Implementing advance care planning in heart failure: a qualitative study of primary healthcare professionals

Background Advance care planning (ACP) can improve the quality of life of patients suffering from heart failure (HF). However, primary care healthcare professionals (HCPs) find ACP difficult to engage with and patient care remains suboptimal. Aim To explore the views of primary care HCPs on how to improve their engagement with ACP in HF. Design and setting A qualitative interview study with GPs and primary care nurses in England. Method Semi-structured interviews were conducted with a purposive sample of 24 primary care HCPs. Data were analysed using reflexive thematic analysis. Results Three main themes were constructed from the data: ACP as integral to holistic care in HF; potentially limiting factors to the doctor–patient relationship; and approaches to improve professional performance. Many HCPs saw the benefits of ACP as synonymous with providing holistic care and improving patients’ quality of life. However, some feared that initiating ACP could irrevocably damage their doctor–patient relationship. Their own fear of death and dying, a lack of disease-specific communication skills, and uncertainty about the right timing were significant barriers to ACP. To optimise their engagement with ACP in HF, HCPs recommended better clinician–patient dialogue through question prompts, enhanced shared decision-making approaches, synchronising ACP across medical specialties, and disease-specific training. Conclusion GPs and primary care nurses are vital to deliver ACP for patients suffering from HF. HCPs highlighted important areas to improve their practice and the urgent need for investigations into better clinician–patient engagement with ACP.


INTRODUCTION
Heart failure (HF) is a major cause of mortality and morbidity worldwide with an increasing proportion of patients suffering from refractory HF requiring palliative care. 1 2 Improving quality of life for these patients is a fundamental goal of HF management in national and international guidelines. [3][4][5] Patients with HF often suffer from a high symptom burden, an unpredictable disease trajectory and severe prognosis. 2 3 Studies show that advance care planning (ACP) can improve the quality of life of these patients [6][7][8][9] and is defined as a voluntary process that helps patients in sharing their personal values and goals of future care in order to safeguard their care preferences in case they become seriously ill. 10 As such, ACP is significant in identifying early palliative care needs and preparing for the end-of-life. 11 While ACP is widely advocated in HF, merely 7% of HF decedents compared with 50% of cancer patients had their palliative care needs recognized, 12 and only 8-10% of patients had the opportunity to participate in ACP mainly because healthcare professionals (HCPs) found ACP difficult. 13 14 As a result, the majority of HF patients receive suboptimal palliative care. 3 15 It is widely accepted that General Practitioners (GPs) and primary care nurses are central to engage with ACP because of their pivotal role in the provision of continuous supportive and palliative care in the community. 16 17 Moreover, patients and carers think that GPs and nurses should have a prominent role in advance care planning. 18 However, what approaches and A c c e p t e d M a n u s c r i p t -B J G P -B J G P . 2 0 2 0 . 0 9 7 3 interventions GPs and primary care nurses perceive as helpful to improve their engagement with ACP in HF has not been investigated to date.
Previous research into implementing ACP in HF has primarily focused on interventions targeting patients to engage with ACP [19][20][21][22][23][24][25] but not on primary HCPs. The few studies which have examined primary care professionals' experiences with ACP in HF demonstrate concerns about the timing, initiation, conduct and recording of ACP conversations, worries about competency of disease specific communication skills and the lack of resources. [26][27][28] To increase HCPs' engagement with ACP in HF, we need to identify and develop effective ways that improve clinical practice. 29 30 This paper explores and reports on the views and ideas of GPs and community nurses on how to improve their engagement with ACP when working with patients suffering from HF in England.
The specific objectives of this aspect of the study reported were:  To explore how GPs and community nurses working with patients suffering from HF understand ACP and their role within ACP.
 To identify factors that may facilitate or impair primary HCPs' engagement with ACP in palliative care for HF.

Design
We undertook an interpretive and descriptive study using semi-structured interviews with primary HCPs in the South of England including rural and urban communities. Semistructured interviews are seen as a suitable approach to generate personal in-depth findings based on HCPs' experience in sensitive topics such as end-of-life care. 31 We explored the A c c e p t e d M a n u s c r i p t -B J G P -B J G P . 2 0 2 0 . 0 9 7 3 experiences of GPs and primary care nurses on how to improve their engagement with ACP in HF. The interview series involved HCPs who delivered clinical services and ACP consisting of full and part-time GPs, district, practice, end-of-life care nurses as well as HF specialist nurses. We focused on primary care staff because evidence suggests that primary HCPs are considered central professionals in the management of continuous and end-of-life care. 32 The study was approved by the first author's institutional ethics committee (R44601/RE001). Written informed consent was given by all study participants. Research participants and their corresponding data were pseudonymized.

Recruitment
Likely participants were identified from publicly accessible information on GP practice websites in the South of England, the local Clinical Research Network, by word of mouth and through professional contacts. Purposive sampling aimed for a maximum variation in participants 33 , seeking a range of HCPs (e.g. full and part-time GPs, salaried, academic and Out-of-Hours (OoHs) GPs, community HF specialist nurses, practice and district nurses, a variety of settings (urban and country practices), a degree of experience ( e.g. HCPs with little or no experience in ACP or looking after patients suffering from HF); and age, gender and ethnicity. We approached 29 potential participants with a letter of invitation, the

Data Collection
Semi-structured, in-depth interviews were conducted between April and September 2016 by one researcher, who is a GP and clinical academic with experience in qualitative research.
One-off audio-recorded interviews were undertaken at a place of participants' choice, which was usually their surgery, lasting between 35 and 60 minutes. An interview topic guide (SF. 3) explored participants' experience with ACP in HF, perceived barriers to engaging with ACP, their management of ACP in recent cases, and suggestions on what might help them to improve their engagement with ACP. The interview guide was piloted and refined during the first three interviews of the study. Initially, multiple questions about their experience with ACP, and barriers and facilitators to its implementation were developed. Pilot testing was performed by qualitatively testing the questions using a form of cognitive interviewing. The purpose of cognitive interviewing and testing the topic guide was to investigate how well questions perform when asked of participants. This allowed us to select the optimal question for each topic and refining the wording to produce a field-test version. 35 The pilot topic guide required only minor adaptations and qualitative data during the pilot interviews was included in the final data analysis.

Data Collection and Coding
All interviews were conducted face-to-face at the participants' place of work. No new themes were identified after 21 interviews with a high rate of recurrence of topics and no new development of codes or themes. However, three further interviews were undertaken to ensure data saturation. All interviews were audio recorded, and then independently transcribed verbatim by an external professional transcription service soon after recording.
All transcripts were checked for accuracy by the interviewer. A sample of transcripts was A c c e p t e d M a n u s c r i p t -B J G P -B J G P . 2 0 2 0 . 0 9 7 3 independently checked for accuracy and coded by a sociologist and non-clinician (JM) with significant expertise in thematic analysis to enhance the credibility of data collection and analysis.

Data Analysis
Interview data were interpreted inductively using Braun and Clarke's six phases of thematic analysis: 36 data familiarisation; generating initial codes; constructing themes; reviewing potential themes; defining and naming themes; and producing the report. Transcripts were entered into NVivo (version 11), coded and thematically analysed. A clinical academic (MS) led the analysis. He kept a reflexive diary, reflecting on the influence of his own professional background on data analysis including deviant case analysis.
To increase rigour, three members of the research team with backgrounds in sociology, nursing and extensive experience of end-of-life research (JM, AB, BW) had ongoing involvement and input in the data analysis. 37 38 JM and BW compared and reflected on the early and subsequent coding decisions together with the interviewer (MS). A third researcher (AB) reviewed findings and provided new insights throughout the coding process. We regularly attempted to identify deviant cases by actively seeking out those interview responses that did not conform to the views of the majority of interviewees. These iterative steps informed the interviewer's interpretative analysis.

RESULTS
Twenty-five participants out of twenty-nine invitees replied to express an interest in taking part. A total of 24 HCPs participated in the interviews. Data saturation was ensured after 24 interviews making the participation of the remaining participant redundant. The majority of interviewees were GPs, females and had a white British origin ( Table 1). They displayed a varying level of experience in general practice (Figure 1). Nearly one third of interviewees were nurses including practice nurses, HF specialist and district nurses. ACP as integral to holistic care in HF •

Providing a broader scope of health
Many HCPs saw end-of-life care in HF as synonymous with providing holistic care which, in itself, represented a strong benefit and motivation to do their job. One GP said: "I think for GPs on the whole, the thing that floats our boat, is the idea of holistic care. It is the idea of not just firefighting the moment, but actually thinking about a broader definition of health." (R21)  A defining reason to go to work Being able to provide holistic care was not only a strong but a defining reason for some nurses to go to work. One district nurse explained:    In order for training to be effective and sustainable, it needed to fit into existing educational programmes of GPs and nurses. Ideally the training should take place at their local surgery.
In this way, learning activities would also reach those professionals, who normally would not attend a palliative care event.

 Shared decision-making tools
When asking a HF nurse, what would help her most in deciding whether to start ACP for a patient or not, she replied: "I very much believe in using templates before consultations in specific areas. But the templates don't always give you the wording to communicate well with patients. And so, that's perhaps something that could be developed. … When you are discussing a patient, a template that covers each individual patient's palliative care needs which can be completed during the meeting -that would be a useful tool." (R3) Another GP suggested: "I think that shared decision making tools would be excellent. Not least of all to facilitate the conversation, because it is so much easier. So that's one of the pluses of a form.
You could give it to the patient in advance and they could write quite a bit on it. And that gives them the opportunity to put as much or as little and tick boxes about what they did or didn't want to discuss. That really helps the clinician to know where the patient is at." (R21)

 Synchronised communication across medical specialities
A c c e p t e d M a n u s c r i p t -B J G P -B J G P . 2 0 2 0 . 0 9 7 3 Some GPs suggested that the signal for the right timing to have an ACP conversation should come from the cardiologist or the HF specialist nurse since they were seen as the subject experts. One GP (R8) said: "I would find it incredibly helpful to receive from a cardiologist or HF specialist nurse the information or prompt that they are happy for me to have an end-of-life conversation with a patient. …Having the sort of permission to do that would actually be really helpful, because it would make me say, ''Okay, good, they think that there is not much else that can be done. She's on maximum medical therapy to help and this is about, you know, the focus is on managing symptoms and that's just so much of an easier conversation if it feels like we are all sort of singing from the same hymn book." Possible solutions for creating such a prompt were discharge summaries, patient notes or patient passports that highlighted the importance of having an ACP conversation. A nurse suggested: "The discharge summary can have a box which serves both as a prompt and as a means of communication between the cardiologist and the GP saying, 'Has any end-of-life conversations gone on?' Wouldn't that be good?" (R5) All participants highlighted the importance of being able to have telephone or face-to-face conversations with colleagues to keep updated on patients' end-of-life care. Yet, structural changes to clinical services meant that some nurses were no longer based in the same building as GPs and communications were less direct.

 Prioritising time for ACP
A c c e p t e d M a n u s c r i p t -B J G P -B J G P . 2 0 2 0 . 0 9 7 3 A common option for GPs to create more time was to have an ACP conversation in the context of a home visit: "I think having two hours after surgery to go and see this family was vital to have a successful end-of-life conversation. I see these kinds of patients after work or on the way home. It was helpful to have as much time as we needed to discuss these things. We continued with the discussion until it naturally came to an end. This is incredibly valuable."

Strengths and limitations
The views of 24 HCPs with diverse roles in primary care and a range of experience in ACP and HF generated detailed insight into identifying ways of improving clinical practice. The combination of rural and urban practice settings generated valuable findings that might be transferable across the UK.
The interviewer's clinical role as a GP helped participant-researcher rapport by an understanding of existing working culture. 37 However, to ensure rigour, a second, nonclinical researcher (JM) contributed to data analysis. 38 Two other researchers (BW, AMB) participated in the key stages of data analysis in order to obtain different perspectives and therefore to enhance the credibility of the analysis. 36 A limitation of the interview sample was that it included only HCPs from primary care.

Implications for research and/or practice
Developing shared decision-making tools in HF or a prompt list for patients was advocated by HCPs. These tools might mitigate against HCPs' fears of causing unnecessary alarm and provide a platform for ACP conversations to take place. Given the paucity of shared decisionmaking tools in HF, 46 the need for their development seemed supported by this study. While these tools might be helpful, the literature also cautions against using them rigidly by allowing the conversation to degenerate into a tick box exercise. 47 ACP conversations need to remain person-centred and tailored to the individual patient 48 .

CONCLUSIONS
This clinician interview series provided some key suggestions from GPs and nurses on how to overcome some barriers to their engagement with ACP in HF. HCPs recommended a prompt list for patients, a shared decision-making tool, a communication prompt between primary and secondary colleagues, and practice-based ACP training as approaches to improve their current practice. Findings from this research can contribute to the design of corresponding interventions to improve the implementation of ACP in HF.