Introduction

During the past 20 years, increasing attention has been given to the benefits of specialized palliative care (SPC). Previous research has shown that early implementation of SPC has a significant positive effect on quality of life (QOL) and survival [1, 2]. For patients with advanced cancer, SPC is advocated within an integrated, multidisciplinary care framework in which it can be considered alongside life-prolonging treatments or as the main care approach [3–5].

Despite growing evidence that SPC should be a crucial component of advanced cancer care, it is often considered as terminal care. Standard oncology care today remains focused on curing and life-prolonging treatments and is often offered to patients without realistic discussion about prognosis, the potential benefits and limitations of treatments and the potential benefits of palliative care [6]. Several studies confirm that oncologists frequently refer their patients to SPC late in their disease course [7–9], or even not at all [10]. Preliminary studies of the barriers to appropriate referral to SPC for physicians have uncovered numerous factors related to the disease, the patient, the family and the referring physicians themselves [11].

Before appropriate referral to a palliative care service can be undertaken, an honest and realistic conversation about disease stage, prognosis, possible therapeutic options and the disadvantages of proposed treatments is crucial. Ideally, the advantages of palliative care as a complementary approach should also be discussed at this point [6, 12]. Initiating this conversation is a difficult task for the medical oncologist. Even experienced physicians often struggle when initiating such complex and emotionally laden discussions [13, 14]. Very few studies have been carried out to explore the essence of their struggle. This paper reports findings from an explorative interview study among medical oncologists to gather in-depth understanding of their experiences in communication with advanced cancer patients. The aim of this paper is to identify the barriers to introducing palliative care into discussions with patients with advanced cancer.

Methods

Between Sept 2013 and July 2014, we conducted face-to-face interviews with 15 certified medical oncologists working in academic as well as non-academic hospitals in Flanders, Belgium. All were involved in daily patient care, in in-patient as well as out-patient clinic. In these settings, the medical oncologists are considered the core members of a multidisciplinary care team that collaboratively develops a treatment plan for each cancer patient and in all stages of the disease. It is generally accepted that the medical oncologist is responsible for all medical decisions and to discuss these with the patient. Consequently, it is assumed that the medical oncologist is responsible for the initial conversation about palliative care.

The contact information of potential participants was obtained through the membership list of the Belgian Society of Medical Oncology (BSMO). Only Dutch-speaking members were taken into consideration; 82 letters of invitation were sent out and a follow-up telephone call was made to those showing interest in the study to make an appointment for the interview.

All interviews were conducted in Dutch by one of the authors (MH) and varied in duration according to each participant’s time restrictions.

The interviews were explorative with open-ended questions and were supported by a topic list compiled according to the results of an extensive literature review on the subject. The topic list was dynamic and was revised continuously during intermediate analysis. Complete anonymity of participants was guaranteed and the interviews were audiotaped with the consent of all participants.

Data collection ended when saturation was reached (i.e. when additional data did not reveal new ideas, concepts or themes) [15].

Analysis

Interviews were transcribed verbatim and the transcripts were analysed using the grounded theory principles of open and axial coding [15] and were coded with the qualitative data analysis software NVIVO v10. The Grounded Theory approach entails an iterative process of coding and categorizing the codes into major themes and patterns relevant to the research question [15].

The analysis began by reading and re-reading the transcriptions to get a preliminary understanding of the main themes emerging from the interviews. One of the authors (MH) coded the transcripts independently in open codes. During regular meetings with the co-authors, with expertise in oncology, social science and psychology, interpretation of the open coding process was verified and reconciled where necessary by comparing codes within and between transcripts. Subsequently, the open codes that had common elements were grouped together into subcategories and categories. The authors LD, KP and SVB re-read and re-coded random transcripts and consistency of the subcategories and categories was discussed until approved by all authors. The categorization of codes lead to the identification of clusters of barriers at different levels. In this study, we only focused on identifying barriers for each identified category.

Results

Of the 82 who received letters of invitation, 15 certified medical oncologists responded and were included in the study, 11 male and 4 female. Their ages ranged from 37 to 70 years and they varied in work experience from 8 to 44 years (Table 1).

Table 1 Characteristics of the medical oncologists who participated in the study
Full size table

The average duration of the interviews was 47.93 min with a range varying from 26 to 82 min.

Throughout the interviews, the participants reported various and heterogeneous barriers discouraging them from initiating discussions about palliative care, which we categorized as shown in Table 2. Illustrative quotes of the medical oncologists for the categories and subcategories are presented in Table 3.

Table 2 Categorization of the barriers discouraging oncologists from initiating palliative care in discussions with patients with incurable cancer
Full size table
Table 3 Illustrative quotes for the barriers discouraging oncologists from initiating palliative care in discussions with patients with incurable cancer
Full size table

Oncologist-related barriers

Emotional bond

Oncologists reported the difficulty they have initiating discussions on palliative care if they have known the patient and his or her relatives for a very long time (close emotional bond) (s012) or, conversely, if they do not know the patient at all (no emotional bond). A close emotional bond sometimes results in personal identification with the patient or with their personal situation. They explained, for example, that if the patient was someone young or with young children, they found it very difficult to bring up palliative care (s002).

Emotional discomfort

For a considerable number of oncologists, palliative care is still equated with terminal care and death (s012). They were reluctant to bring it up as it is considered as admitting defeat and personal failure (s001). Some of the oncologists experienced their own life-experiences of having lost a relative as a barrier to discussing end-of-life issues (s009).

Experience and competency

Oncologists noted that initiating difficult discussions is something you get better at over time (s028) and that inexperience could be a barrier, mainly reported by younger oncologists. They also explained that during their medical education, they had received very little or no formal training on how to handle or initiate difficult discussions (s009). Additionally, the medical education that they had received focused mainly on treating and curing and ignored the key principles of palliative care (s006). Some oncologists reported that being able to handle difficult discussions with patients and their relatives may simply be a matter of personality (s009).

Patient-related barriers

Patient characteristics

In hospitals with an ethnically diverse patient population, oncologists reported language and culture as important barriers. In some cultures, talking openly about death, dying or even palliative care is not acceptable because it is considered disrespectful or as causing loss of hope. Language barriers also make it difficult to have the sort of meaningful and deep discussion that is essential when talking about end-of-life issues (s012). Some oncologists brought up medical co-morbidities as a complicating factor, especially in the case of geriatric or psychiatric patients (s010). A few experienced the patient’s level of intelligence or education as a barrier to discuss palliative care because they fear that these patients are unable to absorb the information properly (s025).

Emotional reactions

Oncologists repeatedly reported difficulty in handling a patient’s emotional reactions. Sadness and anger were indicated as the most difficult emotions to deal with (s025). Patients often tend to reject bad news or to deny it, which obstructs the oncologist from introducing or continuing the discussion (s005). They also indicate that patients have an infinite trust and faith in medicine (s001), which gives them unrealistic hope and expectations about their life expectation or the potential for cure (s001).

Family-related barriers

Interactions

Most patients are accompanied by one or more close relatives during the consultation, which was generally reported as supportive. However, relatives can be very concerned to protect the patient from information that might be painful for them and therefore may directly or indirectly impede the oncologist from talking about palliative care (s025). Some reported family disputes as a barrier (s013).

Individual traits

Patients often tend to protect their relatives by presenting themselves physically better than they actually are. This creates or nourishes the unrealistic expectations of the relatives (s022). According to the difficulties they have in handling the emotional reactions of patients, oncologists also feel hampered by the emotional reactions of relatives (s013).

Barriers related to the referring physician

Referral

There are very few existing guidelines or protocols on when to refer a patient to a medical oncologist (s025). The vast majority of referrals are for the purpose of decisions regarding therapy rather than for diagnostic purposes. Sometimes patients are only referred for phase 1 trials or for a second opinion, and thus far too late in the disease trajectory (s006).

Attitude

Not uncommonly patients are referred to the medical oncologist with very little or even no communication about the patient’s disease stage (s022). This makes it almost impossible for oncologists to bring up palliative care in a first consultation.

Barriers related to disease and treatment

Disease trajectory

Oncologists noted that talking about palliative care was sometimes difficult because the patient is still in good physical condition or responds well to the anti-cancer therapy. As long as the patient feels well, despite poor prognosis, a discussion of palliative care is considered to be rather abstract because there are still some or more therapeutic options (s027). Due to the unpredictable development of some cancers, the patient does not always go through the conventional model of curative, life-prolonging, palliative and terminal phases of medical care (s006). In some cases, patients have an unexpectedly rapid progression, which, in addition to a delayed referral, can prevent the chance to have discussions about palliative care (s022).

Therapeutic decisions

Patients with advanced cancer may be referred to a medical oncologist for enrolment in a phase 1 trial, especially in academic centres. This may create overly optimistic expectations or bring up new hope for a treatment, putting the oncologist in a difficult position from which to bring up palliative care (s028). Another major barrier reported by oncologists was not having the capacity to identify the ideal moment in the disease trajectory at which to discuss palliative care because they cannot estimate the prognosis of the patient (s010).

Institutional/organizational barriers

Availability of palliative care services

Some oncologists refrain from discussing palliative care because there is no palliative care service available in their own hospital or because of the long waiting list for a palliative care service (s013).

Characteristics of the setting

For some, the pervasiveness of a hospital culture directed toward curative treatments with very little attention given to palliative care was experienced as a barrier, particularly in academic hospitals (s028). The possibility of phase 1 trials was often linked to this specific culture (s022).

Team dynamics and responsibilities

Lack of consensus across the care team had a perceived impact on the decision of the oncologist to discuss palliative care with the patient (s023). Instead of feeling empowered by the team, they feel solely responsible for the decision to introduce palliative care to the patient (s006). Oncologists’ own perception of their responsibilities in the care of the patient had an influence on their motivation to discuss palliative care; a few defined their mandate in terms of curative rather than palliative and therefore tended to avoid mentioning palliative care in their discussions with the patient.

Practical organization

Some oncologists mentioned in particular the lack of an available space to talk quietly and openly (s001), the practical organization of their consultations and the lack of available time due to a too high rate of consultations (s007).

Workload

For many of the oncologists, there is a wide range of responsibilities such as managing and supervising patient treatment, keeping patients and relatives informed, filling in electronic clinical records, checking medical and laboratory tests, managing the multidisciplinary care team, overseeing care of hospitalized patients and being on call and standby (s013). The combination of these activities can be demanding and stressful resulting in psychological and emotional exhaustion (s028). This may result in a barrier for oncologists to discuss difficult issues, especially when they have many ‘bad news discussions’ in one day.

Support and coaching

They all agreed that they have few opportunities for support, in particular after highly emotional discussions (s002). Their position at the top of their team, as they perceive it, reduces opportunities for the sharing of experiences or for sitting together discussing and reflecting on specific problems or stressful aspects of their job within the team or with colleagues from another team (s006).

Societal/policy barriers

Palliative care stigma

Oncologists frequently reported that talking about palliative care was avoided because of the stigma surrounding it and its associations with terminal care, death and giving up hope (s022). Among patients and even among the general public, there is an important misunderstanding of the concept ‘palliative care’ which withholds the oncologists to openly use the word ‘palliative’ (s012).

Education and training

All oncologists reported that they had received no formal training in communication skills (s012) or palliative care (s001) during their medical education or specialist training. Additionally, they were all educated in the curative model meaning that their medical training is geared to the imparting of information and skills with the goal of cure and disease control rather than the attitudes and competencies required to provide high-quality palliative care (s006).

Discussion

Palliative care in Belgium is largely delivered by the primary health care teams, with referral to SPC when some aspects of patients or needs are too complex and are unable to be met within the current care framework. In this model, palliative care is—not yet—standard for every advanced cancer patient despite scientific evidence of the advantages of palliative care [2, 5]. Additionally, this study reveals that the medical oncologists still entertain the ‘old’ concept of palliative care in their minds in which palliative care is thought to be the very last option once patients have exhausted all potentially curative or life-prolonging treatments and have reached the terminal phase of their disease. This concept of palliative care makes it difficult for oncologists to estimate when it is appropriate to initiate palliative care, with a high risk that is often initiated late in the disease trajectory. Moreover, we see that this concept of palliative care is still predominant among patients, families and on societal level. This reinforces the idea that referring a patient to or even talking about palliative care is giving up hope, or giving up on the patient. Many oncologists in this study have expressed the concern that a referral to palliative care would destroy a patient’s hope. In other words, the concept of palliative care, as stated by the WHO [16, 17], is not yet accepted among medical oncologists nor among the general public. These findings are supported by the results of a study by Granek et al. (2013) showing that the stigma surrounding palliative care services is an important barrier for oncologists to talking about end-of-life care and palliative care.

It is important, however, to recognize that patients do not necessarily need to choose between cancer treatments and palliative care but can take advantage of both in optimizing their quality and even quantity of life under a simultaneous care model [2, 5, 18]. In this model, palliative care starts at the moment of cancer diagnosis, together with the curative or life-prolonging treatment improving QOL throughout the entire course of the illness [18]. Shifting the mind-set of oncologists, patients, families and society to this ‘new’ model could remove the stigma and allow palliative care to be openly discussed as from the start of the disease trajectory. This change of mind-set would also allow medical oncologists working in academic settings or engaged in clinical research and burdened by the pressure of recruiting patients for phase 1 trials to overcome the barrier they experience when working in this setting. If palliative care is introduced at the moment of diagnosis, and included throughout the entire treatment trajectory, there should no conflict of interest for the medical oncologist when enrolling the patient in a phase 1 study [19, 20]. More education in palliative care, or structural implementation of palliative care in the medical education, could be an important first step to achieving this change of mind-set.

In this study, the medical oncologists repeatedly spoke about lack of training in communication skills during their medical training. Within oncology, communication is a core skill but research indeed confirms that few oncologists or cancer specialists have received formal communication training [21]. Hence, oncologists often feel uncertain about how to communicate with patients effectively and how to respond to the challenging and intense emotions that emerge in difficult discussions such as those about palliative care. The specific result of our study was also supported by the study of Granek et al. (2013). Specific training programmes to improve oncologists’ communication skills could overcome this barrier. However, in every medical discipline, bad news or difficult information must be given to patients and families; recognition that physicians in any medical specialty may require further training to improve their communication skills is therefore essential.

For many medical oncologists, oncology is a rewarding profession but also very demanding and stressful. Their intention to initiate a palliative care discussion is often hindered by accumulations of bad news, treatment failure, patient suffering and the prospect of imminent death, in combination with increasing administrative burdens, time constraints and practical imperatives. This may lead to personal distress manifesting as depression, anxiety and fatigue and eventually leading to burnout or depersonalisation (i.e. psychological detachment from the job) for which they report to receive little or no support. Indeed, in a survey of cancer care professionals in Flanders, 51 % of the oncologists suffered from emotional exhaustion and 31 % from depersonalisation [22]. In another survey among US oncologists, 38.3 % had high emotional exhaustion, 24.9 % had high depersonalization and 44.7 % had at least one symptom of burnout (e.g. a high emotional exhaustion score or high depersonalisation) [23]. Detachment has long been promoted as a necessary condition for good medical practice; however, recent research indicates that the acceptance by oncologists of their own emotional and empathetic responses does not necessarily influences their capacity for objective clinical reasoning but is in fact an appropriate attitude in of end-of-life decision-making [24]. It is important to note that it is the oncologists’ perception that they have very little opportunities to seek support, when in fact this could also be a consequence of their training and of the stigma of the physicians seeking emotional support following conversations about palliative care with patients and their families. Nevertheless, the improvement of the quality of communication about palliative care in oncology will also require efforts in supporting and coaching medical oncologists.

Strengths and limitations

Although this study provided a number of important results, it also has some methodological limitations.

During analysis of the interviews, we applied The Grounded Theory approach which entails an iterative process of coding and categorizing the codes into major themes and patterns. However, we only focused on identifying the essential subcategories and categories. The classification into subcategories and categories and the in-depth description of the barriers will eventually help us to develop a middle range theory on how barriers at different levels can enhance the communication on palliative care. It is clear that this study is only a first step inventory and follow-up work is needed on developing a ‘grounded’ theory for communication on palliative care in oncology.

We only interviewed medical oncologists in Flanders, so the sample of this study is not representative for all Belgian medical oncologists. Similarly, our findings cannot be generalized to physicians from other disciplines such as organ-oriented oncologists, surgical oncologists or radiation oncologists.

Additionally, all oncologists interviewed in this study participated voluntarily which may indicate that despite the reported barriers, they are aware of the advantages of palliative care but do initiate discussions of palliative care; we could expect the barriers to initiating palliative care to be even more pronounced among oncologists who are not convinced of its advantages and tend to avoid it.

Despite its limitations, this study provided an opportunity to get inside the minds of the medical oncologists. To our knowledge, it is only the third study of his kind; most existing studies focus on the barriers for referral to palliative care services while much less is known about the very first discussion on palliative care prior to the referral process. It might be possible that some of the barriers for referral to palliative care are related to the barriers oncologists face when initiating it.