Methods
We performed a single-centre, qualitative study using semi-directed interviews. All the medical and paramedical staff (physicians, nurses, nurses’ aides) working full-time in the Cardiology Department of the University Hospital of Besançon, France at the time of the study were invited to participate.
To be eligible, physicians had to be full-time, qualified staff cardiologists whose practice included delivery of care to patients in any of the units composing our Department, namely: (1) the acute cardiac care unit that admits cardiac emergencies directly (i.e. patients do not transit through the emergency room of the hospital); (2) a short-stay unit for scheduled invasive procedures; (3) a long-stay unit for hospital admissions for reasons other than scheduled procedures and for post Cardiac Care Unit care; and (4) an out-patient unit for ambulatory procedures. Clinical trainees who were not yet qualified MDs and visiting practitioners were excluded. Physicians were invited to participate by personal invitation.
To be eligible, nurses and nurses’ aides had to be full-time qualified staff, and could be from any unit, and on either day or night duty. Nurses were invited to participate through posters calling for volunteers in the common areas of each unit, plus oral information during staff meetings.
Interviews were performed between 30 March and 17 July 2017 by a qualitative researcher with experience in qualitative research, biostatistics and epidemiology (FE (female)) and by a sociologist with experience in clinical research and ICU care (NMB (male)). About the interview guide, participants were asked to describe how they experienced end-of-life situations in the Department. As with all qualitative interviews, the questions were open ended and intended as a prompt to get the respondent to talk about the aspects that were most important to them, and voiced in their own words. Interviews were performed in a private office within the Cardiology department. No other persons were present during the interviews.
All interviews were recorded in full and transcribed for later analysis. Data were encoded to guarantee the anonymity of the participants. Interviews were subsequently coded using thematic analysis as previously described elsewhere [
8]. Briefly, interviews were coded independently by 2 of the coauthors (FE, NMB), the aim being to identify and categorize the different themes occurring in a cross-sectional manner across all interviews, (i.e. topics common to several individuals). Each theme is then considered as a meaningful independent unit of discursive language. The different themes that arise during the interviews are recorded, taking into account major themes (significant points that are of major importance and well developed by the participants) and secondary themes (less well developed by the participants). This first level of analysis was performed individually by each researcher, then meetings were held to harmonize and decide on the major and secondary themes to be retained, and their regrouping into subject categories. Differences in interpretation were resolved by discussion and consensus.
The study was approved by the Clinical Ethics Committee of the University Hospital of Besançon, France at its session on 8 February 2017. Informed consent was implied by the fact that all participants volunteered to be interviewed. Participants were made aware that quotes from their interviews may be used in scientific publications to substantiate the discussion. NVivo software was used to manage data (NVivo, version 11, QSR International, London UK).
Discussion
This study reveals several interesting findings regarding the perceptions that physicians and nurses have about end-of-life issues in a large, university hospital cardiology department. Firstly, physicians present an overwhelmingly curative frame of mind, and view the discipline of cardiology as a very active one, with a large therapeutic armamentarium. Their aim is to provide curative care at all costs. Second, nurses, on the other hand, clearly take a different but complementary approach. They have a less invasive attitude, and are more inclined to acknowledge that patients are at the end-of-life and should be labelled as palliative care. The nurses seem eager for there to be some guidance, or official consensus about when a patient should be considered to be in palliative care, and how this decision should be materialized, since the changing attitudes of different physicians who successively take care of the patient can cause incomprehension and as a result, distress among the nurses. Overall, our findings show that practices are highly individual, differ between medical and paramedical staff, and in any case, are not at all standardized.
Practices are currently far from the recommendations of professional societies regarding advance care planning and end-of-life discussions [
1,
3], although few concrete indicators exist as to how this should be materialized in the practice of cardiology. Indeed, in their review of existing quality indicators in national Swedish policy documents relevant to palliative and end of life care, Lind et al. reported that no indicators relevant for palliative or end of life care were present in guidelines in the field of cardiology [
9]. The Australian National Heart Foundation guidelines for multidisciplinary care for heart failure patients are useful in this regard, as they lay down clear directions with checklists for practical use in implementing their recommendations for best practice [
10,
11]. Similarly, there has been some discussion recently regarding the triggers that should prompt physicians to initiate end-of-life conversations or advance care planning. For example, Denvir et al. reported from an interview study with patients and carers that an estimated 1-year mortality risk of 20% or higher should prompt physicians to initiate end-of-life planning [
12,
13].
The active and curative attitude of the physicians reveals what they consider to be the boundaries of the discipline of cardiology. Indeed, they describe a therapeutic approach akin to a conveyor belt of care: once the patient is admitted and steps onto that conveyor belt, they are moved along as the physician proposes one therapy, intervention or device after another, until all possibilities have been exhausted. Indeed, in a secondary analysis of qualitative data collected through the Leadership Saves Lives initiative, Flieger et al. described this phenomenon as the “power of momentum”, whereby once the patient presents with an acute need, that increases the likelihood of intervention [
14]. With the outstanding progress in medical and device therapy in recent years, there is now a very large choice of therapies available, and thus, for a long time during the course of disease, physicians can offer a new therapy every time the previous one has failed. In a recent qualitative study investigating the reasons why doctors provide futile treatment at the end of life, Wilmott et al. reported that one of the main drivers of futile end-of-life care was the characteristic of treating doctors of being oriented towards curative treatment (“the therapeutic imperative”), and the desire to satisfy the patient, the family and the medical professionals themselves that “everything possible had been done”. [
15]. The result is that the physician is always in a state of expectation, waiting to see the effects of the most recent therapy – and very often, it may work, and the patient may recover. But the successive proposals also serve to postpone the moment when it will become necessary to talk about the end-of-life. Indeed, for a variety of reasons, palliative care is not often proposed to cardiac patients, or may be proposed too late in the trajectory of disease, thereby denying cardiac patients access to end-of-life care [
16,
17].
For the physician, once that patient has reached the end of the “cardiology” conveyor belt, and the only remaining option appears to be inevitable death, then most physicians believe that it is someone else’s job to take over the patient’s management beyond that point, i.e. from there until death (e.g. the nurses, or the intensive care unit, or the palliative care department [
17]). In a survey of clinical attitudes and self-reported practices regarding end-of-life care in heart failure, Dunlay et al. reported that among 95 clinicians interviewed, many reported discussing end-of-life wishes when the patient’s health status worsened, and that the most common reasons for referral to palliative care were that the patient had no other heart failure therapeutic options [
16].
The American sociologist Everett Cherrington Hughes, in his work entitled “Men and Their Work”, describes how professionals pose the limits of their professions, and stipulate the division of labor within this context [
18]. Our cardiologists appear to consider death to be beyond the boundaries of their profession, and not to be part of their role in the division of labor. As Hughes also stated from sociological observations, “…the delegation of dirty work to someone else is common among humans” [
19], and this appears to be a basic human behavior at play here. The physicians of our Department believe their role is to cure and to offer therapeutic solutions, and they would gladly have someone else take care of the patients who are dying and can no longer be saved, and who are, as such, no longer candidates for the highly active and interventional opportunities the cardiologist can offer. This finding is in line with the findings of the qualitative study by Flieger et al., where respondents indicated that a lot of cardiology professionals are afraid of palliative care, and that cardiologists “don’t believe in it” [
14]. The authors relate this rejection of palliative care to the tendency towards intervention among cardiologists.
With all the therapeutic solutions at their disposal, cardiologists are not yet accustomed to thinking about when these therapies might reasonably be de-escalated or scaled back. Indeed, the “trained to treat” attitude described by Willmott et al. in their sample of 96 physicians from various disciplines underpins the perception of many physicians that death is a failure, leading to a reluctance to acknowledge the role of palliative care [
15,
20]. Physicians in our study cited lack of time as a reason for not engaging in end-of-life discussions. In a recent telephone survey of physicians in the United States, Fulmer et al. reported that the number one barrier to advance care planning conversations, cited by two-thirds of respondents, was lack of time [
21]. In the case of the respondents in our study, this may just be an excuse, since most people are naturally averse to talking about death. Instead of saying they do not have the time, it might be more accurate to say that they do not take the time. It has previously been reported that discomfort with end-of-life conversations leads physicians to avoid addressing the topic with the patient [
15], while overall, 30% (28 of 94) of clinicians reported “low” or “very low” confidence in initiating EOL or hospice discussions in the survey by Dunlay et al. [
16].
At this juncture, the nurses’ role takes on its full importance, as they take over the care of the dying patients. Indeed, they are often instrumental in bringing the patient to this point, in that they coax physicians to stop curative therapies, or invasive monitoring, and increase comfort care, with a view to leaving the patient to die in peace. When there is incomprehension about the physician’s intent, or prescriptions for curative interventions for a patient whom the nurse feels to be dying, then this creates distress among the nurses. This is coherent with previous reports of high moral distress among nurses in critical care units, particularly due to futile care or physician-related factors, such as having to assist a physician who the nurse feels is incompetent [
22]. The incomprehension can be due to lack of communication, as the physicians always believe they have a sound scientific rationale for their prescriptions, but the nurses may have a different perception from having spent more time with the patient, and often also with the family. In a prospective opinion survey of critical care providers, Frick et al. reported that nurses gave more pessimistic judgments and considered withdrawal more often in dying patients than doctors did [
23]. These authors also relate this finding to the fact that, as observed in our study, nurses spend more time with the patients than the physicians, often accompanying them through emotionally charged circumstances. For this reason, they may be more aware of the suffering of the patients, and more inclined to admit treatment failure [
23]. Families and patients often confide things in nurses that they won’t say to the physician. The role of the nurses in communicating the wishes and values of the patients and their families when discussing individual cases is thus vital, and the information they can provide is complementary to the physician’s predominantly medical evaluation. This is coherent with a previous report of a grounded theory study among nurses in intensive care units in nine countries, which reported that although nurses do not make the final end-of-life decisions, they engage in the core practices of consensus seeking, which involves coaxing, relaying information and giving voice to patients’ feelings, and emotional holding, which involves comfort-giving [
24]. Although the nurses seem willing to take on the role of having end-of-life discussions with the patients, they do feel restricted in the scope of their conversations with patients by what the physician has already told the patient. In this regard, in line with efforts to involve all healthcare professionals in improving end-of-life management across the spectrum of the healthcare pathway, it is essential that nurses be involved in the discussions regarding end-of-life decisions, and they may also be involved in the communication with the patient about end-of-life matters. According to a survey performed by Aleksova among cardiologists, cardiology trainees and cardiology nurses providing care for heart failure patients, nurses were more willing than physicians to initiate and engage in end-of-life discussions with patients [
25]. In this same study, among non-physician clinicians, advanced practice nurses were deemed to be most acceptable to be involved in decision-making relating to goals of care [
25]. However, it is important that the physicians do not shirk responsibility by having nurses play a role that the patient traditionally expects the physician to play. This could be perceived by the patient as being abandoned by the physician.
Advance care planning clearly remains underutilized in our region. Advance care planning is a dynamic process that aims to encourage and empower the patient to discuss their disease course, prognosis and likely outcome with their family and with their physician, especially the healthcare trajectory they wish to follow [
26]. The patient can consign his/her preferences and desires for end-of-life care in written form, such as AD, in case he/she subsequently becomes decisionally incapacitated. Advance care planning aims to ensure that patients receive care that is in accordance with their wishes, particularly at the end-of-life, and also helps to reduce unwarranted interventions [
26,
27]. Despite evidence that advance care planning positively impacts on end-of-life care [
26,
28], a range of barriers remain to effective and systematic implementation of such services in clinical practice [
28,
29]. In France, despite a generally positive attitude towards AD among the general public, reportedly only 2.5% of patients have actually drafted AD [
30,
31]. In a recent review of AD from hematology departments, Trarieux-Signol [
31] et al. describe the current French legislation in detail, noting in particular that AD are binding, except when “the content of the AD appears to be manifestly inappropriate or inconsistent with the patient’s medical condition”. This is coherent with the findings of our study, where physicians said that while they will enquire if AD exist and take note of what they contain, they might actually pursue a different course of action if they judge the content of the AD to be “off the mark” or blatantly irrelevant to the patient’s condition. This echoes the findings of Aleksova et al., who reported that family members’ difficulty accepting the poor prognosis of a loved one, and the patient’s own difficulty understanding their prognosis and therapeutic options represent significant barriers to goals-of-care conversations [
25].
This attitude, whereby the physician and family and/or other bodies, such as the state, may override the individual’s choice in the name of a greater good, is more common in European countries, and in clear opposition with the American attitude whereby the patient’s autonomy prevails above all [
32]. Indeed, in their investigation of the challenges and opportunities for engaging palliative care after myocardial infarction, Flieger et al. reported the difficulty of delivering care that is aligned with the patients’ preferences when there are limited opportunities to ascertain what these preferences may be, once treatment is under way [
14]. In this context of different societal attitudes to such key questions as patient autonomy, it is likely that culturally-specific models of end-of-life care are warranted to ensure that practices are in harmony with the prevailing principles of healthcare delivery and perceptions of the doctor-patient relationship. Similarly, it may be difficult to implement “one size fits all” recommendations for end-of-life care in the presence of such starkly different world views. Efforts to correct misperceptions about the meaning of palliative care (equating it with end-of-life care and pending death) may be warranted. Also, a multidisciplinary team approach with coordinated collaboration between healthcare professionals could help to ease transitions between healthcare providers, make appropriate care and symptom relief available in a timely manner, and achieve the desired outcome of a peaceful death in the location of the patient’s choice [
10].
This study suffers from several limitations that deserve to be underlined. Firstly, it is a single-centre study and therefore, may not be generalizable to the whole population of physicians and nurses/nurses’ aides working in cardiology. However, the volume and type of activity, as well as the profile of the physicians and nurses is similar to that observed in other, similar sized university hospital cardiology departments, so it is plausible that while practices may vary according to local culture, general attitudes among this professional group are likely similar elsewhere. Although our results are not wholly unexpected, they nonetheless open avenues for further, multicenter studies that would help inform about practices in other, similar sized institutions. Secondly, other hospital personnel likely to be included in end-of-life discussions, such as intensive care unit physicians, palliative care physicians, social workers and psychologists, were not included in this study. Thirdly, we cannot rule out the possibility that previous memorable experiences of participants, either positive or negative, and in either their personal or their professional lives, may have influenced the discourse substantially. Fourth, a wealth of other topics were raised by both groups of professionals that we are unable to address here for lack of space. These points warrant further exploration and analysis. Finally, it would be interesting to complement this analysis by interviewing families who have experienced the end-of-life of a loved one in our Department, in order to understand the impact of the caregivers’ attitudes on the families.