Advance care planning conversations with palliative patients: looking through the GP’s eyes

A qualitative research design was chosen in order to obtain in-depth insights into GP experiences with timely ACP discussions with palliative patients. Our study comprised semi-structured interviews with GPs located at different practices throughout the Netherlands. Data were collected from May to December 2016.

We recruited GPs who had participated in an ACP training programme recently developed by the Dutch College of General Practitioners (Nederlands Huisartsen Genootschap; NHG). This training consisted of two modules with a total duration of six hours. It comprised theory and peer-review practice and focused primarily on discussing EoL issues (Additional file 1). Purposive sampling of trained GPs was employed to ensure the sample represented a diverse range of GP and practice characteristics such as age, sex, experience and size and location of the practice (Table 1). Non-participants mostly declined because of time issues. Recruitment was stopped after saturation was reached and confirmed in three additional interviews.

An interview guide was developed starting from the research gap identified in the introduction, and it was inspired by the framework by Boer et al. [25] The guide was discussed and amended if needed by ABW, SG, HvD and TB. It consisted of four topics, each comprising multiple open questions (Additional file 2). Nine interviews were conducted by two authors (ABW, HVD and/or SG) and the remaining eight by one author (ABW or HVD). See Table 2 for more information regarding researchers’ characteristics. At the start of each interview, participants were informed that the goal was not to evaluate the training itself, but to talk about their experiences with ACP with palliative patients.

The first few interviews took place face-to-face at a location most convenient for the interviewee, so that the interviewers could get a good feeling with the matters discussed. Subsequently, interviewees were given the choice between a face-to-face or a telephone interview. The latter was chosen in all cases for geographical or other reasons making the interview more comfortable for interviewees. The interviews were conducted in Dutch and lasted from one to one-and-a-half hour and were recorded, for which all participants gave consent. The interviews were transcribed verbatim by an official agency, and names of participants were anonymised.

Since the goal was to understand the experiences and motivations of GPs, an interpretative approach was used, in which the focus is understanding the world as others experience it [26, 27]. A constant comparative method was chosen for data analysis the data [27, 28]. First, two authors (ABW and HVD) independently read, reread and openly coded the first seven transcripts. These codes were discussed and compared until an initial codebook was established. Two researchers (ABW and HVD) independently analysed the rest of the interviews by applying the coding framework and modifying it through an inductive, iterative process. In this phase, codes were discussed, added, modified or merged if necessary. Finally, these codes were discussed, axially coded and combined into categories and themes (by ABW, SG and YE). The software programme Atlas.ti 7 was used in managing and analysing the data. In reporting the data the Standards for Reporting Qualitative Research was used [29].

The same interviewee suggested governments could play a role in normalising communication about death and dying, “for example through television spots. Along the lines of ‘you organised your house, your holiday, etcetera. Have you also considered how you would like to organise the last years of your life, or discussed it with your GP?’” Other GPs subsequently warned of a possible negative image of ACP, for example arising out of feelings of negligence. “What is the government’s objective? Are they concerned I’m getting too old? Am I too expensive?” (R12) Proactively mentioning the approaching death might be experienced as an efficiency measure for patients as well as for the general public. This may lead to a negative social perception of ACP.

The third theme regards the initiation of ACP. Categories within this theme concern starting the conversation and issues in the relation with the hospital in initiating ACP.

Although fairly obvious, GPs emphasised it is very important that ACP is tailored to each individual patient. After all, just as every person has their own lifestyle, everyone also has his or her own ‘dying-style’. ‘Everyone does it his own way. If someone fights until the end, sure there’s probably a reason for that too… That’s just how he is.’ (R5) In this context it was also mentioned that being explicit, e.g. about future wishes and needs, and preparing for what might be coming simply might not fit one’s style. ‘Some have no desire at all to discuss EoL issues. Why should you then bring up all kinds of issues, if that doesn’t fit his or her personality at all?’ (R12).

An interesting mix of ACP being a hot topic and death not being regarded as a natural part of life these days was found. Just as interviewees in this study, the Economist Intelligence Unit called for encouraging people to openly discuss their EoL wishes and normalise the conversation about it [30]. Moreover, it is often recommended that GPs act as facilitators in ACP [31], and multiple ACP trainings have been developed specifically for GPs [18‐20]. A survey among 516 GPs for example found that 97% felt that general practice plays a key role in the delivery of care to people approaching the EoL [32], and the majority of patients appreciate it if their GP takes the initiative [33]. Although a 2013 review found evidence for a GP attitude that patients should initiate discussions [31], our study shows interviewees feel they have a role in addressing ACP discussions, even if patients do not have an active stance towards these discussions. Like others, we found that GPs find it hard to identify and discuss palliative care needs with patients with organ failure [23, 34]. Moreover, our finding that initiating ACP is difficult when patients are still being treated in hospital is also confirmed in other literature [35‐37]. It was mentioned in our study that ‘it is not ideal to ask about their wishes when something happens, i.e. when someone must be hospitalised.’ However, these clear milestones offer the possibility to initiate ACP. Research showed that these events, such as hospital admissions or exacerbations, can serve as a helpful starting point for discussions about wishes and needs [38], and may result in ACP taking place more frequently [39]. Lastly, a tension between ACP’s focus on the patients’ own direction and an in medical ethics recognized right not to know is present. Getting patients to face their approaching death is, in a way, a form of medical paternalism. Such paternalism can be justified as long as GPs acknowledge that not wanting to know (where things might be going or what possible ways of care planning are), can also be a form of autonomy. Nevertheless, this can be quite difficult for GPs, as ACP is mainly conducted in the context of anticipated deterioration. I.e., if a patient loses capacity, GPs should make use of information obtained from the ACP process to guide their decision making [40]. The concept of genuine consent, in which provision of information is cyclically adjusted to current wishes and abilities of patients, might better fit the context [41].

A strength of this study is that interviewees had already looked into the matter closely, as they had taken part in the ACP training of the NHG. At the same time, participants had been sufficiently interested in ACP to pursue ACP training, which might have biased the data. Also, the know-how gained from the training might have influenced their experience with actual ACP discussions. However, although generalisability of qualitative research is always an issue, this study reflects similar findings in other studies. Furthermore, although conducting interviews by telephone has been considered inferior to face-to-face interviews [42], recent research showed that rich narrative data collection can be achieved using this method [43]. Telephone interviews may even allow respondents to feel more relaxed and able to disclose sensitive information [42]. We have paid proper attention to the special character of interviews on the phone, but found no indications that in those interviews any aspects would we paid less attention than needed. Moreover, although it was repeatedly emphasised that the palliative stage is broader than the terminal phase, interviewees predominantly spoke about patients in their terminal phase. Despite their training, multiple GPs experienced a barrier in starting ACP conversations, the cause of which might be their focus on the terminal phase. The fact interviewees oftentimes regarded palliative patients as being terminal underlines the awareness needed for what palliative care entails.

A focus on the terminal phase and death might result in difficulties with starting ACP, and in ACP conversations taking place too infrequently. The goal of ACP is to define, discuss and review goals and preferences for future medical treatment and care, in order to improve the quality of the remainder of patients’ lives [1]. Death and EoL arrangements are in most cases only a small part of remaining life. Interviewees indicated that, mainly with chronically ill patients, clear milestones like exacerbations or hospitalisations can be used to start ACP. This finding dovetails beautifully with, when initiating ACP, staying close to the now is helpful: asking what is important for patients at this moment might facilitate conversations about needs and wishes in an accessible way. Useful example questions are described in the Supportive and Palliative Care Indicators Tool [44], as well as by the Mount Vernon Cancer Network. For example: ‘What in particular are you concerned with at this moment?’ and ‘What or who gave you support in previous situations?’. These questions can serve as a good basis for timely initiation of ACP.

We recommend taking the above into account in future ACP training programmes, as it will lead to open and transparent conversations about realistic expectations, and subsequently wishes and needs can be tailored accordingly. Patients can benefit from this: research shows that the decline of QoL in palliative patients with false hope was larger than the decline in patients with realistic hope [45‐47]. Regarding the lack of collaboration between GPs and medical specialists, which might result in overtreatment and associated costs in terms of patients’ QoL as well as societal costs, interviewees suggested regular multidisciplinary meetings or simply picking up the phone more often could help. Recent research indicated that integrated care and multidisciplinary training in ACP might bridge gaps [48, 49]. Another suggestion is better data sharing: an English initiative enabling ACP and improving communication and coordination for example, had successful outcomes (such as 77.8% of patients dying in their preferred place) [50]. Finally, research showed transitional care pathways can be effective in reducing hospital readmissions [51, 52]. We suggest future research to further look into finding ways to achieve a multidisciplinary approach in ACP, and the role of medical specialists in ACP.