Optimising planned medical education strategies to develop learners' person‐centredness: A realist review

Abstract Context Person‐centeredness is a stated aim for medical education; however, studies suggest this is not being achieved. There is a gap in our understanding of how, why and in what circumstances medical education interventions that aim to develop person‐centredness are successful. Methods A realist review was conducted with a search of Medline, Embase, HMIC and ERIC databases and the grey literature using the terms ‘medical education’ and ‘person‐centred’ and related synonyms. Studies that involved a planned educational intervention in medical education with data on outcomes related to person‐centredness were included. The analysis focused on how and why different educational strategies interact with biomedical learner perspectives to trigger mechanisms that may or may not lead to a change in perspective towards person‐centredness. Results Sixty‐one papers representing fifty‐three interventions were included in the final synthesis. Nine context–intervention–mechanism–outcome configuration (CIMOc) statements generated from the data synthesis make up our refined programme theory. Where educational interventions focused on communication skills learning or experiences without person‐centred theory, learners experienced dissonance with their biomedical perspective which they resolved by minimising the importance of the learning, resulting in perspective endurance. Where educational interventions applied person‐centred theory to meaningful experiences and included support for sense making, learners understood the relevance of person‐centeredness and felt able to process their responses to learning, resulting in perspective transformation towards person‐centredness. Conclusion Our findings offer explanations as to why communication skills‐based interventions may be insufficient to develop learners' person‐centredness. Integrating experiential person‐centred learning with theory on why person‐centredness matters to clinical practice and enabling learners to make sense of their responses to learning, may support perspective transformation towards person‐centredness. Our findings offer programme and policymakers testable theory to inform the development of medical education strategies that aim to support person‐centredness.


| INTRODUCTION
Person-centredness is recognised as an essential attribute for clinicians to deliver effective health care 1 and is an espoused outcome for medical education. 2 Person-centred communication skills training is an established part of many medical education curricula 3 ; however, there has been concern that person-centred attitudes do not increase in learners as they progress through medical education and may decline. [4][5][6][7][8][9][10] In recent years several studies have offered evidence on how certain types of medical education, such as longitudinal integrated clerkships (LICs), [11][12][13] can support person-centredness. However, we still have insufficient understanding as to how, why and in what circumstances the range of planned medical education programmes that aim to develop person-centeredness are successful or not. This knowledge is important if we are to optimise medical education for the development of person-centred practitioners.
Person-centredness is based on an interpretivist epistemological framework 14 which guides one's perspective on what constitutes health and one's clinical role. It is often described in contrast to a positivist biomedical perspective in which patients' reports of illness are taken to indicate the existence of disease processes and the clinician's 'expert' role is limited to accurate diagnosis and management of disease to cure or improve the patient's illness. 15 A person-centred perspective recognises that the person is a unique individual whose health is a complex interplay of biopsychosocial and other factors and whose health goals will be informed by their unique context and in terms of what matters to them. 16,17 It recognises patients' subjective knowledge and expertise and seeks to work with patients as equal partners in a shared decision making process. 16 It also recognises that the quality of relationships influences health outcomes and that health care professionals' self-awareness is a vital component of a personcentred approach. 18, 19 Mead and Bower's review of the empirical literature describe five conceptual domains of patient-centredness: biopsychosocial perspective, 'patient-as-person', sharing power and responsibility, therapeutic alliance and 'doctor-as-person.' 20 Personcentred theory involves the conceptual and evidential rationale for a person-centred approach to clinical practice. 21 The terms 'patient-centred' and 'person-centred' are often used interchangeably, and there is considerable overlap in the conceptual domains underpinning these terms. In recent years, there has been a move towards the term 'person-centred' to acknowledge the personhood of an individual and not just their status as a patient. 22 We therefore chose to use the term 'person-centred' in our review.
Teaching of person-centredness in medical education is often conceptualised in terms of skills and competencies. 23 However, studies have shown that biomedical clinician perspectives may be a barrier to delivering person-centred care 24,25 and it has been argued that a person-centred approach requires a perspective shift from a focus on 'what is the matter with you?' to 'what matters to you?' 26 Various policy papers recommend that a workforce committed to personcentred practice is an essential pillar in delivering person-centred care 1,27 and that person-centred values and principles are essential to guiding person-centred behaviours. 28 We queried whether there is a gap in medical education strategy and if the current focus on communication skills training may overlook person-centred theory.
Our review sought to answer the following questions: What are the key mechanisms, triggered in particular contexts, that lead to the success or failure of planned medical educational interventions in developing person-centredness? How might these explanations influence educational policy, practice, and research? The objective of our review was to produce testable theory useful to medical educators and policymakers. For this reason, we chose to focus on the planned education curriculum as opposed to the experienced curriculum which occurs in the wider health care context and over which medical education may have limited influence. We conceptualised the outcome of person-centredness as a perspective and so chose to focus our review at the level of the learner.

| METHODS
We chose to conduct a realist review as this is a practical methodological approach designed to inform policy and practice. Realist research is also particularly suitable for investigating complex interventions, such as those in medical education, 29 and for synthesising qualitative, quantitative and mixed-methods research. Unlike standard systematic reviews that seek to understand whether an intervention works, realist reviews have an explanatory focus and seek to understand what works for whom and in what circumstances. 30 Our research focused on understanding outcomes at the level of perspective and so our use of the concept of 'mechanism' is at the level of human reasoning, 31 where it represents learners' responses to the educational programme they are offered. This paper reports the review according to the Realist and Metanarrative Evidence Syntheses: Evolving Standards (RAMESES) publication standards. 32 The review was conducted by a six-person team which included four doctors with extensive experience in undergraduate and postgraduate medical education (AB, JR, SP and CM), a realist methodologist (KS) and an information specialist (SR).

| Step (i): Developing an initial programme theory
Our first step was to describe an initial programme theory to guide data collection and analysis. We adopted three of the strategies outlined by Shearn et al. 33 : extracting tacit theories held by our research team, using concepts from substantive educational learning theory and extracting tacit theories from exploratory searches of the academic and grey literature. Details of our scoping searches are available at the open science framework (https://osf.io/qnkfh/?view_only= d1d36e578dd449a78ec62b13e2efebc5). These sources produced theory concepts that were configured to form an initial programme theory consisting of three draft programme theories based on the types of learning they represented: cognitive (theory of personcentred care), constructivist (transformative learning environments) and experiential (clinical placements). The detail of our initial programme theory is in Appendix S1.
Our initial programme theory was checked with stakeholders planning, delivering and receiving undergraduate and postgraduate medical education. Their feedback confirmed that the initial programme theory provided a useful starting point for answering the research question. The stakeholder group discussed the scope of the review in terms of professional groupings and agreed to focus the review on medical students and doctors given that the diagnostic and management responsibilities of doctors are not shared with all health professionals and may have a specific impact in the development of a person-centred approach. It was agreed that there were likely to be mechanisms in common across undergraduate and postgraduate learning and so papers from across the continuum of medical education should be included.

| Step (ii): Searching for evidence
Our initial programme theory provided the framework to develop our systematic search strategy. Earlier scoping searches allowed us to refine our search strategy. For example, most papers found under the term 'professionalism' were concerned with the impact of the hidden (experienced but not planned) curriculum, whereas our review focus was on the planned curriculum. We searched multiple electronic databases (Medline, Embase and HMIC via the OVID interface and ERIC via the EBSCO interface on 15 July 2019) through Boolean combinations of free-text and database-specific subject heading terms for 'medical education' and 'person-centred' and related synonyms. Full search strategies for each database can be found in Appendix S2.
We did not limit the search by including search terms on outcomes as we did not want to miss useful papers that may include data on person-centred values, attitudes and beliefs without using these terms. In addition, relevant studies identified from a previous search for grey literature at the scoping stage were included. The search was limited to the English language and to results published since 2000 as we reasoned that this was when the concepts of patient-centredness and person-centredness became established in medical education.
Additional papers relating to highly relevant studies were found using the CLUSTER method 34 including citation tracking and contact with authors. As the synthesis progressed, a new element of theory emerged on 'professionalism being understood as maintaining emotional distance', and as is recommended in realist synthesis, 35 we did additional searching, using Google Scholar, to look for evidence to support, refute or refine this element of theory.

| Step (iii): Selection and appraisal of documents
Selection of articles was based primarily on the RAMESES identified principle of relevance, whether the article illuminated the research question and contributed to theory development. 35 There was no restriction on the type of study eligible for inclusion.
Screening was led by AB with a random sample of 10% of titles double appraised by JR, CM, SM, KS or SG at both abstract and full-text stages. We included papers with a planned educational intervention for medical students and doctors with data on outcomes related to person-centredness. In realist synthesis, the inclusion of relevant data from across disciplinary boundaries is encouraged to support theory development. We therefore agreed to include three papers that were returned in our search, which related to other health professionals as we felt they supported the overall development of programme theory. We excluded papers that were about the hidden curriculum, where there was no intervention or where the description of the intervention was too brief to derive explanation. Health care interventions that were specific to a particular policy context and not part of a formal medical educational programme, were also excluded (e.g. person-centred medical home interventions in the USA). We also excluded papers that only reported outcomes at behavioural level (e.g. improvement in communication skills). Full inclusion and exclusion criteria can be found within the protocol published on the PROSPERO database (CRD42020168197).

| Step (iv): Extracting and organising data
Before extracting the data, first read all the included full-text papers to familiarise herself with the data. Data on study characteristics and summary findings for each paper was extracted into an Excel spreadsheet. In addition, each paper was assessed for rigour (high or low) based on whether its findings were considered coherent and plausible. Coding was both inductive, i.e., codes were identified from the data, and deductive, i.e., they were informed by the initial programme theory. Each paper was uploaded onto NVivo 12 (QSR International, Warrington, UK) qualitative data analysis software, to facilitate coding of text at a granular level. These coded data were grouped into concepts and then developed into 'if-then' explanatory statements. Explanations of each paper's findings were then transferred to the Excel spreadsheet for ease of looking at explanations across papers. These explanations were colour coded so we could easily see if they supported, refined or refuted the draft programme theories or whether they inspired new theory. A table with details of studies included in our realist synthesis is available in Appendix S3.

| Step (v): Data synthesis
The data synthesis process involved several iterative cycles where we continually moved between data and theory to help us develop realist causal explanations for outcomes. The emerging findings were subject to critical appraisal at regular research team meetings to ensure the coherence and validity of the explanations. To develop our realist explanatory theories, we looked across papers to see if there were repeated patterns (demi-regularities) between particular educational strategies and outcomes. We then looked for results and/or author suggestion and used abductive reasoning to see if we could explain how these outcomes occurred (mechanisms).
Data synthesis then involved looking across the data for confirmatory and disconfirmatory evidence for our theories and, indeed, novel explanations that might explain the change in learner perspective or lack thereof.
A learner's perspective prior to learning was inferred from data indicating their response to the learning. We found that all interventional strategies that aimed to support person-centredness validated the perspective of participants who leaned towards a person-centred perspective. Learners who leaned towards a more biomedical perspective had variable responses to learning and this was the predominant learner perspective reflected in both our qualitative and quantitative data. At this stage, we chose to focus our review on the subset of data that provided explanations of the impact of educational strategies on learners who leaned towards a biomedical perspective. In line with RAMESES guidance, 35 we checked with our stakeholders who agreed this focus would be most useful to them as educators and policymakers in terms of addressing the question of how planned medical educational strategies can be optimised to support learners' person-centredness.
Several models are available for articulating theory configurations. 36,37 We found that separating the learner context from the educational strategy helped us to isolate the mechanisms leading to changes in perspective. We therefore chose to use the context-intervention-mechanism-outcome configuration (CIMOc) for our realist causal explanations. We grouped CIMOcs based on their outcome and working backwards were able to investigate the key features of the educational strategies that triggered the mechanisms leading to a change in person-centred perspective (or not) and so to construct an overall programme theory.
The initial search yielded 4217 results. Figure 1 illustrates the filtering process that resulted in 61 papers or documents being included in the final realist synthesis. Eight papers or documents were looking at the same 'intervention', so the number of educational interventions looked at was 53. In 37 interventions, learners were undergraduates; in 11, they were postgraduate doctors-in-training; and in five, they were experienced clinicians or multidisciplinary teams in work. The

| Refined programme theory
Our refined programme theory consists of nine CIMOcs that explain how educational interventions that aim to develop personcentredness may or may not succeed (Table 1). Our analysis found that these explanations applied to leaners across the medical education continuum (undergraduate, postgraduate and continuing professional development). Figure 2 is a visual representation of our refined programme theory. Abbreviation: CIMOc, context-intervention-mechanism-outcome configuration. In both situations, the feeling that they had no meaningful role with patients whose needs did not fit with an expertise in disease diagnosis or management, resulted in perspective endurance. For these medical students and (maybe some or even) many medical scientists, the discursive nature of social science often makes it seem less certain, more contingent than natural scientific 'fact' and therefore less valid. 51 A lack of knowledge of person-centred theory, such as the rationale for a bio-psycho-social formulation of health, was identified as a key reason that learners struggled to engage with person-centred learning.

|
Many students also reported scruples about exploring the patient's life circumstances, in particular, social relations and working conditions. The students revealed a lack of knowledge about the relevance of psychosocial factors to health and disease. Thus, they perceived a holistic perspective as unnatural. 44 Papers with quantitative data that compared cohorts with different interventions showed a decline in person-centred attitudes in cohorts where person-centred theory was not clearly integrated into learning. 59  Another student wanted to be both academically skilled and empathetic but felt that being bio-medically skillful is given higher priority throughout medical school. In his opinion, one learns only to diagnose, refer patients, and relate to a time schedule during medical school. 43

| Perspective transformation: How educational interventions may succeed in developing person-centredness
Four explanatory statements from our data synthesis (CIMOCs 6-9) explain how educational interventions may lead to perspective transformation towards person-centredness (Table 1). Critical analysis of these four statements identified two overarching mechanisms that prompted change towards person-centredness: greater clarity on the relevance of person-centredness to clinical practice and support to process emotions and challenge assumptions. Thirty-six papers contributed to explanations on perspective transformation towards person-centredness. 49,58,59, Out of these papers, 31 interventions offered a combination of the components that enabled both mechanisms to be triggered: person-centred theory, meaningful experiences and opportunity to make sense of one's responses to learning. Only five papers that contributed to these explanations offered experiences without theory or sense-making opportunities. 63,68,73,84,87 The experiences offered in these five interventions were either longitudinal in nature or involved an active person-centred role in caring for patients, both of which are likely to have enabled students to personally reflect on the relevance of person-centredness to clinical practice.

| Mechanism 1: Clarity on relevance of person-centredness to clinical practice
CIMOc 6: Understand health as holistic and value personalisation of care Interventions that applied person-centred theory to meaningful experiences helped learners understand the relevance of a person-centred approach to patient care. 53,[58][59][60]65,66,69,71,72,74,75,77,78,82,83,86,87,89,90,94,96 The contribution of theory to these interventions varied from didactic introductions on the concepts and rationale for a person-centred approach 60,71,72,74,82,94 to regular discursive and interactive engagement with theory and application. 58,65,70,77,85,89 Interventions with brief theoretical framing could support significant changes to perspective if the rest of the intervention supported the application of this theory to practice through discussion or through experience. 60,75,82,94 For undergraduate students, experiences that enabled them to engage with patients over time 63,65,71,73,76,84 and/or take an active role in patient care, 74,83 were more compelling and memorable.
In Students remarked that despite their understanding of the pathophysiology of diabetes and its complications, they 'had no appreciation how it really impacts someone on a daily basis and how they really have to fit it into all their daily activities'. The students described themselves as 'being floored' by their discussion. 65 The interventions with the clearest qualitative data on perspective transformation towards person-centredness all offered opportunities for learners to make sense of their responses to their learning. 49,53,58-60,66,67,70,72,76,78,80,82,  A recent study of 16 medical schools curricula in Canada questioned if learners are receiving teaching on person-centred concepts. 108 A lack of explicit person-centred theory in medical education may help to explain why studies have shown that both students and doctors hold superficial and unclear understandings of personcentredness. 109,110 Learners may regard person-centred practice as 'implicit' and 'obvious' 109 in contrast to a paradigmatic shift in practice described in academic and policy papers. [15][16][17]20,27,111,112 Superficial understandings of person-centredness may explain the gap observed between clinicians' belief that they are person-centred and their clinical practice. 110 Furthermore, studies that use validated scales to attempt to measure learners' person-centredness consistently show higher scores in 'caring' compared to the 'sharing'. 4,24 It may be that 'caring' is more easily integrated into a biomedical perspective where one still sees one's role as 'doing for', whereas 'sharing' involves a shift in perspective to 'working with' which requires a facilitative, partnership approach with patients. 113,114 Without an explicit conceptual framework and evidential rationale for the theory of person-centred practice, 21 it may hard for learners to commit to person-centred professional practice.
Our results showed that when person-centred theory was applied to meaningful experiences, this triggered an understanding of the relevance of person-centredness to clinical practice. Meaningful experiences included engaging with real patient stories, opportunities to get to know patients as people over time, as well as opportunities to take an active person-centred role in their care. Our findings fit with a recent realist review that showed that access to patients' whole illness trajectories was an important context for learning about patient-centredness. 101 Our findings also support literature that calls for an increase in patient-educators 115 and active opportunities to care for patients. 116 Our results help to extend an understanding of the literature on LICs, an educational approach that has been demonstrated to enhance person-centredness, 11,12 by highlighting how these clinical placements may act as a resource for meaningful experiences.

| Support for sense making
The opportunity for learners to process their emotions and critically reflect on their responses to learning, including dissonance, allowed them to develop new meaning frameworks and self-awareness. Dobie has argued that learners can miss the fact that self-awareness and self-knowledge are crucial to person-centred practice. 117 She has called for education that supports the emotional work and critical reflection needed to develop self-awareness to be the foundation for medical educational curricular reform. More recently, incorporation of dialogic learning into medical education has been proposed to enhance person-centred practice. 118 Dialogic learning involves regular opportunities for teachers and learners to meet in non-hierarchical settings to reflect on patient experiences. In the wider health care context, Schwartz rounds, which provide multidisciplinary teams the opportunity to process their emotional and cognitive responses to clinical practice, have been shown to support person-centred attitudes. 119 The characteristics of interventions most likely to support sense making were regular small groups with relational continuity of learners and group facilitator. This continuity allowed relationships of trust to develop which enabled participants to feel safe enough to express their emotions, engage in honest appraisal and challenge their assumptions. These conditions mirror those known to foster transformative learning, which adds explanatory power to our findings. 120 Transformative learning theory states that for adult learners to change their existing meaning framework, they need support to critically reflect on their values, meanings and purposes. 121 In our review, we found that critical reflection and emotional processing took place together and this is supported by more recent work on fostering transformative learning which shows that the capacity for critical reflection may depend on the ability to process emotions. 122

| Relationship of educational interventions to whole curriculum
Our review did not find any person-centred interventions that were perceived to be in line with the wider focus of the curriculum. All interventions, including those that applied theory to meaningful experiences and support for sense making, were perceived to be at odds with the biomedical focus of the curriculum and led some learners to minimise the importance of the learning for their future clinical practice. Our recognition of the importance of epistemological dissonance reinforces calls for a whole curriculum approach to support personcentred practice. 123

| Implications for educational practice
Our findings explain why a skills-based approach may be insufficient to support the development of person-centredness in doctors. In line with constructivist learning theory, our review finds that educational interventions interact with learners' existing meaning perspectives.
Therefore, in order shift from a biomedical to a more person-centred approach to clinical practice, medical students and doctors need to understand why person-centred practice matters to health. This may be achieved by integrating explicit learning on the theory of personcenteredness with opportunities to experience its relevance in clinical practice through meaningful experiences.
Educational approaches also need to recognise and address the emotional work needed for perspective transformation. Regular opportunities to process emotions and critically reflect on responses support the creation of new meaning frameworks that enable a shift in perspective towards person-centredness. Regular, supportive small group learning, with continuity of peers and facilitators, needs to be integrated throughout the curriculum. Finally, to optimise the development of person-centredness, these components need to be integrated at both pedagogical and whole curricular level.

| Implications for research
Our realist review provides insights into evidence-based strategies that may be effective in medical education settings, and our refined programme theory offers a testable theory for medical educators and policymakers to implement in practice. Further research should test and refine our theory through empirical work to develop and evaluate educational models in practice. In particular, we call for evaluation of whether implementation of our programme theory can lead to persistent person-centredness in practice. Understanding the relevance of person-centredness to practice is closely aligned with the concept of autonomous motivation in self-determination theory. 124 Autonomous motivation develops when learners understand the value of an activity and it aligns with their sense of self, and several empirical studies have shown that autonomous motivation is highly related to persistence of an activity. 125,126 Longitudinal studies to test if applied person-centred theory can support a sustained change towards person-centred approaches would be a valuable area for future research. Furthermore, although the focus of our review was on medical students and doctors, our findings may be relevant to other health professions.

| CONCLUSION
Our findings offer explanations as to why communication skills-based educational strategies may be insufficient to develop personcentredness. Integrating person-centred experiences with theory on why person-centredness matters, and enabling support for sense making, may support perspective transformation towards person-centredness. Our findings offer programme and policymakers testable