What Makes Oncologists Feel Compassion Fatigue? Qualitative Findings From Korea

This study explored oncologists’ compassion fatigue in their practice with cancer patients and families. Qualitative in-depth interviews were conducted. A purposeful sample of nine oncologists was recruited from three hospitals in Korea. We used a hybrid of the theory-driven deductive approach and data-driven inductive approach based on the transactional model of physician compassion. Oncologists shared contexts from which their compassion fatigue stems based on four categories: physician factors, clinical factors, patient and family factors, and environmental and institutional factors. Given the nature of their job, oncologists should receive interventions that improve resilience and reduce compassion fatigue. We suggest interventions that promote appropriate levels of compassion and reduce interfering barriers.


Original Research
Advances in medical oncology have led to notable improvements in the 5-year survival rate for cancer in Korea, increasing from 41.2% in 1993to 1995to 70.6% in 2012(Korea Central Cancer Registry, National Cancer Center, 2018. Although the survival rate has improved, approximately one third of Koreans are likely to be diagnosed with cancer in their lifetime (Korea Central Cancer Registry, National Cancer Center, 2018), and cancer is the leading cause of death in Korea (Statistics Korea, 2019). Oncology health care professionals play a significant role in managing and treating the burgeoning cancer patient population.
Oncologists face unique challenges in working with cancer patients and family members (Shanafelt & Dyrbye, 2012). Medical care is delivered over a long term, often resulting in close relationships between the medical team and patients. Families affected by a cancer diagnosis and treatment are often overwhelmed and may require significant reassurance and emotional support. Requirements for oncologists to convey compassion and empathy in times of patient crisis are emotionally demanding (Le Blanc et al., 2001). Compassion and empathy, which are often used interchangeably, are distinct from each other (Fernando & Consedine, 2014). Empathy refers to a capacity to experience another's perspective or state of being (Wilson & Lindy, 1994), whereas compassion involves the desire to help or relieve the suffering of others (Goetz et al., 2010). Oncologists often witness patients' unbearable pain and death . In a study that explored medical oncologists' challenges in managing the emotional labor of their work, oncologists acknowledged that although compassionate care is central to the delivery of authentic care, they face challenges in balancing "the importance of intimacy versus the art of detachment" (Broom et al., 2016).
Previous studies reported that oncologists experience a heavy workload (Grunfeld et al., 2005) and high levels of stress and burnout across the globe (Blanchard et al., 2010;Girgis et al., 2009;Grunfeld et al., 2000;Lyckholm, 2001;Whippen & Canellos, 1991). One study by Dimoska et al. (2008) in Australia found that of 134 cancer specialists, 62% experienced stress in all practice situations, especially when discussing high-cost drugs with patients or treatment failure. Also, multidisciplinary health care professionals in pediatric oncology (medical, nursing, and allied health staff members) reported both stress and reward related to patient care and interactions with children (Bowden et al., 2015). In the 1990s, Whippen and Canellos (1991) found that of 598 oncologists surveyed in the United States, 56% reported experiencing burnout in their professional life, describing a sense of failure and insufficient breaks from work. More recent data found that a substantial percentage of oncology health professionals (e.g., nurses, oncologists and palliative care physicians, other health professionals, researchers, and administrators) experienced considerable burnout as well, in all three domains of emotional exhaustion, depersonalization, and diminishing personal accomplishments (Blanchard et al., 2010;Girgis et al., 2009). Notably, the prevalence of emotional exhaustion and feelings of depersonalization were significantly higher among physicians than allied health professionals and support staff members in oncology in Canada (Grunfeld et al., 2000). How compassion fatigue relates to the three domains of burnout (emotional exhaustion, depersonalization, and decreased sense of personal accomplishment (Maslach et al., 1986) or the composite term burnout needs to be further examined.
Compassion fatigue as a cost of caring among allied health care professionals has emerged as a crucial topic among professionals directly involved in the supportive and mental health care of oncology patients (Figley, 1995;Najjar et al., 2009). The perspectives of oncology nurses (Cho & Jung, 2014;Gentry & Shockney, 2018;Hunt et al., 2019;Potter et al., 2013;Sullivan et al., 2019) and oncology social workers (Simon et al., 2006) have been described in the oncology literature. However, oncologists' compassion fatigue has not been adequately studied or understood. Research on oncologists' quality of work life has focused on stress and burnout across the globe (e.g., Ghali et al., 2019;Girgis et al., 2009;Shanafelt & Dyrbye, 2012;Sherman et al., 2006). Recently, Kleiner and Wallace (2017) explored key predictors of burnout and compassion fatigue among oncologists in Canada. But the voice of oncologists is missing from the research. Qualitative understanding of oncologists' experiences related to compassion and how their compassionate response generates fatigue needs more academic attention.
The Transactional Model of Physician Compassion proposed by Fernando and Consedine (2014), which explains how physicians' compassion is influenced by physician, patient and family, clinical, and environmental and institutional factors, provides a valuable theoretical framework to understand how compassion fatigue among oncologist forms and is experienced. This model posits that as the main agent of compassion, the physician's personal ability, and motivation to provide a compassionate response toward patients is critical. Also, how patients and families interact with the physician is crucial in understanding the physician's level of compassion. In addition, the dynamics of clinical features are at play, such as working with cancer patients, who often face lifethreatening situations and severe pain. Finally, the physical environment and institutional contexts likely influence compassion. Building on contextual factors related to physician compassion identified from this theoretical model (Fernando & Consedine, 2014), the purpose of this study was to understand the dynamics of oncologists' compassion fatigue.

Inclusion Criteria and Recruitment
Our study inclusion criteria were oncologists (a) currently working in oncology units in hospitals in Korea and (b) providing oncology patients or families with medical services in the field of oncology. We recruited participants in Seoul and Gyeonggi Province, both metropolitan areas of Korea, because oncology care is mostly concentrated in these areas in Korea, at major hospitals. First, the research team searched websites of the oncology department of major hospitals providing oncology care for patients in Korea and generated a list of potential participants. Then, we requested recommendations from several oncology health care professionals regarding participants who could provide in-depth information toward our study purpose. We contacted potential participants via email to provide our information sheets with study announcements. Finally, voluntary participants responding to the research team were selected as study participants. A snowball sampling method was primarily used by asking potential participants to identify oncologists providing care. As a result, of 12 oncologists contacted, 2 indicated that they mainly did not work with oncology patients, and 1 oncologist declined to participate. As a result, nine oncologists were recruited from three hospitals with oncology departments. Recruitment ended when data were saturated (new themes were no longer found) and a determination of consensus was reached through discussions among the research team (Strauss & Corbin, 1998).

Interview Procedures
Ethical approval was obtained from the institutional review board of the university with which the principal investigator (second author) was affiliated at the time of this study. Before the interview, participants initially completed a brief survey that collected sociodemographic and work-related information, such as age, gender, education level, marital status, religion, years of experience as an oncologist, and years of work in their current hospital. The second and third authors, who had worked in oncological social work either as a clinician or academic professional for approximately 10 years, conducted in-person interviews in Korea at or near the participant's workplace based on their preference from July 2015 to January 2016. We emphasized that participation was voluntary and confidentiality was assured. Participants provided verbal and written consent to participate in this study after receiving a careful explanation and an information sheet.
Each interview was 60 to 90 min. The present study, as part of a larger research project, examined the participants' narratives on the open-ended guiding question: "Can you tell me about your experiences of feeling compassion and compassion fatigue when working with your oncology patients and family members?" Field notes on notable contexts or impressions from the interviews were taken during or immediately after the interviews and shared with the research team. The participants received a $40 bookstore gift certificate for their participation.

Data Analysis
Data were analyzed concurrently with data collection. With the permission of the participants, all interviews were audio recorded and transcribed verbatim. All transcripts and a priori codes were entered into qualitative data analysis software (ATLAS.ti; Muhr, 2009). We used a hybrid approach of qualitative methods for thematic analysis, inspired by a blend of the data-driven inductive approach with the theory-driven deductive approach (Fereday & Muir-Cochrane, 2006).
First, theme categories (physician, clinical, patient and family, and environmental and institutional factors) were developed a priori based on the theoretical model of physicians' compassion (Fernando & Consedine, 2014). The four theme categories (with detailed definitions, descriptions, and examples) were created in the code manual, which guided the analysis. In generating the code manual, the authors discussed how the theoretical framework applies to this study. Second, the second and third authors independently summarized the significant statements, phrases, and sentences from each transcript and compared their findings to ensure the applicability of the predetermined theme categories to the data. Third, the authors completed the inductive process by openly coding the transcripts line-by-line and grouping codes into clusters that corresponded with the a priori theme categories. Fourth, through constant comparisons and discussions in the research team, we organized the relationships between the open codes, clusters, and theme categories. Fifth, all authors independently or collectively reviewed and finalized the theme categories and their sub-themes to ensure the correspondence of the data with the core elements of the theoretical framework (Fernando & Consedine, 2014). Finally, we selected quotes that best illustrate each subtheme. All quotations were verbatim, but some quotations have been truncated (denoted by ellipses in the text) for ease of reading. Pseudonyms (Korean names) were generated for all participants during the data analysis to maintain their anonymity. Table 1 features the individual characteristics of the participants: nine oncologists (four male and five female) working in oncology units in hospitals in Korea. They were age 32 to 60 (M = age 39.4; SD = 8.6). Most (n = 7) were married, and they all were either current graduate students or had earned a graduate degree. More than half (n = 5) identified with a religion. All participants worked with inpatients and outpatients. Five participants worked with pediatric oncology patients and four worked with adult oncology patients for issues such as hepatobiliary pancreatic cancer, stomach and colon cancer, breast and gynecologic cancer, and other cancers requiring palliative care. Their experience in the oncology specialty was 3.7 to 31 years (M = 9.8, SD = 8.25); 3 had been working in this specialty for more than 10 years. Participants had worked in their current hospital for 10 years (SD = 10.06) on average, with a range of 31 months to 35 years.

Qualitative Findings
As shown in Table 2, this study identified eight subthemes relating to the experience of compassion and fatigue. We categorized these subthemes using the framework proposed by Fernando and Consedine (2014) in their Transactional Model of Physician Compassion, which describes the dynamic interplay between physician compassion and factors related to physicians, the clinical situation, patient and family, and environmental and institutional factors.

Physician factors
Strong empathy and identifying with the patient. Oncologists developed strong empathy toward their patients and experienced emotional suffering. The degree of their identification varied depending on their characteristics and background, including age, childhood experience, and family values. The more they identified with the patient, the more they considered the patient's perspective when making medical decisions. Dongmin (male, age 37), a surgeon for hepatobiliary pancreatic cancers, which have high recurrence rates, said he devotes attention to the patient's suffering and considers the patient's age and individual situation when he recommends surgery. Yeon (female, age 39) stated that she intentionally expresses how she identifies with the patient to the patient and family members. She wants them to feel that she is doing "her best in the most humane way" and hopes to alleviate their suffering.
Becoming a family member instead of a doctor. All the oncologists reported building long-term relationships with patients and their family because cancer treatments are usually long-term. Because this rapport is established over a lengthy period, a patient's suffering or death often has a deep impact on the medical providers, as if the event occurred in their own family. They experienced seeing "juxtaposed images" of the dead patient in their own children, as illustrated by Sunmi (female, age 34), a mother. Participants said such experiences prompted deep sadness. They also found themselves transitioning "from the mind-set of a medical provider to that of a caretaker," as exemplified by Yeon. This profound involvement affected not only their professional life, but also their personal life. Boundaries became blurred; a patient's death felt like more than losing a patient, and when a family member was lost, this further aggravated their suffering.

Clinical factors
Burden of life-or-death decision-making. As their most difficult task, oncologists noted the continuous burden of making critical decisions that would determine the life or death of a patient. It is a typical and crucial part of work for any medical provider to make an accurate assessment of the patient's status and decide on the most effective treatment and medicines for recovery. Yet for cancer patients, making even a minor decision could lead to consequences that result in life or death. Such a heavy burden weighed on the oncologists. The participants mentioned the burgeoning fatigue from constantly making decisions about the most effective treatment plan in the shortest time possible, which would likely have a major impact on the patient's life.
Feeling guilty about unexpected negative patient outcomes. Oncologists identified with their patients' suffering and channeled emotional pain even more profoundly when a patient's condition suddenly worsened or unexpected outcomes such as sudden death occurred. Such feelings varied based on the types of cancer in which they specialized; those who treated patients with cancers with poor prognoses reported feeling shock and even guilt when observing protracted recovery periods, suffering from side effects, or sudden death. When the patients developed negative responses to treatment and side effects, especially if not expected, the oncologists reported experiencing extreme pain. Although the doctors did their best to decide on the treatment they judged most prudent, they felt personally responsible, guilty, and regretful about their decision if the patient worsened or did not recover. They also said they felt as if they betrayed their patients, who implicitly trusted them. Relevant quotes include: "What if I did better? I don't know what, but something"; "I could have delayed it longer, or the death could have been more peaceful, less painful"; and "If I didn't do what I did, could he have recovered?" Experiencing pain from patients dying or in palliative care. Oncologists face ongoing deaths and become personally pained, especially when observing dying patients and having to explain death to patients and family members. Witnessing the extreme suffering that patients and families experience while receiving anticancer drugs or CPR until the last second of life, holding onto hope for extending life, caused intense personal suffering among the oncologists. Doctors specializing in pediatric oncology reported that even after a lengthy period in the field, they continue to re-experience serious traumas, including vivid flashbacks of children being severely bruised from CPR and having to console physically and emotionally devastated families. Facing frequent deaths, they described manifesting guilt and distress, and that these emotions do not resolve but instead alarmingly accumulate.

Patient and family factors
Mistrust from patient and family. Oncologists faced many challenging situations and found themselves emotionally exhausted when a patient or family displayed mistrust or a hostile attitude toward the oncologists and their treatment decisions. Cancer treatment requires long-term treatment and surveillance and even then, it often recurs and causes side effects. For successful cancer treatment and recovery, intimate interrelationships and collaborations among the patient, family, and medical team are crucial. Lack of trust Most patients come to the hospital thinking that a surgery will remove the cancer completely and they are cured. I also meet the hopeful patients with confidence that I can lead the situation in the best direction. Some actually get so well, but some live several months or several years, then they need another major treatment for recurrence. Also, some die. Mistrust from patient and family Doctor, patient, family, we all should stand at the same starting point looking in the same direction. If all of us stand in different places, the whole process is exhausting. If the family understands where we intend to go and responds in collaboration, I can explain the same things hundred times and I don't get tired at all. If the family appreciates me and my effort, I become so grateful, comforted, and full of energy. (Jaemin) Because I am an oncologist and deal with cancer, it is hard to keep a good relationship with the family. Even when the relationships started so good, if the kid's health gets worse, the relationships go to hell. . . . When the family does not accept what I try to explain and deliver, or they do not understand how I care so much for the kid, I get hurt, sad, and angry, all at the same time. (Sunmi) Some patients are hostile and critical from the beginning. They get suspicious of anything that I do. I explain things as they are, but they do not take them as face value. I am also a human being. I get hurt when I get treated like this. (Dongmin) Blame and guilt about negative medical outcomes I try to treat my patients in the same way. I have their best interest in mind. There are kids, of course, I want to treat them well so they can survive and thrive. But, depending on the outcomes, I get opposite responses from the families. I had two leukemia patients at similar ages. One of them got suddenly worse and was sent to ICU with hemorrhage. The other patient responded well to treatment. I did exactly same things to both of them. The family of the patient who did not fare well treated me as if I am a murderer. They tell me, "You made my kid like this. It is all because of you." The family of the patient who is doing well say, "Thank you so much, so so much. You saved my kid. You are a savior." I absolutely didn't do anything different for each kid. I feel, maybe how caring or well-intentioned I am does not matter at all. . . . I get blamed, criticized, hurt. I get scared of visiting the family. (Yeon) When a kid dies, parents naturally get distressed and angry, so they say things that are negative. Every single word that I hear from them stays with me. . . . At first, I think, "Can it be true what they told bad about me?" Then, after ruminating about the same words, I get to think, "What they told me how bad I was must be true." (Sunmi) Environmental and institutional Inadequate compassion due to work overload When I was a fellow, I remember every second was hard. I was the only fellow there, so on call 24/7. I received 100 phone calls. My ears were glued to the cell phone without a break. At the same time, I was acutely aware that every decision that I made in this crazy situation will make the patients live or die. It was too hard. (Sunmi) It was code blue. A kid died in ICU. I announced death and cried with the family for 2 seconds, and it was time to do a round. I cannot cry to other kids, so I run to the bathroom to wash my face and wipe out all the tears and go to other kids. I smile and talk with them. "How were you today? You will do so well soon." This kind of situation repeated every day and it was so hard. I never have enough time to collect myself and my feelings. (Yeon) and frequent conflicts between patient and doctor can lead to high levels of compassion fatigue and emotional distress to the extent that doctors "want to drop everything and run away." Blame and guilt for negative medical outcomes. Oncologists experience high levels of compassion fatigue when negative medical outcomes such as side effects, recurrence, and death occur. They reported feeling responsible for and guilty about the consequences, despite doing their best. Such suffering was aggravated when the patient and family blamed and criticized the doctor. Participants reported feeling deeply hurt when a family emotionally severed their long-term relationship if the patient experienced negative outcomes.

Environmental and institutional factors
Inadequate compassion due to work overload. Although they are aware of the need for compassion and close interactions with patients regarding their suffering and challenges, oncologists said they found it difficult to devote such attention to individual patients due to excessive workload (long working hours and volume of work). This conflict between their awareness of the need and desire to provide adequate care and the high workload put the oncologists in an emotionally exhausting situation. They even reported having insufficient time to grieve over patients' deaths.

Discussion
This study was designed to apply the Transactional Model of Physician Compassion to understand compassion fatigue among Korean oncologists. Consistent with the model, we found that four factors (physician, clinical, patient and family, and environmental and institutional factors) related to oncologists' compassion fatigue were aligned with the experiences of the participating oncologists in this study.
As the theoretical model posits, the oncologists' personal background and characteristics were identified as a determinant of their levels of compassion fatigue. Oncologists in this study described experiencing compassion fatigue when they had a strong sense of empathy toward their patients. In these cases, the feeling of empathy extended beyond professional sympathy, reflecting a feeling as if they were a family member. Although compassion is perceived to be stimulated by the provider's personal qualities, such as love and kindness (Sinclair et al., 2018), a compassionate response toward patients could be compounded by how the oncologist's personal situation, including factors such as age, gender, and family values, are related to those of the patient. When these doctors felt more empathy and personally related to the situation and context of their patients, they felt more compassion fatigue. Empathy is considered a crucial trait in providing quality patient-centered care. Compassionate care, which includes core elements of relationships based on empathy and emotional support, improves health outcomes (Hojat et al., 2011) and patients' care experiences and thus, should be prioritized among cancer health care providers (Lown et al., 2011). However, as shown by Hunt et al. (2019), more empathy seems to be related to higher secondary traumatic stress. Similarly, in this study, when oncologists experienced empathy to the extent of being overwhelmed by compassion fatigue and over-identifying with the patient, they expressed immeasurable distress. When professional compassion becomes compassion fatigue, their professional quality of life can be seriously compromised, which could then negatively impact the patient's quality of life.
The results of our study highlight the emotional distress experienced by the oncologists over difficult clinical situations that are inherent to the clinical nature of oncology work. This is a finding also reported in other studies  and has been shown to be one of the key predictors of compassion fatigue (Gribben et al., 2019). In this study, oncologists often faced life-threatening situations and reported feeling responsible for patient outcomes and even death. When oncologists experience the deaths of long-term patients or unexpected deaths, they often struggle more in life . Also, oncologists working with pediatric cancer patients experience significant secondary trauma (Bowden et al., 2015). Oncologists become heavily affected, because they cannot meet their ultimate goal of compassion-namely, ameliorating the suffering of their vulnerable patients (Sinclair et al., 2018). In addition, patient deaths may result in consequences for the physician, including personal consequences such as feeling disconnected from family members and friends and becoming desensitized toward death, as well as professional consequences such as concern about burnout or compassion fatigue at work (Granek et al., 2015).
We also explored how negative interactions with patients and families could influence oncologists' compassion fatigue. When oncologists were not trusted by patients and families, they reported more fatigue. Challenging relationships with patients and families may lead to oncologists bearing the blame of and feeling more guilt for negative health outcomes, which can increase their compassion fatigue. Emotional fatigue among physicians often results from conflicts with patients and families, including interactions with difficult patients (Schwenk & Romano, 1992). Oncologists in this study reported being deeply hurt by negative interactions with patients and families. These reports of experiencing personal emotions in adverse situations-whether a suffering patient or a difficult and unappreciative family member or colleaguemay suggest that common perceptions of physicians as detached or unemotional is outdated and inaccurate regarding the realities of physicians today.
Consistent with previous studies (Fernando & Consedine, 2014) the oncologists in our study responded that environmental and institutional factors such as work overload did interfere with them being able to fully express compassion toward patients and families. When doctors wanted to devote more compassion to their patients but could not, they felt guilt. Guilt can cause fatigue, creating a cycle in which guilt about having compassion fatigue may increase fatigue. The heavy workloads and limited work-life balance of physicians has been well documented in previous studies (Kaliannan et al., 2016;Leigh et al., 2011;), but subjective perception of time pressure at work was a key predictor of burnout and compassion fatigue among oncologists in Canada (Kleiner & Wallace, 2017). Future studies could examine their perceptions of compassion fatigue and self-blame or guilt related to not being compassionate.
The public expectation of the medical doctor as a cold and rational professional without much emotional attachment (Flores, 2002) may be outdated. Our study shows that doctors experience pain acutely and palpably because they are also human, as they repeatedly emphasized in interviews. Their effort to channel attitudes of detachment toward patients' painful situations while maintaining a professional level of concern for them (Halpern, 2001) can be viewed as coping strategies related to compassion fatigue and thus, appropriately providing a compassionate response toward patients. Emotional detachment was once considered the preferred way to remain professional (Merton, 1957). However, detachment and concern for patients are not necessarily dichotomous (Cadge & Hammonds, 2012), and as recommended more recently, emotional attunement is more desirable (Halpern, 2007). Instead of assuming oncologists must be different than other people and must accept without consequence the devastating situations occurring around them, we should focus on their distress and how to alleviate their suffering.

Limitations
Our study participants were recruited from three major hospitals in Korea. Although this sampling may limit the generalizability of our results, it was valuable to include oncologists with extensive experience working with cancer patients, and cancer care in Korea is concentrated in major hospitals. In addition, the oncologists in our study had a variety of expertise in oncology, such as radiation, surgical, and hematological oncology. The nature of their clinical work is different, so it may lead to different levels of interactions with patients and family and responsibility for life decisions, thus leading to different levels of compassion fatigue. Future studies could also examine if personal background factors (age, gender, years in the profession) have different effects on compassion fatigue.

Implications for Psychosocial Oncology
Various interventions for improving resilience and reducing compassion fatigue among oncology staff members have been delivered in oncology units (e.g., Fetter, 2012;Slater et al., 2018;Sullivan et al., 2019). Building on concepts of oncologists' compassion fatigue based on a transactional perspective, we suggest interventions that promote appropriate levels of compassion and reduce barriers that interfere with it. Although the emotional demands of oncology work are inevitable, our study findings indicate that interventions should target diverse factors at the micro (individual) and macro (institutional) levels. In one study, oncologists suggested institutional supports, such as validation of grief and provision of psychosocial support, as well as more vacation time . Similar to these recommendations , providing training and education for oncologists on balancing compassion toward patients and self-detachment for personal well-being is essential (Kearney et al., 2009). Improving physicians' communication skills with cancer patients, such as rendering a poor prognosis, for example, can improve the physicianpatient relationships and reduce compassion fatigue (Baile et al., 1997). Given the nature of oncology, namely managing life-and-death situations in daily work, career-long education in medical knowledge and skills as well as strategies to deal with the accompanying existential pain is necessary. Also, time and opportunities to decompress and express the impact of patients' death and difficult decisions through debriefing sessions in a formal oncology unit or informal support group, alongside professional help and care for psychosocial difficulties, are imperative. Particularly, these supports should be available when the events occur. Finally, providing regular vacations or even short breaks for oncologists could mitigate compassion fatigue.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.