Care of the dying: a qualitative exploration of Foundation Year doctors’ experiences

Foundation Year doctors (FYs), who are newly qualified, are expected to provide care for dying patients. Experiences at this early mandatory stage of training may form the foundation for future encounters, but little is documented about what these experiences involve. The aim of this research was to explore the experiences of FYs in caring for the dying, using the recently published Priorities for Care of the Dying Person as a conceptual framework, to identify areas for improvement in education and clinical practice.

METHODS Semi-structured group and individual interviews were conducted to explore the experiences of FYs and how these relate to the five aspects of Priorities for Care of the Dying Person: 'recognise', 'communicate', 'involve', 'support' and 'plan and do'. All FYs in the North Yorkshire and East Coast Foundation School (n=335) were invited to participate and 47 FYs were recruited from five sites through convenience sampling and snowballing. Recordings were transcribed verbatim and a framework analysis approach was used with the published Priorities for Care of the Dying Person guidelines as a conceptual framework.
RESULTS Five main themes and 13 subthemes emerged from the data. The five main themes, which mapped to the conceptual framework, were: recognition that the patient is dying; communication with the patient, family and other staff; involvement of the patient and family in their care; support for the dying person and their family; and planning and carrying out good care of the dying. Examples of when things are done poorly or done well were shared, giving context to experience. INTRODUCTION Care of the dying patient is an integral responsibility of Foundation Year doctors (FYs). In their first year as a doctor, FYs will look after around 40 patients who pass away, and approximately 120 patients nearing the end of their life. 1 The General Medical Council (GMC) recognises care of the dying as an integral part of medical practice, and their guidance is set out in their 2010 publication, Treatment and care towards the end of life: good practice in decision making. 2 Recently, the Leadership Alliance for the Care of Dying People (LACDP) published a pivotal document, Priorities for Care of the Dying Person, 3 which outlines five key priority areas to be addressed by the health care team. These five areas are 'recognise', 'communicate', 'involve', 'support' and 'plan and do.' 3 The LACDP contains 21 national organisations, including many major stakeholders such as the Department of Health and Health Education England. 4 A national care of the dying audit occurs regularly 5 but the experience or performance of FYs is not addressed.
The Foundation Programme is a 2-year broad training programme for newly qualified medical graduates working in the UK. Doctors on this programme are known as Foundation Year 1 (FY1) or Foundation Year 2 (FY2) doctors, and their duties consist of both service provision and clinical training. In 2015, 15 355 of 59 250 doctors in training posts in the UK were FYs. 6 The Foundation Programme curriculum includes competencies relating to care of the dying. 7 Following on from this, the majority of training and career pathways for doctors will involve some aspects of care of the dying. Therefore, this common stage of training is a key opportunity to instil principles of good practice of care of the dying. Although there is much literature regarding these issues for medical student education and broader stages of training, 5,[8][9][10][11][12] it is important to understand the experiences of FYs specifically.
Studies focusing specifically on FYs' experiences of caring for the dying are limited. [13][14][15] One Scottish study showed that FYs felt inadequately prepared to care for a dying patient, a thought echoed by senior medical staff. 13 Another study concluded that 'undergraduate medical education is currently failing to prepare junior doctors for their role in caring for dying patients'. 14 Similarly, Linklater identified that there are areas within caring for the dying that FY1s found difficult, such as communication, lack of support from senior medical staff and providing inappropriate medical treatment that is futile. 15 There may be emotional distress and responsibility felt by all staff in these situations, alongside uncertainty and ethical dilemmas. 2 A pathway intended to aid in care of the dying, which was in widespread use across England, was recently withdrawn and a review undertaken 16 following alarming media reports and criticism of poor care. A 2016 Cochrane review found limited evidence of the effectiveness of pathways to assist with providing this care. 17 Care of the dying is emotive, important and closely scrutinised, and has potential to be a particularly challenging area for FYs.
In view of the significance of care of the dying and the literature suggesting that doctors are underprepared for this role, [13][14][15] the aim of this research study was to explore the experiences of FYs in care of the dying. The Priorities for Care of the Dying Person 3 was used as a conceptual framework. This framework allowed the exploration of experiences against a reputable and established national source, to provide a unique insight into the experiences of FYs, which may then help identify areas for improvement in training and clinical practice. The research questions for this study were: (i) What are the experiences of FYs? (ii) How do these relate to the five Priorities for Care of the Dying Person 3 ? (iii) How can any findings inform under-and postgraduate medical education curricula?

Ethics
Ethical approval was granted by the Hull York Medical School (HYMS) Ethics Committee. This included the method of invitation, obtaining consent, data use, storage and dissemination. As the project only involved NHS staff, NHS Research Ethics Committee review was not required, 18 but institutional approvals were gained prior to recruitment.

Design
Semi-structured group interviews were used to enable the participants and researchers to discuss complicated subjects and explore FYs' experiences, perceptions and reasoning. 19,20 Because of the practical necessity of scheduling around clinical commitments, some participants elected to contribute via individual interviews instead. Individual interviews allow for a more in-depth exploration of issues 19 and can be used alongside group interviews. 21 Therefore, a pragmatic mixture of group and individual interviews was used to maximise the opportunities for FYs to contribute. Interviews were peer facilitated and the strengths and weaknesses of this are discussed in the limitations section.
As the Priorities for Care of the Dying Person 3 was used as a conceptual framework to generate the topic guide, framework analysis was used. An inductive approach was exercised whereby the five priorities 3 were pre-selected as key themes. 22 A topic guide for the group and individual interviews was developed using the current literature and the document Priorities for Care of the Dying Person. 3 It was then tested in a pilot group interview before use, which allowed the refinement of stem questions. The topic guide consisted of open questions regarding FYs' experiences of care of the dying to stimulate discussion, followed by stems based upon Priorities for Care of the Dying Person. 3 These stems and prompts (see Table 1) were used for all interviews to reduce interviewer bias and enable transparency, whilst still allowing for an inductive response based on participants' narrative and feedback.

Study setting
FYs from across the North Yorkshire and East Coast Foundation School, which covers five sites (York, Hull, Scunthorpe, Grimsby and Scarborough), were invited to participate via convenience sampling through e-mail, social media and snowballing. No participant relationship was sought prior to study interviews.
Participation was voluntary. Consent, information and short demographic data forms were completed. Participants were offered light refreshments and a certificate of participation.

Data collection
Eight group interviews and 21 individual interviews were conducted between January and March 2016 and were all audio-recorded. The interviews were largely conducted by one author (JP) and MR, SG and GF each conducted at least one group interview. Group interviews contained between two and five participants and were facilitated by a single researcher (JP, MR, SG or GF) who introduced her role at the beginning of the session; no observers or other individuals were present. HYMS facilities were used at each of the five listed hospital sites and sessions were conducted outside of participants' working hours. Individual interviews were held over the phone or using videoconferencing facilities. No repeat interviews were conducted. The researchers were . From a reflexive standpoint, it is worth acknowledging steps taken to minimise researcher bias, namely the use of standardised stems and prompts (see Table 1), an agreed framework, and data analysis steps as described below. Transcripts were not returned to participants for feedback. Participant validation was not used as different events may have occurred between the time of interview and transcription and because of the excessive burden of time this would impose on busy participants. 23 No participants withdrew from the study.

Data analysis
The procedure for data analysis included verbatim transcription, familiarisation with the interview through reading the transcript, open coding, developing the framework, charting and interpreting the data. 22 Transcripts were coded using an agreed framework, with comparison across the transcripts. The results were coded and data were managed using NVivo qualitative data analysis software (Version 11). 24 A combined approach was used; although the five Priorities for Care of the Dying Person 3 were used as a conceptual framework and as pre-selected themes, coding was inductive and sub-themes were found under each priority. 22 Care was taken to include contradictory views during the coding process. Participants were not involved in checking the findings. All authors contributed to interpretation of the findings, including MJ, who is a senior clinical academic in palliative care, involved in under-and postgraduate palliative care education in HYMS. In this paper, the results section will be restricted to presenting the data, then the discussion will be used to present the analysis of the findings in the light of other relevant literature.

RESULTS
A total of 47 (14%) FYs participated out of 335 (see Table 2 for demographic data) (mean age 25.7 years, range 23-38). There was an even split between FY1s and FY2s. Participants gained their medical degree from 17 medical schools, of which two were outside the UK.
Thematic saturation was reached. Thirteen main sub-themes arose from the five overarching themes. These are shown in Table 3.
Unsurprisingly, given the topic guide used the Priorities of Care of the Dying Person 3 as a frame, the five main themes directly mapped onto each priority. However, the 13 sub-themes were not confined by those topics and were included to minimise reporting bias. An inductive process of exploring participants' discussions allowed for further sub-themes to be uncovered. In this paper results are therefore presented using the conceptual frame provided by Priorities for Care of the Dying Person. 3 Each of the five priorities is defined and then explored in turn with themes and illustrative quotes from participants. Quotes from participants in group interviews are labelled with their 'G' (group) and 'P' (participant) numbers. Quotes from individual interviews are identified by their 'I' (interview) number. No participant demographics are linked to these in order to preserve anonymity. Recognition that the patient is dying 'The possibility that a person may die within the next few days or hours is recognised. . .. 3 ' Participants generally found it difficult to recognise a patient was dying, but found that this became easier with experience: . . . it's difficult to recognise kind of the signs, such as things like agonal breathing, which we get told about in medical school but haven't seen yet. G7, P3 . . . it was challenging initially . . . recognising that they are deteriorating . . . at the moment I think with experience I am getting a bit better, I wouldn't say I'm the best yet but getting better. I16 Participants also discussed differences between specialties when recognising a patient was dying; surgical specialties were highlighted as an area where this was particularly poor: I think in my surgical job patients that did end up on palliative care pathways it tended to be much closer to the point of death than on medical [jobs] . . . G6, P4 There was also some exploration about whose role it is to recognise when a patient is dying. Participants questioned if it was the role of the FY or the responsibility of a more senior doctor or a more experienced nurse: .  This also brought pressure, because these interactions were seen as critical: . . . it's quite intimidating because . . . they always remember the discussion with the doctor who told them that their loved one is going to pass away [die] and they are really critical kind of doctor-patient/doctor-relative moments that you really don't want to mess up and it's like really really scary. . . G7, P3 However, when communication was sensitive and effective, participants found it could be a valuable part of their role: . . . it's quite rewarding though I think . . . when you pick up on that and you do have these good like communication discussions with the family and they, even though it's horrible to give that sort of information they were, felt so much better informed about what was going on . . . I2 Just as the role of FYs was sometimes unclear regarding recognition of dying, there was also uncertainty about their role in addressing patients' and their families' expectations, which could be practically challenging but potentially rewarding.

Communication with the team
Written and verbal communication within the multidisciplinary team were both reported as crucial. Multiple examples were given, but good communication at handover and documentation in the clinical record were particularly highlighted.

Handover
At the beginning and end of some shifts, doctors have the opportunity to give their colleagues written and verbal information about patients that the following doctor needs to be aware of. It was noted as particularly important to include information about deteriorating patients and expected deaths in that handover process: . . . even at handovers . . . with the stroke ward . . . Even if people weren't necessarily for end of life . . . they were quite good at making strict escalation plans. . . they were quite good at setting out "this is the limits of what this person should be and it was all discussed with them", with the patients, but it was quite a good, and so it meant handovers and things over the weekend that they, it was also quite clear. G3, P1 Doctors may start a shift and suddenly find themselves responsible for the care of a dying patient or communication with his or her family. Therefore, communication between team members is highly important: . . . it's really important that it's that the on call doctor knows who the patients are that are expected to die because sometimes you get asked to see relatives of the patients that you don't know . . . there wasn't a specific area on our handover for expected deaths . . . there is something in there as well about handing over expected deaths, plans, and a little bit of background information for those patients to the on-call team. G7, P3 Effective handover of patients who are expected to die may better prepare FYs when they begin a shift.

Documentation and terminology
Participants described the importance of accurate and specific documentation of all aspects of patient care, ensuring access to relevant information from the patient's notes: . . . so a lot of it is not just communication with the whole team but putting that communication into words in the documentation so it can be passed between different teams that don't have chance to communicate verbally. G7, P1 However, participants described a lack of clarity about the manner and content of documentation: . Sometimes the patient's lack of capacity led to the teams relying more heavily on input from families: . . . I feel like we don't really get the wishes from the patient themselves because most of our patients are quite elderly and cognitively impaired so . . . they can't always communicate what they want, so most of the time we speak to the family I think, to get that from the families and from their friends and if they are close to them and that's when we find out, okay, you know, is this what this person would want or what would they prefer or, you know, would they prefer to die in a hospital, would they prefer to die at their own home or in a care home or? I9 As families may be expected to be more involved in those situations, there was an emphasis on the delicacy of such conversations. There was a dilemma about the involvement of families in decision making, as this was sometimes an uncomfortable role for the FY: . . . some people turn it around on to the relatives and say . . . "What would you want them to do", it's, it's really not that fair because then they do kind of feel like "well I want them to stop treatment, maybe that's why they died and it's my fault, maybe I shouldn't have said that". . . ask them . . . but then you kind of need to make, make them feel that you're the one in control and you're just involving them, you're not making them make the decision. I19 Involving patients is positive where capacity allows. Advanced planning could be considered. Involving families is beneficial but requires careful consideration in order to not make them feel overburdened.

Support for the dying person and their family
'The needs of families and others identified as important to the dying person are actively explored, respected and met as far as possible.' 3 The discussion around support included comments that overlapped with issues raised about communication. Several FYs felt support was largely provided by nurses and palliative care teams and time can be a challenging factor: . . . So in a way you never know whether people want to be left alone or whether they want to talk through things or not, the nurses would always like to give that option but you don't always have the time to do that. . .

I19
It was noted that it may be difficult to know how to best support family members, partly because of the variety in individuals' reactions: . . . I think it's more dealing with people who've got different kind of reactions and grief responses, as I say grief reactions and things within a family and trying to . . . explain things appropriately . . . but normally you're having to do it all in the same conversation, it's quite difficult to get the time to be able to have all those conversations separately with all those different people . . .

G3, P1
Different cultural, linguistic and faith backgrounds also challenged FYs' ability to provide appropriate support: . . . . once you're aware that a patient is dying, it's been agreed, I think that the management is, is very good in the hospitals . . . I18 Many participants noted the usefulness of the palliative care team when coordinating and delivering plans around end of life: . . . if you have easy access to the palliative care team that makes things a lot easier . . . G4, P1 However, it was also clear that sometimes, because of the speed of deterioration or delayed recognition of dying, it can be difficult to allow for proper planning: . . . when they deteriorate in surgery there's less of that kind of being able to plan for things and it's more reactive rather than proactive, which is what I found with medicine. G5, P2 There were also comments that the role of the FYs is more to deliver the plan, rather than make the decisions: My job is more to carry out the management plan, I was not involved in decision making, I was involved in prescription of the end of life drugs and actually writing those things out. . . G4, P2 FYs commented that up-to-date prescribing guidelines or appropriate protocols are not always accessible, thus indicating organisational issues that need to be addressed: . . . it does vary from trust to trust but at the time the Trust guidelines weren't up on to our system so I had to go on to Google. G1, P2 Therefore, palliative care team support and timely recognition of dying were important for allowing proper planning, but barriers such as sudden deteriorations, inaccessible resources and the limitations of the role or knowledge of an FY may affect the FY's experience. Up-to-date guidelines for prescribing should be made available.

DISCUSSION
In this review of FYs' perspectives on caring for the dying, experiences were found to map to the five Priorities for Care of the Dying Person 3 expected of health care staff. However, within that experience, consistent with findings from the End of Life Care Audit (EOLCA), 5 both good and bad examples of care and some areas for improvement were identified. Good experiences were characterised by an integrated team with excellent communication, documentation and handover. Clear plans for a ceiling of care and documented discussions with patients and families, with proactive recognition of dying, particularly out of hours and at weekends, prevented persistence with futile and often invasive treatments, and allowed appropriate care of the dying. Where care was poor, this was linked with late recognition of dying or persistence with futile treatment, poor communication within the team or with the family, lack of understanding of and engagement with the patient's or family's wishes, waiting too long to seek the patient's view (until they no longer had capacity), and inadequate support and management. Where the five priorities 3 were not met, it is unsurprising that this reflected negatively in the FY's experience. It was clear that not all FYs feel comfortable recognising patients who are dying or communicating with them, as previously described, 13 and that there are associated learning needs. 14,15 Despite recommendations that a senior clinician is involved in recognising dying, the EOLCA identified that for a quarter of patients (24%) this did not happen. 5 The data highlight the importance of senior support for FYs in this regard, which is echoed elsewhere for juniors further along in their training. 12 Where senior clinicians remain actively involved in the care of the dying, better care is provided by the whole team, who are supported in this. Where seniors 'abdicate', then decision making and care is left to the most inexperienced and poorly supported member of the team. 8 Medical students and postgraduate trainees need systematic and experiential training about care of the dying as part of this support. Models exist but are variably implemented. 9,10,25 The palliative care team was identified as an important source of support, which is reinforced elsewhere. 12,14 Clinical, practical and emotional support for all doctors caring for dying patients has been identified by the British Medical Association as important. 11 Therefore it is crucial that this support is also provided for the most junior, least experienced doctors, the FYs.
Poor experiences reported by FYs were also linked to delayed recognition of dying. Delayed recognition of dying is common. One centre's audit found that for 87% of dying patients, recognition occurred less than 72 hours prior to death, 26 and EOLCA found that for half of all cases, recognition occurred less than 34 hours prior to death. 5 Qualitative research has also identified a culture barrier to accepting death as an acceptable outcome. 27 It is important that recognition occurs as early as possible, to deliver good care 3

Limitations
The three main authors (SG, JP and MR) of the study were all FY2s at the time of research. The experiences of these three FY2s may have subconsciously influenced the facilitation of the interviews. However, their role may also have reduced the power differential between facilitator and participant, and there is less chance of peers misinterpreting the discussion. 20 There were advantages to being medically trained in terms of familiarity with the language and concepts discussed, but by contrast, some participants may have felt constrained by this peer-facilitated discussion 20 and there may be limitations around perceived lack of neutrality or doubts regarding confidentiality.
The research was conducted as a combination of both group and individual interviews. 21 Fewer group interviews and more individual interviews were conducted than initially expected, largely because of availability of participants, who were busy FYs with different working patterns. There was a relatively large sample size and similar themes were found across both group and individual interviews, both of which help negate any impact of mixing the methods. With interviews, there is no participant interaction and this may have had implications for themes that were discussed, but does allow for more in-depth discussions of topics. 29 One limitation was the time constraints of the study period and the length of interviews, as there was a rich depth to both and further questions could have been asked in order to explore the topics.
The research was conducted at one Foundation School (NYEC), but graduates from 17 medical schools participated. There was a variety of teams and hospitals and duration of experience (6-18 months), which will have affected FYs' experiences.
Although recall bias may have affected the accuracy of the events described, the experiences of the FYs remain valid. In the context of the aims of the study, the perceptions of the FYs are more important than a factual account.
Inevitably, research volunteers may be those with strong views and experiences in the area, either positive or negative. However, the broad range of 47 participants (see Table 2) would indicate wide representation of FYs at NYEC.

Clinical and training implications
To improve the five aspects of 'recognise', 'communicate', 'involve', 'support' and 'plan and do', 3 recommended areas for improvement obtained from exploring the experiences of FYs are listed in Table 4.

CONCLUSIONS
Using the Priorities for Care of the Dying Person 3 as a framework, these data highlight both good practice and areas for improvement. Experiences of care of the dying varied in quality depending on several factors. Good experiences tended to be in circumstances where the five priorities 3 were being addressed by the whole clinical team. Gaps where the five priorities 3 were not met can be addressed through training, support and alignment with the five priorities, 3 to then improve the experience of care for patients, families and the FYs caring for them, and to instil principles of good care of the dying in these doctors to take with them throughout their careers.
Contributors: MR, JP, SG, MJ and GF all contributed to the design of the work by meeting together at the start of the project and planning the way forward. The conception of the work was largely due to MR, JP and SG, but honed and focused by MJ and GF. The acquisition of the data was mainly by JP but also heavily involved MR, and SG and partly involved GF. The analysis and interpretation of the data was mainly by MR but also involved JP, SG, MJ and GF. The drafting was performed by MR and was critically revised by MJ, but also to a lesser extent by JP, SG and GF. MR, JP, SG, MJ and GF all gave final approval of the version submitted.