“With a grain of salt”? Supervisor credibility and other trainee concerns when seeking clinical oversight, support and advice: a focus group study of Australian general practice trainees.

Background ‘Ad hoc’ help-seeking by trainees from their supervisors during trainee consultations is important for patient safety, and trainee professional development. General practice trainee decisions about seeking (and responding to) supervisor oversight, support and advice may be influenced by a number of concerns, including reservations about the credibility of their supervisors. Methods Focus groups with general practice trainees were undertaken to explore their help-seeking decisions during consultations, and their assessments of the credibility and value of the help provided by their supervisors. All data was audio-recorded and transcribed, coded using in-vivo and descriptive codes, and analysed by the constant comparison of provisional interpretations and themes with the data. Findings are reported under the over-arching categories of help-seeking objectives, activities and outcomes. Results Early in their general practice placements trainees need advice about practice facilities, and the “ complex maze ” of local patient resources and referral preferences, as well as unfamiliar clinical presentations and management decisions. Subsequent help-seeking tended to be characterised informally as “ having a chat ” or “ getting a second opinion ” so as not to “ miss anything ” when trainees were “ not 100% sure ”, although trainees continued to emphasise the importance of being (and demonstrating that they were) safe. Time and workflow constraints had a powerful influence on trainee help-seeking decisions, and patients also exerted considerable influence. Trainees assessed the credibility of supervisors in terms of their approach to risk and their clinical expertise. The latter was often believed to be uneven. Several trainees reported seeking other opinions when they were uncomfortable with supervisor advice. Conclusion How a supervisor responds when their help is sought strongly influences the trainee’s subsequent help-seeking. Trainees prefer to seek help from supervisors who respond promptly and observe the etiquette of providing help in front of patients, and whose advice they trust. Trainees learn through help-seeking to make their own clinical calls but remain curious about how other general practitioners practise. Several trainees expressed great respect for their supervisors and valued opportunities to observe their clinical expertise. These important learning opportunities should

general practitioner expertise still found in some hospital cultures (14).
The contested nature of general practitioner expertise itself is also a factor. There have been a number of attempts to capture general practitioner expertise in terms of 'whole person medicine' (15), 'the integration of the biological and the biographical' (16), and 'narrative medicine' (17), but even with these efforts it remains unclear how to recognise and evidence such expertise. A further factor is that supervisors may downplay their expertise in their scripts for providing in-consultation assistance to trainees, as they try to promote trainee autonomy and avoid undermining patient impressions of trainee competence (18).
In summary, general practice trainees may have a number of concerns, including reservations about the credibility and expertise of their supervisors, which are likely to have an impact on their helpseeking behaviour. The aim of this study was to explore general practice trainee help-seeking decisions, particularly ad hoc help-seeking from supervisors during patient consultations, and their assessments of the credibility of their supervisors and the value of the help provided.

Methods
The study used focus group discussions to collect data from general practice trainees. Focus groups were chosen to reduce the impact of any social distance between the facilitator and participants on the discussions, and gain more insight into trainee group identity, talk and dynamics (19). There is a spectrum of views on the value of focus group data. At one extreme are claims that participants throw little light on how things really are: this is both because there is considered to be no objective reality in social worlds which can be so illuminated, and because participant accounts are constructed for a particular audience in a particular context (20). At the other extreme are views that research participant accounts directly describe how things are, or at least how participants experience and make sense of the world. In this study we chart a middle path. We acknowledge that participants do not have direct insight into their own cognitive processes, and that they fashion their accounts for a particular audience, in view of their self-presentation and other agendas. However we take a constructivist realist position (21) that there are links between these accounts, and internal, external and social worlds which are useful to explore and understand.
Five focus group discussions (mean duration 43 minutes, range 35-57) were held in November 2017 with a total of 16 trainees (median group size 3, range [2][3][4][5] in Brisbane, Australia. The groups were held immediately prior to, or following, trainee education release days, at premises rented by the training organisation. Information about the study was provided by medical educators at previous education sessions, who distributed written participant information (including the study focus on trainee help-seeking) to trainees expressing interest in participation. All trainees who contacted the focus group facilitator (the corresponding author) by email to confirm interest took part in a focus group. Sampling was by convenience and no record of non-consenting trainees was kept. Several participants may have recognised the facilitator from her previous role as coordinator of undergraduate medical programme general practice placements, but none had a current relationship.
No-one other than the participants and facilitator was present. See Table 1 for participant demographic information. See Table 2 for the focus group guide, which was not pilot tested. The group discussions were experienced by the facilitator as engaged, generous and authentic. All focus group discussion was audio-recorded and transcribed by a professional service, and handwritten jottings were made concurrently by the facilitator. Transcripts were not returned to participants.
Analysis proceeded after all transcripts were read in full by all the authors. Initial coding was performed by the corresponding author, using in vivo and descriptive coding (22), and Atlas.ti software to assist with data management. Analysis proceeded using a constant comparative approach, to ensure grounding of analysis in the data as initial codes were clustered into higher level codes and developing concepts. For example, data coded as 'pushback', 'good teacher' and 'losing face' were clustered under 'supervisor style', which fell into the overarching category of 'help-seeking outcomes'. An inductive and interpretive approach was used to identify major and minor themes (23).
Analysis continued until no new findings emerged.
The corresponding author is a practising general practitioner undertaking a PhD in medical education, CJ is an academic with relevant expertise in patient safety and MP is an academic with expertise in medical ethics, law and professional practice. The authors met twice to discuss their analysis and interpretations of the data and reached a consensus on the overarching themes over multiple redraftings of this paper.

Results
Findings are discussed below under three major categories: help-seeking objectives; help-seeking activities; and help-seeking outcomes.

Help-seeking objectives
Participants consistently reported having sought help frequently during their initial transition from hospital practice into "the wild safari" of general practice, and this was contrasted with reduced helpseeking after the first "two to eight weeks".. Initially trainees needed assistance to obtain information about practice software and other facilities, and the "complex maze" of local patient resources and referral preferences. For many trainees, general practice was clinically unfamiliar, as was their responsibility for patient management decisions. Trainees reported that it was "comforting" to have confirmation of their plans in situations which they were encountering for the first time, such as calling an ambulance to transfer a patient to hospital, instead of "second guessing" themselves.
It was common to contrast this initial stage with subsequent help-seeking. Instead of "coming across as really needy" and "having no idea",, trainees described their later help-seeking as "getting a second opinion" and "double checking" when they were "not 100% sure".. In general, participants described their help-seeking in informal, non-hierarchical, lay language, using terms like "just having a chat", "having a quick look", "having a feel of her tummy" and "helping each other work out a plan".
Trainees also often described their help-seeking as "making sure" that they were "not missing anything".. They talked about the importance of being (and demonstrating that they had been) safe: P2: And you want to make sure that you are safe. And that you have shown that you have been safe by discussing with your supervisor about something that you-you are unsure about.
Q: Showing to the patient or to the supervisor or?
P2: I think to yourself. To the medical legal systems, to the patient and to your supervisor.
What everyone wants to know is that you are safe in your practice as a registrar and that, if you are unsure, then you will talk to someone FG3 (Focus group 3) Help-seeking stories included several scenarios in which decisions had to be made about whether or not to admit a patient to hospital. These clinical contexts included febrile children, adults with chest pain and patients with suicidal ideation. Participants often talked about having to "sleep at night" to convey their sense of the risk of managing these patients at home, relying only on patients and their families for monitoring. Other scenarios for help-seeking which were mentioned by several trainees included skin conditions, 'drug-seeking' patients, requests from patients for another opinion, and regular patients of another general practitioner in the practice with ongoing problems. Trainees also reported that it was sometimes valuable to take "time out" away from patients to collect their thoughts: Sometimes Participants talked about being "here to learn" and "soak in" as much knowledge as they could from supervisors who they perceived to be particularly clinically astute. They also commented appreciatively on senior general practitioner colleagues who discussed their own cases, with "noone's thinking they know it all". However there was some ambivalence about it being "OK that you don't know the answer" and several trainees mentioned particular general practitioners or "very, very smart" senior general practice trainees who might in fact know all the answers: You ask the person who you really trust who's just going to give you a quick opinion and who's always bloody right! FG5P4.

Help-seeking activities
Help-seeking typically involved an initial attempt to solve the problem independently, for example by reviewing patient records for previous management strategies or searching online. A more accessible source of advice than the supervisor, such as a practice nurse, was sometimes contacted. If these steps were unsuccessful, a decision was then made about whether help-seeking could be postponed to a more convenient time, in order to reduce the disruption to their own and their supervisors' consultations and enable more relaxed discussions. If "the answer" was required during the consultation, trainees attempted to locate a supervisor who did not need to be interrupted with their own patient, and who was likely to provide a timely response. This activity included studying the appointment schedules and patient waiting times of other general practitioners, and "hovering" or "loitering" in corridors and outside consulting room or treatment room doors. Several trainees reported that they distributed their help-seeking to reduce the load on any one general practitioner, and several described "cherry-picking" help providers based on perceived areas of provider expertise and their approach to managing risk, although a few participants reported having only one supervisor.
Some trainees also reported seeking further advice from other sources if they were uncomfortable with, or dubious about, the advice received. Participants seemed somewhat ambivalent about whether this was disrespectful or justified: Trainees typically phoned their supervisor from their consulting rooms, presenting their request over the phone within the patient's hearing. However they also reported a number of reasons for preferring to seek help outside the patient's hearing, including the following: not wanting information in the case presentation to upset, worry or offend the patient; concealing the extent of the trainee's uncertainty from the patient; and avoiding the patient overhearing supervisor advice which the trainee might choose not to follow. One participant described "jumping up" to leave the consulting room as she noticed the supervisor approaching, in order to speak to the supervisor before he came in. Trainees described supervisors wanting them to be concise and "direct upfront" when presenting patients, "showing that they had given it some thought",, and to propose a management plan if possible.
Several reported using medical terminology in these presentations, although one participant reported using lay terms within patients' hearing.
While much in-consultation advice was sought via brief phone conversations, trainees also reported that it was often useful to witness the supervisor interacting with their patient: You know, so he could, you know, look at his patient and sum the whole patient up and work that out, just with that clinical experience. You know, that was right and that was a really useful experience, to actually see that interaction FG4P2 One trainee even commented that "in an ideal world" they would always prefer the supervisor to Help-seeking outcomes Many trainees reported positive outcomes from help-seeking, including picking up 'learning points' and coming to terms with ongoing clinical uncertainty, as something that they had to "learn to live with" as general practitioners. Several trainees reported being particularly reassured and impressed when they witnessed their supervisor admitting and managing his or her own uncertainty in front of their patient. Managing trainee uncertainty was also assisted by "readjusting their perspective" to the general practice setting, and staging management over several consultations rather than trying to definitively solve "all a patient's problems at one sitting": Trainees also discussed a number of adverse or mixed outcomes from help-seeking during consultations, and several reported avoiding in-consultation assistance from supervisors with whom they had had a negative experience. Adverse outcomes from help-seeking included delay and disruption to trainee workflow, pushback from supervisors, etiquette breaches by supervisors, and uncomfortable advice. These outcomes are expanded upon below.

Delay and disruption to workflow
Waits of up to 20 minutes were reported while supervisors completed their own consultation before providing help. Several stories were told of the trainee moving forward with patient management decisions while waiting for the supervisor, risking the embarrassment of having to "wind back" plans and "backpedal" once the supervisor arrived. A "good teacher",, on the other hand, could provide help efficiently: He's in and out in three minutes but it's quality, not wanting the whole consult re-done but giving the impression of being unhurried and thoughtful, and he interacts with the patient and reveals his thinking. FG4P4

Pushback
Trainees did not report any explicit discouragement from their supervisors to seek help, except for a single mention of audible supervisor sighs on the phone (which seemed to surprise the other trainees). However participants in one group reported knowing other trainees in less accommodating practices who avoided asking questions and "cut corners" because of workflow pressures. Another trainee commented that he would have liked his supervisor to "regularly check in... to kind of say 'Is this working for you?'" rather than "assuming everything is all right"..

Some ambivalence was expressed in a number of discussions about supervisors asking trainees to
propose a plan and "trust your judgement".. On the one hand, trainees were sometimes dissatisfied with "just hav(ing) to call it",, in a few cases framing this response as "pushback":: Several trainees reported advice to prescribe antibiotics which conflicted with their understandings of antibiotic stewardship (although one trainee reported accepting advice to prescribe antibiotics for a patient he had intended to manage conservatively 'because it wasn't such a bad argument actually').). Several trainees reported always following their supervisor's advice, although several others reported having disregarded this on occasions, sometimes after seeking another opinion. This was justified in terms of the trainee having "to sleep at night",, and in terms of the uneven nature of supervisor expertise so that in some areas it could be taken "with a grain of salt":: I think it's probably the GPs that have been GPs for many, many years, with patients that expect certain things, it's hard for them to perhaps start to change their practice in a way that's more in line with antibiotics stewardship... the necessity for antibiotics sometimes, you'll take that really with a grain of salt, and see whether you're reasonable in not prescribing and having a good return plan

FG3P2
Most trainees who reported having disregarded their supervisor's advice did appear to find this situation awkward, however, and several reported concealing this from their supervisor. One trainee described feeling obliged to follow a supervisor's practice, and to "be a certain type of GP" at that training practice, due to the supervisor's position as his employer and assessor.
A number of stories were also told of advice which was less than fully satisfying, where it seemed to address only one aspect of a more complex problem. Contexts included restricting the prescription of opiate analgesia to an opiate dependent patient, and referring a patient for assessment of their cognitive competence in a situation which appeared to involve financial abuse by a relative of an elderly patient.

Discussion
This study reports on the objectives, activities and outcomes of general practice trainee help-seeking, particularly clinical oversight, support and advice during patient consultations, and builds on previous work in Dutch general practice (5). New findings include trainee assessments of the credibility of their supervisors and the value of the help provided. There are tensions in any community of practice between newcomers and old-timers, and between the reproduction and the transformation of practice (24) as the new displaces the old. However there has previously been little exploration of the particular ways in which these tensions play out in general practice training.
There are some limitations of our study. There was no field work component to complement the selfreport data, participants did not provide feedback on study findings, we did not collect and analyse data iteratively, and we included a fairly small sample of Australian trainees. These limitations may restrict the transferability of our findings. Strengths include the inclusion of diverse trainees across different stages of training, the use of focus groups, and our attention to deviant cases (20) and complexity and inconsistency (22) in the data.
Although some trainees appeared more comfortable initiating oversight, support and advice than others, and some supervisors appeared to be more amenable than others to trainee help-seeking, and more skilled in responding, overall the authors gained the impression that help-seeking experiences were predominantly positive for most trainees. Some of the adverse outcomes reported appeared to be particularly "dramatic" and atypical episodes (25). The manner and content of supervisor responses to help-seeking requests had a strong impact on trainee assessments of supervisor credibility and their subsequent help-seeking decisions, including whether, and from whom, help was sought in future.
Trainees appeared to assess supervisor credibility in two key areas. The first was their approach to managing clinical risk, and trainees reported seeking help from supervisors whose approaches aligned most comfortably with their own. How trainees position themselves on this spectrum from cavalier to over-cautious may be an important aspect of their developing identity and resilience as general practitioners (26). The second area was their clinical expertise, and many trainees appeared to have the impression that this was patchy, and best approached by 'cherry-picking' the supervisor who seemed most knowledgeable in a particular clinical context, or on occasions seeking several 'second opinions'.
The ability to identify credible sources of advice is an important skill in patient safety (27), and developing reliable individual and local practice 'mindlines' (11,28). The development of this skill seems to have received little attention in the literature, beyond the emphasis on evidence-based practice, although it is believed to be important for learners to be able to observe and recognise expert practice (29,30). Given the limited opportunities for general practice trainees to observe supervisor expertise, ideally help provision would be one such opportunity. However given the time constraints upon supervisors, and the limited information available to them from trainees' brief presentations, supervisors may tend to fall back on simpler coping routines (such as reassuring trainees and patients when trainee plans appear to be essentially safe) rather than more expert proficiency routines (31) The authors were struck by the powerful influence on trainee help-seeking decisions of time constraints and the workflow requirements of moving patients in and out of consultation rooms at frequent intervals. The impediments to seeking help experienced by trainees also contrast somewhat with previously reported supervisor perceptions of being approachable and accessible (18). The trainee's patient also appeared to exert considerable influence on trainee help-seeking decisions, especially if they were a regular patient of another general practitioner in the training practice.
Trainees often preferred to present patients outside their hearing (but have supervisors join their consultations to provide advice, time constraints permitting). There was also some tension for trainees in whether or not to use medical talk in their case presentations in front of patients. These tensions suggest that general practice trainees attend to how they present themselves to both patients and supervisors. Trainees often used informal, non-hierarchical talk about their help-seeking, including the "second opinion" terminology which is recommended to Australian general practitioner supervisors (6,33). However the social distance and power differential between supervisors and trainees remained apparent in discussions about how to manage advice which made trainees uncomfortable, discussions about getting "back-up" from "the boss",, and comments that supervisors had a role in assessing, and often employing, trainees.
The risks of 'loss of face' in front of patients during training, particularly when seeking help or feedback, have been described as costs to ego and self-presentation (8), harking back to Goffman's original writing about the performance of everyday life, and the discombobulation resulting from disruption to social performances (25). The exposure of a character for not being what he claims to be is a powerful disruption, and trainees who implicitly claim to be full general practitioners are particularly vulnerable to exposure as lacking competence by supervisors with a poor understanding of the etiquette which protects trainees from such exposure. Many trainees, however, appeared to have become comfortable with seeking help in front of patients, at least from supervisors who they trusted to respond appropriately.
In talking about their help-seeking, trainees positioned themselves as having appropriate insight into the risks of unrecognised incompetence (34), rather than being under-confident, risk averse or "needy".. Trainees were somewhat ambivalent about the advice of supervisors to "trust their judgement", although they did appear to appreciate validation from their supervisors and preferred to "trust (their) gut" rather than receive unwanted advice. Accepting uncertainty and risk appeared to be understood by trainees as part of becoming a general practitioner. However trainees remained curious about how other general practitioners practised, particularly in the "grey areas" of general practice. It is important for trainee development that supervisors model and discuss approaches to clinical and professional practice dilemmas, providing opportunities for what has been termed 'long loop' learning (5) throughout training.

Conclusion
These findings have a number of implications for general practice supervision and supervisor training both in Australia and internationally, which may extend to other specialty training contexts. Trainees make judgements about the credibility of supervisors in terms of their approach to risk and their clinical expertise. The latter was often believed to be uneven. How supervisors respond when their help is sought has a strong influence on a trainee's subsequent help-seeking, and whether they follow the advice. Trainees prefer to seek help from supervisors who respond promptly and observe the etiquette of providing help in front of patients, and whose advice they trust. Trainees may seek further opinions if they are uncomfortable. They learn through help-seeking to manage uncertainty and make their own clinical calls but remain curious about how other general practitioners practise, and value opportunities to witness their supervisors' clinical expertise especially in managing clinical and professional dilemmas, and the 'grey areas' of practice.
Abbreviations GP general practitioner, FG focus group, P participant Declarations Full ethics approval was obtained for the study from the University of Queensland (2017000867). All participants provided written consent to participate and for dissemination, including publication, of de-identified study findings. Consent was not sought to make full transcripts available outside the investigator team.