supplementary Appendix C. Most studies were conducted in the United States (n = 40). Ten medical specialties were represented, almost half of the articles were from family medicine (n = 33), 45 reported primary research and the remaining 26 were non-empirical papers.

The findings identified in empirical articles were based on qualitative (Johnson and Brock 2000; Cohen-Katz et al. 2003; Pinder et al. 2006; Smith and Anandarajah 2007; Ballon and Skinner 2008; Clandinin and Cave 2008; Antoun et al. 2014; Nothnagle et al. 2014; Salander and Sandström 2014; Veen et al. 2015; Veen and de la Croix 2016, 2017; Player et al. 2018; Antoun et al. 2019; Freudenreich and Kontos 2019; Ng et al. 2019; Sivam and Joseph 2020; van Braak et al. 2021, 2022), quantitative (Brock and Stock 1990; Alexander and Skinner 2002; Locher and Blankenstein 2004; Turner and Malm 2004; Cataldo et al. 2005; Ghetti et al. 2009; Huang et al. 2019; Bar-Sela et al. 2012; Diaz et al. 2015; Bird et al. 2017; Lichtenstein et al. 2018; Hata et al. 2019; Antoun et al. 2020) or mixed data (Brenninkmeijer et al. 1986; Musham and Brock 1994; Brenninkmeijer et al. 1999; Adams et al. 2006; Feld and Heyse-Moore 2006; Graham et al. 2009; Wen et al. 2013; Kung et al. 2015; Foshee et al. 2017; Asan and Gill 2018; Sherman et al. 2019; Shamaskin-Garroway et al. 2020; Holtzclaw et al. 2021), and data sources included questionnaires and surveys (Brenninkmeijer et al. 1986; Brock and Stock 1990; Musham and Brock 1994; Brenninkmeijer et al. 1999; Alexander and Skinner 2002; Locher and Blankenstein 2004; Turner and Malm 2004; Cataldo et al. 2005; Adams et al. 2006; Feld and Heyse-Moore 2006; Smith and Anandarajah 2007; Graham et al. 2009; Bar-Sela et al. 2012; Wen et al. 2013; Diaz et al. 2015; Kung et al. 2015; Bird et al. 2017; Foshee et al. 2017; Huang et al. 2019; Sherman et al. 2019; Antoun et al. 2020; Shamaskin-Garroway et al. 2020; Holtzclaw et al. 2021), individual interviews (Musham and Brock 1994; Cohen-Katz et al. 2003; Pinder et al. 2006; Graham et al. 2009; Nothnagle et al. 2014; Veen et al. 2015; Foshee et al. 2017; Player et al. 2018; Sivam and Joseph 2020; van Braak et al. 2021, 2022) and focus group interviews (Cohen-Katz et al. 2003; Smith and Anandarajah 2007; Antoun et al. 2014; Foshee et al. 2017; Ng et al. 2019; Sherman et al. 2019; Holtzclaw et al. 2021). Quantitative studies investigated a variety of outcomes, including burn-out (Ghetti et al. 2009; Bar-Sela et al. 2012; Hata et al. 2019; Antoun et al. 2020; Holtzclaw et al. 2021), physician empathy (Cataldo et al. 2005; Ghetti et al. 2009; Foshee et al. 2017; Hata et al. 2019) and (self-assessed) professionalism (Adams et al. 2006; Kung et al. 2015). Few studies compared a group that received an educational intervention for GGR with a control group, that did not (Turner and Malm 2004; Cataldo et al. 2005; Adams et al. 2006; Huang et al. 2019) or only infrequently did receive an educational intervention (Musham and Brock 1994), and two studies compared two different educational approaches for GGR (Brenninkmeijer et al. 1986; Pinder et al. 2006).

What and why: Educational approaches for GGR and their aims

Articles described a plurality of approaches for GGR to stimulate PIF. We distinguished articles that described stand-alone approaches (Appendix D) and those that were part of a broader educational program. The most frequently described stand-alone approach was the Balint group (Scheingold 1980, 1988; Brock 1990; Brock and Stock 1990; Margo and Margo 1994; Musham and Brock 1994; More 1996; Johnson and Brock 2000; Turner and Malm 2004; Cataldo et al. 2005; Adams et al. 2006; Pinder et al. 2006; Forssell 2007; Ghetti et al. 2009; Bar-Sela et al. 2012; Mahoney et al. 2013; Antoun et al. 2014; Diaz et al. 2015; Lichtenstein et al. 2018; Player et al. 2018; Antoun et al. 2019; Huang et al. 2019; Antoun et al. 2020; Lichtenstein 2020; Sivam and Joseph 2020). In these groups (Keith et al. 1993; Smith et al. 1993; Smith and Anandarajah 2007; Graham et al. 2009; Salander and Sandström 2014; Asan and Gill 2018), residents explore the dynamics of their patient interactions and gain insight into their own reactions, by presenting and discussing a case from different perspectives. Studies identified aims of creating awareness of one’s own influence in the doctor-patient interaction (Scheingold 1980, 1988; Brock 1990; Brock and Stock 1990; Forssell 2007; Player et al. 2018; Antoun et al. 2019, 2020), burnout prevention or improving empathy (Ghetti et al. 2009; Bar-Sela et al. 2012; Huang et al. 2019). In Supervised collaborative reflection (SCR) residents reflect on experiences to explore their behaviours and underlying attitudes and beliefs regarding their profession (Brenninkmeijer et al. 1986, 1999; Locher and Blankenstein 2004; van Katwijk 2005), aiming to integrate this in lifelong learning. In Exchange of experiences, residents collaboratively reflect on their experiences from practice, with a limited structure and without necessary follow-up of discussed themes over time (Veen et al. 2015; Veen and de la Croix 2016, 2017; van Braak et al. 2021, 2022). In Support groups residents share experiences with peers, guided by a supervisor, aiming to provide a safe place to reflect on personal and professional problems (Alexander and Skinner 2002; Feld and Heyse-Moore 2006). Specific approaches are appreciative inquiry, narrative reflective practice using parallel charts, ten-minute snapshots, reflective team supervision, primary supervision (Jellinek 2007; Clandinin and Cave 2008; Graham et al. 2009; Chien et al. 2012; Chandra et al. 2017; Ng et al. 2019). In some articles, the approach for GGR was not clearly defined, for example small group reflection (Wen et al. 2013; Holtzclaw et al. 2021) (Supplementary Appendix D).

GGR was also part of a broader educational program for example focusing on wellness or resilience, combining GGR with other reflective learning activities (e.g. reflective writing) or educational interventions (e.g. resilience workshops) (Novack et al. 1997; Fins and Nilson 2000; Yeheskel et al. 2000; Cohen-Katz et al. 2003; Ballon and Skinner 2008; Masding et al. 2009; Nothnagle et al. 2014; Bird and Pincavage 2016; Bird et al. 2017; Foshee et al. 2017; Freudenreich and Kontos 2019; Hata et al. 2019; Shamaskin-Garroway et al. 2020).

How, when and who: Educational approaches for GGR and their components

In each paragraph we will describe the identified characteristics of an approach (i.e. components), followed by study findings and conclusions of included papers regarding this specific component: how, when, who (see Supplementary Appendix E for an overview of components for each included article).

How: Starting point for reflection

Components

Balint groups and SCR sessions follow a predetermined format, initiated with a topic raised by the resident, forming the starting point for further reflection and analysis. Exchange of experience sessions also start with a resident’s experience but there is a more open and flexible format (van Braak et al. 2021). Few papers found that common resident-chosen themes include difficult interactions with patients or colleagues and physician self-care (Lamboo et al. 2005; Smith and Anandarajah 2007). Brock (1990) states that topics that are introduced in Balint groups depend on the phase of professional development.

In other papers, GGR sessions were initiated using an external source that introduced a theme, for example a TV series or videos (Foshee et al. 2017; Holtzclaw et al. 2021), a book or an article (Freudenreich and Kontos 2019; Hata et al. 2019). Predetermined topics in included articles vary, ranging from broad (e.g. professional dilemmas or (clinical) topics meaningful to residents (Nothnagle et al. 2014; Chandra et al. 2017)), to specific (e.g. structural institutional racism (Sherman et al. 2019)), and are sometimes combined with clinical topics (e.g. pain and symptom management) (Fins and Nilson 2000).

Findings and conclusions

Smith and Anandarajah (2007) highlight that allowing participants to choose the focus of a session, leads to greater resident’s satisfaction. Feld and Heyse-Moore (2006) found that resident’s chosen issues may be unreal, because residents may be hesitant to honestly address personal difficulties, fearing judgement or that their concerns would be interpreted as weaknesses. Van Braak et al. (2021) found that collaborative reflection adds educational value for all if the interaction is inclusive and diverse, meaning that all residents have the opportunity to bring up something for discussion. However, compulsory contribution may reduce authenticity and compromise a safe learning environment, which in turn depreciates the educational value.

How: Mandatory nature of GGR

Components

Several papers described whether sessions were all mandatory (Brenninkmeijer et al. 1999; Locher and Blankenstein 2004; Forssell 2007; Rietmeijer et al. 2008; Ghetti et al. 2009; Graham et al. 2009; Kung et al. 2015; Player et al. 2018; Hata et al. 2019), mandatory for a limited period (Musham and Brock 1994; Cataldo et al. 2005) or non-mandatory (Brenninkmeijer et al. 1986; Keith et al. 1993; Smith et al. 1993; Adams et al. 2006; Feld and Heyse-Moore 2006; Smith and Anandarajah 2007; Bar-Sela et al. 2012; Wen et al. 2013; Nothnagle et al. 2014; Salander and Sandström 2014; Bird and Pincavage 2016; Asan and Gill 2018; Freudenreich and Kontos 2019; Ng et al. 2019).

Findings and conclusions

According to Scheingold (1988) and Musham and Brock (1994), in mandatory sessions, intrinsic motivation of residents could vary, which could hinder group cohesion and a safe learning climate. However, Scheingold (1988) and Veen and de la Croix (2016) also indicate that an advantage of mandatory attendance could be that the group composition does not change, and that the group is already cohesive due to other joint training activities. If sessions are not mandatory, emotions (e.g. anxiety about self-disclosure) or personality traits (e.g. introversion) of a resident, could lead to non-attendance (Musham and Brock 1994).

How: Assessment and evaluation of GGR

Components

Only one study described that poor attendance would be flagged (Sivam and Joseph 2020), and assessment on content for training purposes was done in one study (Yeheskel et al. 2000).

Findings and conclusions

Several authors emphasized that assessment, other than obligatory attendance would undermine the learning process and hinders sharing of experiences (Feld and Heyse-Moore 2006; Kim et al. 2016; Lichtenstein et al. 2018). Scheingold (1988) indicates that Balint seminars have traditionally been intended for psychoanalytic rather than educational purposes and therefore are not amenable to systemic evaluation (Scheingold 1988). Van Katwijk (2005) highlights that setting criteria and performing an evaluation can contribute to learning. However, to maintain a safe learning environment, personal vulnerabilities and challenges during the process should not be explicitly communicated to the educational institution as an assessment.

When: Timing of GGR

Components

GGR was offered to residents during different years of training, ranging from first-year to fifth-years residents. In some studies GGR was implemented during one training year, whilst in other studies it was a longitudinal program during whole residency (Keith et al. 1993; Yeheskel et al. 2000; Veen et al. 2015; Veen and de la Croix 2016, 2017; van Braak et al. 2021, 2022).

Findings and conclusions

Margo and Margo (1994) suggest that the first year of residency seems to be the most suitable for Balint groups, since that is the time that PIF gets increasingly more attention in medical training, whilst others indicate that some essential topics are not addressed until later in residency (Brock 1990). Fins and Nilson (2000) target residents because they have mastered basic clinical skills but remain receptive to educational experiences that might alter their lifelong practice patterns. Other included articles did not address the timing of GGR.

Who: Group characteristics

Components

Groups were homogeneous (same specialty and same level of training) or heterogeneous (different specialty or different level of training or different healthcare professionals) (Smith et al. 1993; Chien et al. 2012; Chandra et al. 2017; Huang et al. 2019; Shamaskin-Garroway et al. 2020). Most papers did not describe group continuity. Some described groups as open (Huang et al. 2019) or closed (Brenninkmeijer et al. 1986, 1999; Veen et al. 2015; Veen and de la Croix 2016, 2017; van Braak et al. 2021, 2022). In the first, members could join in and leave over time, whilst in the second, the group started and ended in the same composition.

Findings and conclusions

Brock (1990) suggests that Balint groups should ideally be at the same level of training and that mixed groups (different levels of training) may result in conflicting learning needs in PIF. According to Rietmeijer et al. (2008), mixed groups can be a disadvantage since the presence of both experienced and inexperienced physicians may make it less engaging for the former. However, some authors found that residents appreciated participants of different levels of training, different medical specialties or different healthcare professionals (Chandra et al. 2017; Shamaskin-Garroway et al. 2020), as sessions led to more diverse ideas and interactions, and different problem-solving approaches (Antoun et al. 2014; Holtzclaw et al. 2021). Studies agree that time is needed to establish group cohesion and that changing the composition does not benefit group cohesion, safety (Winstead et al. 1974; Scheingold 1988; Brock 1990; Veen and de la Croix 2016), and depth of discussion (Rietmeijer et al. 2008). On the other hand, if participants already know each other from their work environment, they may not feel free to disclose sensitive experiences (Rietmeijer et al. 2008; Graham et al. 2009). Brenninkmeijer et al. (1999) found that a group composition chosen by residents was seen as stimulating the learning process.

Who: Background and role of the supervisor

Components

GGR sessions are commonly led by one or two faculty members (Fins and Nilson 2000; Adams et al. 2006; Ghetti et al. 2009; Masding et al. 2009; Wen et al. 2013; Nothnagle et al. 2014; Bird et al. 2017; Foshee et al. 2017; Shamaskin-Garroway et al. 2020; Holtzclaw et al. 2021), or by GPs together with a psychologist or behavioral scientist (Scheingold 1980; Keith et al. 1993; Yeheskel et al. 2000; Smith and Anandarajah 2007; Veen et al. 2015; Veen and de la Croix 2016, 2017; van Braak et al. 2021, 2022). In one study, the supervisor was a senior peer, further progressed in training (Chandra et al. 2017) and in some, the group was led by researchers (Alexander and Skinner 2002; Clandinin and Cave 2008).

Findings and conclusions

Van Braak (2021) and Nothnagle et al. (2014) identify a safe learning environment as a pre-requisite for efficient group learning. Other authors view the supervisor as having a critical role in creating this (Smith et al. 1993; Margo and Margo 1994; More 1996; Novack et al. 1997; Yeheskel et al. 2000; Forssell 2007; Ballon and Skinner 2008; Chien et al. 2012; Mahoney et al. 2013; Kim et al. 2016), by establishing confidentiality rules (More 1996; Forssell 2007) and modeling sharing personal experiences and emotions (Ballon and Skinner 2008; Chien et al. 2012). Some studies highlight the importance of a neutral, non-judgmental and independent position of the supervisor, which will allow residents to open up (Winstead et al. 1974; Smith et al. 1993; Asan and Gill 2018). Other responsibilities include facilitating discussion (More 1996; Cohen-Katz et al. 2003; Feld and Heyse-Moore 2006; Wen et al. 2013; Kung et al. 2015; Lichtenstein et al. 2018; Huang et al. 2019; Holtzclaw et al. 2021; van Braak et al. 2021, 2022), stimulating active participation (Yeheskel et al. 2000; Clandinin and Cave 2008; van Braak et al. 2022), structuring the session (Alexander and Skinner 2002; Freudenreich and Kontos 2019; Ng et al. 2019) and fostering group cohesion and identity (Scheingold 1988). Brock (1990) states that the role of the supervisor often changes as the group develops, from leading and structuring to observing and coaching. Awareness of the developmental phase of residents might assist supervisors in choosing the appropriate role (Brock 1990). In one study, Balint seminars were led by supervisors from different cultural backgrounds than the participants. Some participants considered this a hindering factor, as they were required to contextualize the patient’s situation and explain the local setting and policies, which might result in deemed culturally inappropriate or unacceptable solutions (Antoun et al. 2014). Since leadership of GGR requires coaching skills, but also personal maturity, some authors state that a supervisor, especially for Balint groups and SCR might need training or even certification (Scheingold 1988; Kim et al. 2016).

Potential working mechanisms of GGR

We distilled potential working mechanisms of GGR from the findings above, using the components and related study outcomes as input. These are summarized in Table 2.

Table 2. Components and their potential working mechanisms.

Components	Potential working mechanism
A topic raised by the resident as starting point for further reflection	If a GGR session starts with a topic or case raised by the resident, the resident may feel intrinsically motivated to participate since the session more likely matches the learning needs of the resident. This fosters engagement in reflection. If a GGR session starts with a compulsory contribution of a resident, for example an experience or case, residents may feel pressured to contribute, will make something up out of fear of judgement or worry that their concerns may be perceived as weaknesses. As a consequence they will not reflect authentically.
Mandatory nature – Group continuity	If GGR is a mandatory part of the curriculum, not all residents will have equal motivation for participation or will be intrinsically motivated for attendance, which may hinder group cohesion, making residents feel less safe to share their experiences and making them disengage from the reflective process. If GGR is a mandatory part of the curriculum, the group consists of the same participants each session, and participants have the opportunity to get to know each other. This will build trust and rapport in the relationship between participants. As a consequence participants will become more willing to share and reflect on experiences.
Assessment and evaluation	If participants of GGR are assessed by their supervisor on their performance in the sessions, then they will become less willing to be honest about their personal experiences and limitations, as they may be afraid not to pass. As a consequence they will not be open to authentically reflect and will not improve their self-reflecting skills. If participants of GGR are assessed by their supervisor, then this implies that the education is of value, which promotes participant motivation.
Group composition – participants and colleagues	If participants of GGR are working together in clinical practice and come together in an educational setting to collaboratively reflect, they already know each other, which may build mutual trust in the relationship between participants. As a consequence participants will become more willing to share and reflect on experiences. If participants of GGR know each other from their work environment, participants might struggle to be open in sharing their experiences and limitations (e.g. medical errors), fearing they could be perceived as weak by colleagues, or that their disclosures could be used against them in the workplace. Additionally, participants may be hesitant to provide honest feedback on their colleagues experiences, fearing that they hurt their colleagues.
Group composition – mixed groups	If GGR groups are mixed (different level of training, different medical specialties or healthcare professionals), then participants are exposed to a wide range of perspectives on a case, including more diverse ideas and interactions and different problem-solving approaches. This may guide them in future events. If GGR groups are mixed, then participants may have conflicting learning needs, because they face different challenges during clinical practice. This may lead to misunderstanding and impaired group cohesion, making participants less engaged in reflection.
Role of the supervisor	If GGR is guided by an independent supervisor that has no other supervisory responsibilities towards the residents, and this supervisor establishes confidentiality rules, and also shares own personal experiences and emotions, then residents will feel safe to share, because they are not afraid that what they share can be used against them in their training

Discussion

We described a variety of approaches for GGR aiming for the support of PIF, their aims and components and distilled their potential working mechanisms. How components of GGR actually worked and eventually contributed to PIF was not explicitly addressed in empirical papers, since they were not conducted with the aim to unravel working mechanisms. The lack of systematic description of mechanisms behind GGR is consistent with the approach often used in medical education through using a medical model. This model assumes, according to Biesta and Van Braak (2020), that ‘teaching causes learning’ and focuses on what works and not how it works. In this review, we attempted to unveil how GGR works and distilled potential working mechanisms, related to engagement in reflection, group learning and the supervisor’s role. Next, we explore some of these mechanisms in more detail and relate them to educational theories.

First, our review suggests that to promote PIF, residents must feel intrinsically motivated and safe to engage in authentic reflections. Components such as assessment, the mandatory nature and starting point for reflection potentially play a role in this. Our search for group reflection applied in educational settings, yielded little information about assessment of GGR. Two previous reviews suggest that assessment implies real value for reflective activities, which in turn drives learning (Mann et al. 2009; Sandars 2009). However, there are concerns that reflection on personal functioning is not amenable for assessment and may limit reflection, compromising a safe learning climate (Hodges 2015; Ng et al. 2015; de la Croix and Veen 2018). In articles included in our review, other than being mentioned as a concern, this was not explored further. Also, the (non)-mandatory nature was not investigated in depth in included articles. Our synthesis suggests that mandatory attendance might be beneficial, enabling group continuity and thereby group cohesion and safety. However, it may also lead to a variety in experienced motivation, hindering group cohesion and safety. The realist review of Hamilton et al. (2023) suggests that initially mandating undergraduate students attending a small number of group reflection sessions, could help them understand their purpose and value, increasing attendance later. Additionally, our review suggests that GGR sessions starting with a resident’s experience from practice seem more likely to match resident’s learning needs and foster engagement in reflection. This aligns with theories of Mezirow (2000) and Schön (1983) who state that reflection is particularly useful for learning in response to complex clinical problems.

Second, we found in our review that group learning (with other participants and supervisor) might play a role to foster PIF. This can be related to Biesta’s theory of education, indicating that medical education needs to aim for professional qualification, professional socialization and professional subjectification. When hearing perspectives of others, residents get acquainted with the norms and values of the profession, and get to know how the wheels spin. This may help them to become a member of the professional group (socialization) and it also contributes to forming their own ideas on their profession as a resident and eventually becoming an autonomous professional (subjectification) (Biesta and van Braak 2020). Supervisor guidance can play a crucial role here, by role modelling and coaching (Bandura and Walters 1977; Mann et al. 2009; Sandars 2009; Cruess et al. 2014). The supervisor’s role naturally depends on their background (e.g. medical doctor or psychologist), a factor that is not addressed in included articles, but potentially representing a working mechanism as well.

Interestingly, in the included articles objectives of GGR were not specified in terms of conceptualizations of PIF, but were standardized, generic (e.g. preventing burnout), or not stated at all. Previous research indicates that setting clear learning goals is important, since it sends a signal that reflection is meaningful for personal growth and professional development, increasing the resident’s motivation, potentially leading to more successful reflection (Sobral 2005; Sandars 2009; Aronson 2011). Moreover, clearly defined learning objectives in terms of PIF, may help the supervisor choose an approach for GGR and determine the timing (Sandars 2009).

Limitations

Our search string included a description of topics of reflective sessions relating to PIF, according to previous literature and expert opinions. However, we may have missed papers on other topics that were not included in the search. To mitigate this, we used snowballing and citation tracking on the reference lists of the included papers to find additional articles that met our inclusion criteria. In addition, we did not directly gain insight in what educational practice looks like. We could have missed educational approaches for GGR that have not yet been described in research papers. In addition, our review is limited by the available data, as the included studies did not aim to unravel mechanisms.

Future directions

We found hardly any studies on the working mechanisms of GGR. Future empirical research should pay attention to this and refine the potential working mechanisms as distilled in this review, for instance in a realist evaluation, showing how an educational approach works, for whom, in what context (Wong et al. 2012). In line with Biesta (Biesta and van Braak 2020), we suggest not limiting the focus to the process of GGR but also specifying its goals and how it is related to concepts such as PIF, to clarify how GGR can support PIF in terms of educational purposes. The combination of clarifying the purpose, research on underlying mechanisms and providing a framework for systematically evaluating GGR approaches, would benefit future research and educational practice to support life-long learning practices.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Glossary

Components: All characteristics of a learning activity, for instance group composition, group size and supervisor.

Working mechanisms: Explain how components potentially contribute to outcomes of a learning activity (or not). It consists of the interaction between components of a learning activity (e.g. a small closed group), and the participant’s response or reasoning upon that (e.g. feeling safe to share emotions).

Additional information

Funding

This study was supported by ZonMw (Project number 839170003).

Notes on contributors

Frederieke van Oorschot

Frederieke van Oorschot, MD, is GP in training and PhD candidate at the Department of Primary and Community Care, Radboud University Medical Center, Nijmegen, the Netherlands.

Marianne Brouwers

Marianne Brouwers, MD, PhD, is GP and an associate principle lecturer at the Department of Primary and Community Care, Radboud University Medical Center, Nijmegen, the Netherlands.

Jean Muris

Jean Muris, MD, PhD, is GP and full professor at the Department of Family Medicine, Care and Public Health Research Institute (CAPHRI), Maastricht University, Maastricht, the Netherlands.

Mario Veen

Mario Veen, PhD, is philosopher and Assistant Professor Educational Research at the Department of General Practice, Erasmus University Medical Center, Rotterdam, The Netherlands.

Angelique Timmerman

Angelique Timmerman, PhD, is a psychotherapist and Assistant Professor at the Department of Family Medicine, Care and Public Health Research Institute (CAPHRI), Maastricht University, Maastricht, the Netherlands.

Sandra van Dulmen

Sandra van Dulmen, PhD, is full professor in Communication in Healthcare at Radboud University Medical Center, Adjunct professor at Faculty of Caring Science, Work Life and Social Welfare at the University of Borås, Borås, Sweden and research coordinator Communication in Healthcare at the Netherlands institute for health services research (Nivel).