Leadership and followership in the healthcare workplace: exploring medical trainees' experiences through narrative inquiry.

OBJECTIVES: To explore medical trainees' experiences of leadership and followership in the interprofessional healthcare workplace.
DESIGN: A qualitative approach using narrative interviewing techniques in 11 group and 19 individual interviews with UK medical trainees.
SETTING: Multisite study across four UK health boards. PARTICIPANTS: Through maximum variation sampling, 65 medical trainees were recruited from a range of specialties and at various stages of training. Participants shared stories about their experiences of leadership and followership in the healthcare workplace.
METHODS: Data were analysed using thematic and narrative analysis.
RESULTS: We identified 171 personal incident narratives about leadership and followership. Participants most often narrated experiences from the position of follower. Their narratives illustrated many factors that facilitate or inhibit developing leadership identities; that traditional medical and interprofessional hierarchies persist within the healthcare workplace; and that wider healthcare systems can act as barriers to distributed leadership practices.
CONCLUSIONS: This paper provides new understandings of the multiple ways in which leadership and followership is experienced in the healthcare workplace and sets out recommendations for future leadership educational practices and research. Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://www.bmj.com/company/products-services/rights-and-licensing/

This is the first study to explore how medical trainees experience leadership and followership in the healthcare workplace through narrative analysis.

The large number of narratives across multiple UK sites and different specialties, enhances the transferability of our findings.

We acknowledge that our construction of the leader-follower dyad in this study may have influenced our study findings (eg, encouraging further co-construction of leader-follower dualism by participants).

We had relatively low numbers of male, non-white and foundation trainee participants so our findings are most relevant to female, white and specialty trainee doctors.

Introduction

In recent years, there has been an upsurge in the use of the term ‘leadership’ to describe a range of activities connected to the organisation of patient care.1 ‘Leadership’ is no longer attributed solely to those in formal leadership positions, but is seen to be the responsibility of healthcare professionals across all levels of healthcare organisations.1–5 Notions of traditional hierarchical practices have given way to arguments for distributed (or shared) leadership models. Modern theoretical discourses assert that leadership is a process involving leaders and followers acting within a fluid context so that people construct leader or follower identities moment-to-moment.6 7 Suggested benefits of such distributed leadership practices include improved patient experience; reduced errors, infection and mortality; increased staff morale and reduced staff absenteeism and stress.8 Using this ‘leadership lens’, those in non-formal positions of healthcare leadership (eg, medical trainees) are expected to undertake leadership throughout their careers and develop their leader identities.9

While the healthcare literature contends that effective distributed leadership practices are necessary to improve healthcare workplace cultures, patient safety and quality of care, little is known about how these leadership processes are experienced by medical trainees. Within this paper we seek to understand better how the notion of ‘leadership’ has been embedded into frontline healthcare practice through narrative analyses of how medical trainees’ leader and follower identities are constructed.

Researching distributed leadership processes

Until recently, the focus of scholarly activity in leadership has been on individuals as leaders.10 11 In healthcare, many studies have concentrated on the role of leaders rather than the processes of leadership.12 13 However, many leadership theorists criticise leader-centric research for its emphasis on individuals as leaders, how effective their activities are and how others (followers) act in response to their influence.11 14 Rather, leadership theorists argue that leadership can only be understood through exploring the underlying social systems in which leadership happens.6 15 As a product of co-construction, leadership is perceived as an on-going negotiation as part of a multifaceted interaction between social beings.16 Each interaction can be seen as socially and historically bound operating through language, within a socially-constructed context.16 17

Some studies have explored leadership processes and the link between senior clinicians and the wider organisation. For example, combining interviews and observation, MacIntosh et al18 identified the extent to which interactions between clinicians and managers were two-way discussions, finding that each group presented themselves as less powerful than the other group and lacking agency. They described the clinician-manager relationships as having potential to limit the opportunities for distributed leadership processes. In addition, Martin et al5 revealed through interview and observation, that there was the potential for a disconnect between the desire for distributed leadership within healthcare and actual organisational practices.

While these studies have focused on wider organisational leadership processes our focus is on leadership that may occur day-to-day within the clinical context, where medical trainees potentially have their first experiences of leadership. Through this, our study seeks to add to the literature on distributed leadership in healthcare.

Aim and research questions

This study aimed to explore how medical trainees experience leadership and followership by asking two research questions. What are medical trainees’ lived experiences of leadership and followership in interprofessional healthcare workplaces? How do medical trainees construct their identities as leaders and followers within their narratives of interprofessional healthcare workplaces?

Methods

Study design

We undertook a qualitative study using group and individual interviews to elicit medical trainees’ personal incident narratives (PINs) of their experiences of leadership and followership. Ascribing to the notion that meanings are constructed by people as they interact with the world around them, our study draws on social constructionist epistemology.19

We used narrative inquiry methodology. Narrative accounts of the healthcare workplace offer abundant resources for research.20 The narratives referred to in this paper are short, about discrete events and recounted in interactions in various contexts as sense-making tools.21 22 A narrative in this form makes the self the central character (or protagonist), either playing an active part within the story or as Chase describes as an: ‘interested observer of others’ actions’ (ref. 23, p. 657).

A narrative is the shared construction between the narrator and his/her audience.21 23 24 Bound to this is the context in which the narrative is shared; the specific setting, audience and the reason the story is told.23 24 Pivotal to our paper is the concept of the ‘narrative turn’ in that narrators construct events through their story, expressing their feelings, beliefs and understandings about leadership processes.24 As such, the narrative becomes a construction of who the narrator is, who they wish to be and how they wish to be seen.25 In other words, when a story is told, the narrator constructs and presents identities, events and realities in interaction with others.23 24 Thus, paying attention to and asking questions not only about what participants experiences are but also how they narrate their leadership experiences can afford insight into the multiple identities that medical trainees construct as leaders and followers.23

The research team

The research team included three members with health professions backgrounds (one practicing general practitioner; one ex-physiotherapist; and one ex-clinical psychologist) and one social scientist. Team members had various personal experiences of leadership and management covering clinical, research and educational leadership, with all team members teaching leadership in healthcare at undergraduate and postgraduate levels.

Sampling and recruitment

On receiving ethical approval and appropriate institutional consents we utilised maximum-variation sampling to ensure a diversity of medical trainees in terms of their stage of training, specialty and location. Following an initial recruitment drive by email, we recruited further participants using flyers at trainee teaching sessions and snowballing.26

Data collection

We conducted 11 group (with between three and seven participants) and 19 individual interviews with 65 medical trainees (25 male: 40 female, 51 white: 14 non-white) from both early-stage (34) and higher-stage (trainees beyond the half-way point: 31) postgraduate medical training. Our initial aim was to have only group interviews as from a social constructionist perspective they had the potential for participants to build on each other's ideas. However, convenience for participants meant that individual interviews were also necessary. Our sample came from four UK Health Boards and a range of specialty trainee groups including: general practice (GP: 23); medicine (13); surgery (11); and service-orientated trainees such as anaesthetists, radiology and pathologists (10). Our sample also included eight foundation doctors.

We devised an interview guide which was developed in line with our research questions, which provided an aide memoir. The semistructured nature of the interviews allowed for flexibility so that ideas could be pursued and expanded on. The interviews were broadly split into two sections: first, we asked participants to articulate their understandings of leadership and followership (reported elsewhere).27 Following this, narrative interviewing techniques were used to collect narratives of participants’ experiences of leadership and followership in the interprofessional healthcare workplace. All interviews were audio recorded (with permission) and independently transcribed. All interviews bar one were conducted by the primary author. The second author conducted an early group interview and listened to several initial interviews with the primary author to reflect on the structure and relevance of the interview schedule (thus enhancing research rigour).

Data analysis

We began our analysis with thematic framework analysis.28 This allowed us to identify patterns across the data. We constantly familiarised ourselves with the data through repeated reading of transcripts and listening to audio recordings. A team data analysis session was held which provided opportunity to discuss and negotiate possible themes to be included in the thematic coding framework. Prior to the session, a subset of data were analysed separately by each team member. Through an iterative process of discussion, feedback and agreement within the team, a coding framework was developed which was then used to index the data. To identify narratives, we drew on Labov's construction that a narrative is a structured account of an incident that has become part of the biography of the storyteller.29 It is increasingly common within qualitative research to explore patterns across data through the use of computer-assisted qualitative data analysis software (CAQDAS). We used Atlas-ti (V.7.2) in our identification, time-stamping and coding of all narratives.30

Using the premise that identities are formed through talk and interaction we explored the interplay between different thematic groupings.24 To do this, we used a form of structural narrative analysis which pays attention to the ways in which narratives are organised.21 Labov states that a fully formed narrative includes seven elements: (1) abstract; (2) orientation; (3) complicating action; (4) evaluation; (5) most reportable event; (6) resolution; and (7) a coda.29 Not all stories however will contain all elements and often elements occur in different sequences, with narrators moving back and forth, providing further complicating actions and evaluations as they make sense of the story.30 Thus, we were able to explore how participants constructed their identities as leaders and followers within their narrative, what parts of the story participants constructed as important, and how the narrator used language to illustrate how they evaluated the events.31

Results

Across the data set, we identified 171 distinct narratives. Initial thematic analysis identified three different sets (or groupings) of themes. Contextual themes for the narratives provided orientation to the timing of the events; where the events took place; how the narrators positioned themselves in the story (eg, as leader, follower or observer); the type of activity that was being undertaken when the event occurred; and how the narrator evaluated their experience (eg, positively, negatively or neutrally) through their commentary on the events. The second group of themes focused on the content of the story and signposted its gist (ie, the main plotline of the story). Finally, process-orientated themes focused on how the stories were narrated. This set of themes highlighted, for example, linguistic features used by narrators to articulate their stories.

What are medical trainees’ lived experiences of leadership and followership in interprofessional healthcare workplaces?

Contextual themes

Participants most often constructed themselves as followers within the stories (n=80), with around half as many constructing themselves from the position of leader (n=41). Of the 171 narratives, 144 were set in the hospital, with only 12 set in GP practice. However GP trainees offered the highest proportion of narratives across the specialties (sharing 53 narratives, of which 36 were hospital-based).

The activities on which stories centred were wide-ranging: they were most likely to come from the clinical environment and be related to clinical leadership activities (n=119). This included stories about complex patient scenarios, which participants deemed to be extraordinary (n=37). Still related to clinical leadership, were stories about acute emergency scenarios (n=32) and routine patient care (n=29). Data also included stories about formal ward-based activities such as planned team meetings and ward rounds (n=15). Narratives were evenly balanced between positively and negatively evaluated experiences (80 positive; 77 negative).

We identified two overarching themes for the content of the narratives (Static leadership relationships and Emergent leadership relationships) and a series of subthemes as defined in table 1. We also identified three key process-orientated themes: pronominal; emotional; and metaphoric talk. What follows in this section is an overview of each of these themes with illustrative data excerpts presented in box 1.

Table 1

Narrative content themes and subthemes

Theme: Static leadership relationships (n=131)	Definition
Subthemes
Facilitated by supportive dialogue or behaviours (n=25)	Leaders are perceived to take part in supportive behaviours or dialogue through revealing fallibility, listening, accommodating, being fair, responsive or showing empathy
Inhibited by unsupportive behaviours or lack of dialogue (n=21)	Leaders are perceived to be unsupportive and lack dialogue with followers. This is carried out through them being unfair, not admitting fallibility, not listening, being unresponsive or lacking empathy
Abusive (n=21)	Abuse was constructed through the actions of leaders including undermining, verbal abuse, physical abuse, humiliation and/or criticism
Inhibiting team-working (n=14)	Participants described instances of poor team working, often with conflict/disagreement being described or a lack of inclusivity
Conflictive decision-making (n=12)	Trainees described those perceived to be leaders in conflict/disagreement with each other about patient care
Fostering constructive team-working (n=8)	Team-working was described that was collaborative and perceived to be conducive to good patient care
Ineffective due to unclear role definition (n=7)	Described when there was a perceived lack of leadership or when too many people were trying to take on the leadership role
Effective, based on clearly defined roles (n=6)	Roles here were defined often as a result of having time to prepare for the situation. For example, a multiple trauma coming into accident and emergency
Identified through traditional clinical roles (n=6)	For example, Doctor as leader, nurse as follower
Collective decision-making (n=5)	Sharing group goals, all team members working towards the same goal and with an appropriate allocation of tasks
Identified through traditional hierarchies (n=4)	The most senior person present was seen to automatically take the lead. Assumed through traditional hierarchies
Effective, based on practiced protocols (n=2)	This often related to cardiac arrest scenarios in which protocols are practiced and the scenario is seen to ‘run’ ‘smoothly’ due to repeated practice of these scenarios
Theme: Emergent leadership relationships (n=40)	Definition
Subthemes
Facilitated by individual knowledge or experience (n=21)	An individual will ‘step into’ leadership based on previous experience or knowledge. Leadership can sometimes come from unexpected sources and does not necessarily follow traditional hierarchies
Facilitated by lack of engagement of expected leader (n=9)	Trainees described being ‘pushed into’ a leadership role due to lack of engagement of a perceived leader. Sometimes the perceived leader can ‘hand leadership back to the junior’. Trainees are not actively seeking to take on leadership but sometimes circumstance requires them to do so
Facilitated by systems and protocols (n=5)	For example, trainees used protocol to support a change in clinical care and take on leadership
Facilitated by timing (n=3)	Owing to the timing of incidents, trainees take on leadership for example, at night
Inhibited by lack of knowledge or experience (n=1)	Trainees describe an individual who ‘steps into’ the leadership role but is unable to take on that role due to lack of experience or knowledge
Inhibited by systems and protocols (n=1)	Where systems do not allow leadership to emerge (eg, consultant to consultant referral systems.) Often this was linked to perceptions of traditional medical hierarchies

Excerpt 1: “…the registrar who was there…who in that situation [cardiac arrest] was leading the team…was very good at…knowing what everyone's limitations were and…telling you to do things without patronising you or making you feel silly…you could see the medical student was…looking petrified…he [registrar] gave them a job to allow them to feel involved but…not get too involved that they got scared…it was really something to learn from” (Female, foundation trainee)

Excerpt 2: “…a consultant who didn't come up to the ward…when he did come up he was never that fussed if you were with him or not…he would just leave you a list of things to do…it was a bit disheartening…you were never…completely reassured about what you were doing…you never really get to grips with his actual overall plan…so you end up not feeling that important part in a team because it doesn't matter if you were there or not…you've to follow blindly what he wrote…” (Male, early-stage medical trainee)

Excerpt 3:“[during ward round] he [the consultant] asked everyone to leave…which was a bit mortifying…a nurse and student and possibly even somebody else in the room at the time and he asked them to leave and told me off for something…I asked him questions about it because I didn't…understand what he was talking about because it made no sense to me…and so I asked him…but he was extremely emotional and angry about it and stormed out…” (Female, higher-stage surgical trainee)

Excerpt 4“I was there [GP practice] for one day…a patient who come screaming… this patient was in labour and she [the GP] have no experience in doing labour at all… then I came to the room…the GP stepped down and I led the team that I'm doing the examination and assessment and then get each one of them a job… So it doesn't mean that the most senior who can do it. I was not senior, I was trainee….And they agreed, and accepted it…they stepped back and then they let me just give clear instruction…they doesn't know what to do…because I was just finished my gynae training” (Female, higher-stage GP trainee)

Excerpt 5“… although everyone was kind of looking at me…I noticed that one of the nurses was managing to get through to the patient and kind of getting him to listen …and I thought, I think ‘this is when I should be quiet and let this nurse deal with it’ and I just did what the nurse said…” (Female, foundation trainee)

Excerpt 6“…we [junior trainees] were kind of saying,‘well I don't think this person needs resuscitating because they got quite unwell’ and he [consultant] just didn't have a clue what to do. He [consultant] said, ‘well that's not really my, my area’ and kind of then deposited it back to us and said ‘well if you think they're not for resuscitation, then you take the decision’…but then he was kind of not really living up to what we'd hoped…” (Female, early-stage GP trainee)

Static leadership relationships

Static leadership relationships was the dominant content-related theme (n=131). Here, the identity of the leader and follower/s remained static throughout the story and trainees typically narrated from the position of follower. These leader-follower relationships were based on the traditional hierarchies found within the healthcare workplace. From this, we identified 12 subthemes, which focused on leader behaviours within the stories and which were seen to be facilitative or inhibitive to good leader-follower relationships (see table 1).

Here, we talk in-depth about the most common three themes only. Excerpt 1 (box 1) illustrates a facilitative subtheme, where the leader is seen to be entering into a supportive dialogue or behaving in a supportive way to facilitate leadership processes and leader-follower relationships. Here, the leader is perceived to have acted in the best interests of the patient regardless of the outcome. As a follower, this participant describes how she felt valued, respected and supported within this relationship, that it was conducive to learning and that this type of relationship was something to aspire to.

In contrast, excerpt 2 (see box 1) illustrates an inhibitive subtheme in which the leader is seen to be unsupportive and lack dialogue with the trainee as a follower. The trainee reports that these behaviours led to lost confidence and feelings of non-validity and exclusion. This is reported to be detrimental to training experiences and at times detrimental to patient care.

We also identified abuse narratives (see excerpt 3, box 1). This subtheme categorised narratives around direct and indirect experiences of what constituted abuse as perceived by participants (this included undermining and humiliation). As well as being the recipients of abuse, participants witnessed others being abused too. The abuser was most often identified as the consultant but trainees also reported abuse from other more senior trainees and nursing staff. Abuse typically revolved around clinical leadership and during routine patient care (eg, surgical theatre) or formal activities such as the ward rounds or meetings, as illustrated in excerpt 3 (see box 1).

Participants often reported negative emotional responses to these experiences: they talked about feeling humiliated and non-human, sometimes getting angry themselves; the need to keep going and ‘survive’ training; and being careful to avoid situations in which abuse was likely to happen/occur.

Emergent leadership relationships

A smaller proportion of narratives (n=40) were coded to the content-related theme: Emergent leadership relationships. Unlike the previous theme, in which the identities of the leaders and followers were static, here participants recounted how the combination of individuals involved, the context (including the task), the relationships within that context and the wider systems affected who emerged as the leader within the experience. Leadership emergence was more likely to be categorised in narratives that related directly to patient care scenarios, namely: complex patient cases; routine patient care; and acute emergency care. Interestingly, no narratives in which formal clinical activities were being undertaken were coded to this theme, indicating more static (and possibly traditional) leader–follower relationships within more formalised clinical settings.

We identified six subthemes within these narratives which described the factors either facilitative or inhibitive to leadership emergence (see table 1). Unlike the static leadership relationship subthemes, medical trainees typically narrated emergent leadership relationship narratives from the position of leader. Excerpt 4 (see box 1) illustrates an example in which emergent leadership relationships were facilitated by individual knowledge or experience. The trainee describes an incident where, as a junior trainee, her broad-based training experience made her more ‘expert’ than those with more specialised training.

Often stories within this subtheme were interprofessional. Participants narrated incidents in which more experienced nurses and other members of the interprofessional team took on leadership. Participants narrated this as emergent because they were working in the context of traditional interprofessional hierarchies, which meant that they thought doctors were expected to lead. Perhaps unsurprisingly, participants inevitably saw leadership emergence occurring in the best interests of the patients, as illustrated in excerpt 5 (see box 1).

In order for leaders to step forward and out of traditional hierarchical boundaries, participants narrated the process of traditional leaders needing to ‘step back’ (as illustrated in excerpt 5, above). This was sometimes perceived to be difficult. At times, participants narrated experiences in which traditional leaders stepped back through their non-engagement, as illustrated in excerpt 6 (see box 1).

Process-orientated themes

From the position of follower, participants often used the pronouns ‘we’ and ‘us’ to describe themselves and their contemporaries, and ‘them’ and ‘they’ to describe a group of leaders within their narratives, indicating a perceived separation (and potentially adversarial relationships) between followers and leaders. This was particularly apparent within the negatively evaluated narratives (see box 2: excerpt 1).

When followership experiences were evaluated more positively and the leadership process was seen to go well, pronouns ‘we’ and ‘us’ would be used to describe the whole team, including both leaders and followers together (see box 2: excerpt 2). From the position of leader, participants often used the pronoun ‘I’ when describing leadership decisions, which seemed to indicate their agency and autonomy within the situation (see box 2: excerpt 3).

Pronoun use:

“… (Female, early-stage GP trainee)

“…discussion about a decision that had been made…It was like, ‘well (Male, early-stage surgical trainee)

“ …‘bring back information about that [a patient decision] and (Male, higher-stage medical trainee)

Emotional talk:

“it was probably the (Male, early-stage medical trainee)

“it was (Female, higher-stage service trainee)

Participants typically used positive emotional talk within stories that were evaluated positively (box 2: excerpt 4) and negative emotional talk (box 2: excerpt 5) within stories evaluated as negative experiences.

Across the narratives, we identified hundreds of metaphoric linguistic expressions (MLEs).32 Although it is outside the scope of this paper to present a full systematic metaphor analysis, we identified broad groups of conceptual metaphors which revealed participants’ understandings of leader–follower relationships.32–34 We identified eight overarching conceptual metaphors used to describe the leader–follower relationship. These were LEADER-FOLLOWER RELATIONSHIP AS: WAR; HIERARCHY; PARENTALISM; SPORT; CONSTRUCTION; MACHINE; JOURNEY; and TRANSACTION (the convention of cognitive linguistics requires that conceptual metaphors are presented in small capitals: see table 2).34

Table 2

Metaphoric talk excerpts

Conceptual metaphors and examples	Identity construction
Leader–follower relationship as war
“…first thing he [clinician-manger] started to do is just to…attack, attack other consultants…and started to just to stab…even in the same team you're supposed to share the same targets and because if the ship sinks, everybody will sink”(Male, early-stage GP trainee)	Leader as aggressor; followers as victims; department as ‘ship’
Leader–follower relationship as hierachy
“I mean there's one person…if it was monkeys in the zoo, there's one person who's very much the dominant personality, and the alpha male…he's very clear at any kind of whole unit meeting that, you know, this is his view, and he'll shout it from the rooftop” (Male, early-stage medical trainee)	Leader as alpha monkey; followers as subordinate lower-ranking monkeys; Leader is also ‘higher up’ (as in, on the rooftop)
Leader–follower relationship as parentalism
“There's no doubt about it, she gave 110 per cent to her patients. And we used to talk about ourselves as students, and we wanted to be like her when we grew up” (Female, early-stage GP trainee)	Leader as adult; followers as children; leader is also role model
Leader–follower relationship as sport
“And the number of times you felt like a piggy in the middle. You were being batted backwards and forwards. At the end of the day you're just trying to do the best for the patient who is outside your expertise” (Female, early-stage GP trainee)	Leaders as game-players; followers as objects to be played with (ie, ball)
Leader–follower relationship as construction
“Other ways of optimising influence of people around you, you just kind of like learn gradually through working, through building working relationships…” (Male, higher-stage surgical trainee)	Leader as builder; leader–follower relationship as the object to be built
Leader–follower relationship as machine
“…it works really well and very efficiently and suddenly everyone kind of clicks into gear…the senior registrar will be running it…” (Male early-stage GP trainee)	Leader as machine operator; followers as parts of the machine
Leader–follower relationship as journey
“He's been a good enough leader that day, we'll do it, we'll go that extra mile for him.” (Male, higher-stage surgical trainee)	Leader as destination; followers as travellers
Leader–follower relationship as transaction
“…good managers are probably effectively good sales people so you have to treat each person individually…” (Male, early-stage surgical trainee)	Leader as seller; followers as consumers

How do medical trainees construct their identities as leaders and followers within their narratives of interprofessional healthcare workplaces?

Here, we pull together both content and process-related themes to present a more detailed exploration of one narrative. This narrative from ‘Carol’ (a pseudonym, female early-stage GP trainee) comes from an event she experienced during her time as a trainee in psychiatry. It is not explicitly clear what grade of training Carol was at the time of her story but through her use of language and the events narrated she is clearly junior to the narrative protagonists (see table 3). Carol presents a complex patient scenario in which people in contact with a particular patient needed prophylactic treatment for meningococcal disease. The focus of Carol's story is on her personal experience of trying to take leadership as the consultant (a psychiatrist) states that he does not have the necessary experience to do so. Carol describes how attitudes, systems and protocols become barriers to fully undertaking leadership in this scenario. Indeed, the key message of Carol's narrative is that her ability to take on leadership is inhibited by the wider systems in which she works. We identified other content-related themes, which could facilitate and/or inhibit leadership processes. Facilitative aspects included her emergence as leader due to her own knowledge and expertise, in contrast to the consultant's unwillingness to engage in leadership due to his own lack of experience, leading him to support her leadership emergence. Inhibitive aspects include Carol's description of traditional systems, protocols and expectations that the consultant should be leading.

Table 3

‘Where's your daddy?’

Lines	Narrative*
	Evaluation #1
1234567891011121314151617181920212223	Carol: I had an interesting thingOrientation #1when I was doing psychiatry and we had a patient who, they phoned me on a Saturday at lunch timeComplicating Action #1to say that she'd tested positive for meningococcus and it was, eh was in air, in a sputum sample, so it would have aero-cisedMost Reportable EventSo they said everybody who'd been [on] the ward for the past 2 weeks plus relatives plus staff all had to get, erm, prophylactic treatment for meningococcusResolution #1And I mean I was phoning sort of infectious disease and occupational health and public health and everybodyComplicating Action #2And a lot of them were saying, well look we want to, we want to speak to your consultantEvaluation #1But the consultant hadn't done anything that wasn't psychiatry for 40 years and he had no idea what to doComplicating Action #3So literally the consultant was sitting in the corner of the office signing prescriptions for me, while I organised everythingEvaluation #2because he was, he was largely uselessComplicating Action #4Erm, but they kept sort of saying, ‘well does your consultant know what you're doing? Can I speak to someone more senior?’Interviewer: And how did you find that then? How did you find that situation?Evaluation #3It was difficult becauseOrientation #2I, I mean I'd, I'd phoned the consultant first just to let him know, and he just said, ‘well look I have no idea what to do, can you manage this?’Evaluation #5And I was really confident that I could manage it myself, and I did. But it was just sort of, the people looking over your shoulder going, sort of ‘Where's your daddy’ kind of thing. It was just, it's a bit frustrating. And it sort of undermines how you feel as a leader a wee bit*Editing notes: …=speech edited out for brevity; Metaphors underlined; Interesting pronoun use in bold; Emotional talk in italics

Carol describes different interactions with different sets of actors in her narrative. First, there is the interaction between herself and her consultant. Second, there is the interaction she has with a group of people (ie, ‘the people’, line 21) she repeatedly describes as ‘they’ or ‘them’. Throughout the narrative she is not specific about who ‘they’ are. However, in lines 7 and 8, she lists a group of specialties and we assume that ‘they’ are included within this group. In the following paragraphs, we explore how Carol constructs her identities as leader and follower in relation to these two interactions.

Carol as a capable leader, the consultant as supportive follower

When Carol narrates her interaction with the psychiatry consultant, she constructs herself as a confident and capable leader. She describes a discussion with her consultant in which responsibility for dealing with the situation becomes hers (lines 19 and 20). When evaluating this event, she uses positive language with intensifiers, to construct herself as ‘really confident’ that she can handle the situation (line 21) and qualifies this with the short statement that: ‘I did’. Within this interaction, Carol identifies herself as the leader through regular use of ‘I’ to indicate her agency and control of the situation (lines 7 and 12). Despite her describing the consultant as having a part in the process (eg, writing prescriptions, line 12), she chooses not to use the pronoun ‘we’, suggesting that she does not see herself and the consultant as a team. Possibly to reinforce her own identity as a strong leader, she constructs her consultant as a supportive follower who undertakes tasks ‘for me’ (line 12). This is emphasised early in the narrative through her use of derogatory language to describe the consultant as having ‘no idea what to do’ (line 10) and as being ‘largely useless’ (line 14), constructing him as lacking both clinical knowledge and leadership.

Carol as child, consultant as ‘daddy’

The second type of interactions Carol has within this narrative (with other healthcare professionals as part of workplace systems and cultures) reveals a contrasting picture relating to her and her consultant's identity construction. ‘Their’ express desire is to speak to the consultant (lines 9), and ‘they’ ask Carol if her consultant knows what is happening (line 15). Carol's pronoun use at this point places distance between her and this wider group and gives this part of her narrative a confrontational feel. Carol narrates that she is powerless to change systems which expect the consultant to be the leader and thus within this interaction, Carol shifts her own position from leader to follower. Carol uses negative emotional language such as ‘difficult’ (line 18) and ‘frustrating’ (line 23) to express how she finds this and how this interaction affects her identity construction.

Key within this narrative is Carol's use of a metaphoric linguistic expression (MLE), which reveals how she thinks her and her consultant's identities, and their relationships, are ascribed by others as: LEADER-FOLLOWER RELATIONSHIP AS PARENTALISM (line 22). Through her use of this MLE (‘where's your daddy?’), she constructs her identity as ascribed by others as a young child and the consultant as her male parent. This reveals that she thinks others see her as junior within the wider healthcare system. Indeed, Carol states that the consultant is required to sign the prescriptions in order for the task to be fulfilled and thus protocol reinforces this traditional hierarchy (line 12), reinforcing her lowly position within the hierarchy. She finishes the narrative by expressing that this (childlike) identity imposed on her ultimately undermines how she feels as a leader (lines 23).

In summary, this narrative reveals a complex and at times contrasting interplay between individuals, context, relationships and systems, which seem to simultaneously facilitate and inhibit Carol's emerging leader identity. As an individual, Carol feels confident that she can cope with the situation and her relationship with the consultant is such that she feels able to move away from traditional hierarchies to take control, which within this context she feels is appropriate due to her superior knowledge of how to approach the situation. However, through this narrative Carol also describes the frustrations of trying to take on leadership in a wider healthcare system in which protocols and traditional hierarchical attitudes prevent her from fully undertaking the role and ultimately positions her as ‘childlike’ and ‘undermined’.

Discussion and Conclusions

Despite the espoused rhetoric of distributed leadership relationships within the healthcare literature, our data indicate that static and hierarchical leadership relationships remain the norm.4 8 9 We argue that a workplace perpetuating leadership processes embedded in static leader–follower relationships has the potential to be prescriptive about the division of labour and may be inflexible to innovation with potential adverse implications for patient care.6 8 11 Within the ‘static leadership relationships’ theme, the focus of the content-related subthemes was often on evaluation of whether individuals were ‘good’ or ‘bad’ leaders. Schyns and Miendl suggest that leaders are evaluated through followers’ ideas about leadership that have been formed through previous experiences as part of professional socialisation.35 Leaders are thus linked to pre-existing prototypes.11 36 Included in this were narratives about abusive leader–follower relationships. Experiences of abuse are reflected in findings from both undergraduate and postgraduate studies, in which abuse within the healthcare workplace has been narrated by medical and other healthcare students and trainees.37–40 It would appear that abuse continues to be experienced in the postgraduate sphere.

Our findings revealed that participants most commonly drew on their experiences of clinical leadership in hospitals. Despite the traditional notion that leadership is focused on organisational change, ‘leadership’ within these narratives was about ‘influential acts of organising’ (IAOs) that happened day-to-day in the healthcare workplace.41 This difference could perhaps be explained by the recruitment of participants from out with traditional positions of organisational leadership, which meant their focus was on ‘everyday’ leadership experiences. In previous leadership research, narrative inquiry has been limited to the broader narratives of an organisation or the life story of a leader.42 43

Less than a quarter of the narratives were about ‘emergent leadership relationships’, most of which involved complex patient scenarios. It can be suggested that these scenarios could be seen as ‘non-linear’ or ‘wicked’ problems, requiring emergent leadership relationships that can happen out with traditional hierarchies.17 44 Key to this was the assumption that actions were in the best interests of the patient (thus, these emergent relationships can be seen as patient-centred). However, our findings show that at a local level, there is some work to be carried out on cultural shifts toward what could be described as distributed leadership patterns.

In the leadership literature there is a lively debate around the broad metaphors of ‘leadership’ as an overall phenomenon.16 45 However, our analysis was concerned more with the conceptual metaphors used as part of language-in-interaction and what this revealed about leader-follower relationships. This, along with pronominal and emotional talk, gave insight into how participants evaluated these relationships. To the best of our knowledge, no one has explored previously metaphoric talk identified within leadership narratives specifically, although similar talk has been found describing the student/doctor–patient relationship and the student–doctor feedback relationship at the undergraduate level.31 33 34

Carol's narrative revealed an unpredictable situation in which change (and learning) was required in response.46 Through her narrative, Carol constructed her identity as a leader out with her formal position and thus leadership was ‘emergent’.46 Carol's narrative also revealed the ‘enabling’ identities of her consultant, who had to act as a bridge between the administrative structures within the organisation and the adaptive leadership required to solve the issue faced.46 Our narrative analysis also revealed the potential for ‘disconnect’ between the emergent leadership expectations of the immediate context and those of traditional hierarchies within the wider organisation, which firmly positioned Carol as a follower. These barriers could explain why static leader–follower relationships perpetuate as the norm and echo the work of Martin et al.5

Narrative inquiry is not novel in the field of medical education. It has been utilised to analyse, for example, professionalism dilemmas; prescribing experiences; and feedback experiences.30 31 38 47–49 Through narratives, in the current study, participants had the opportunity to develop their own voice as they constructed their stories, others’ voices and multiple realities.23 This approach was particularly valuable when considering that some narratives including those describing workplace abuse were also evaluated negatively. Indeed, participants may not have had the opportunity to discuss such distressing matters before the study and had this study not been conducted. In fact, we suggest that these narratives, at times, became ‘acts of resistance’, which challenged traditional hierarchical conceptualisations of leadership.50 Such ‘resistance’ was possibly a conscious act to ‘subvert’ asymmetrical power relationships that were constructed through traditional healthcare hierarchies (ref. 50, p.433). Narrative analysis has also highlighted the potential use of narratives for educational purposes. Through story-telling, participants repeatedly evaluated their experiences. Using narratives in this way would provide opportunities for learners to evaluate and make sense of their leadership experiences and reflect critically on their developing identities as leaders, followers and doctors, exploring opportunities for on-going development and building on their understandings of leadership processes.51

Our research is not without its methodological challenges. We acknowledge that we chose ‘leadership’ as a specific lens for our study. Recent critical leadership literature argues that the rise of ‘leaderism’ in healthcare discourse has meant that many things that were more traditionally aligned to, for example, ‘interprofessional relationships’ or ‘clinical decision-making’ are now being branded as ‘leadership’.1 52 Asking participants specifically about their leadership experiences may have perpetuated this leaderism discourse in our findings. We acknowledge that prior to eliciting narratives from participants, we drew participants’ attention to leader–follower dualism as a concept through the preinterview participant information sheet and questioning them about their understandings of ‘leadership’ and ‘followership’. We therefore potentially influenced participants through our socially constructed conceptualisations of leadership and followership. In particular, the notion of ‘followership’, an uncommonly used or understood term in healthcare is something we sensitised participants to and they often struggled to define (reported elsewhere27).

As a research team, although we had diversity in terms of our professional backgrounds (see methods), we were not diverse in terms of other identities such as gender and ethnicity (we are all female and white). Our interpretation of the data will be influenced by own understandings and experiences of leadership, and these will inevitably be coloured by things like our gender. We also acknowledge the lower proportion of male, non-white and foundation doctors within our sample, meaning that our findings may be less transferable to these groups. While interview methods helped to reveal a complex picture of the interprofessional healthcare workplace it also exposed a limitation in our study, in that we sought only to interview medical professionals. Broadening our narrative interviews to take into account the whole interprofessional team would have enriched these data and should be considered for future interview research. In addition, although our study was multisite, it was conducted in the UK so the findings might be less transferable to other countries with different healthcare and healthcare education systems.

We used a qualitative, process-orientated approach to our research. We acknowledge that the use of numbers within the presentation of our data has the potential to draw criticism. However, it is not unusual for some qualitative researchers to draw on numbers to look at patterns in large qualitative data sets.30 We used numbers in our study to elucidate patterns (similarities and differences) that would not have been apparent with text alone.

The large number of hospital-based narratives in comparison to community-based narratives in our study did not allow us the opportunity to explore whether clinical settings made a difference to participants’ experiences, something that should be considered for further study. Indeed, although GPs were a large proportion of our participant group, in the UK they spend 18 months of their 3-year training in hospitals, following a 2-year (largely) hospital-based foundation programme. This time spent in hospitals will have undoubtedly influenced the setting of their stories.

In addition, although narratives were identified in the same way for individual and group interviews, we did not note, nor formally explore any differences in data between group and individual interviews (as this was not relevant to our research questions). However, this could warrant exploration in further research. Finally, our cross-sectional data did not allow for exploration into how experiences of leadership and the formation of leader identities might change over time as doctors move through training and this should be considered in further (longitudinal) research.

In conclusion, this study has led to better understandings about participants’ multiple, constructed realities of leadership and followership in different healthcare contexts. The findings from this study reveal that many factors influence developing leader identities; that traditional medical and interprofessional hierarchies persist within the healthcare workplace; and that wider healthcare systems can act as barriers to distributed (or shared) leadership practices. Collecting and analysing narrative data provided us with new understandings of the multiple ways in which leadership and followership is experienced in the healthcare workplace. Exploring the interplay between both what the narratives contained and how the narratives were told, provided unique insights into how narrators constructed their identities as leaders or followers against the backdrop of a complex healthcare workplace.