Sixteen physicians were recruited with ages ranging from 28 to 42 years, and all participants received interview one time. Nine were attending physicians, and their years as practicing physicians ranged from 3 to 14 years. The details of their demographic and work characteristics are presented in Table 1
. Four themes were identified, including (1) multiple roles of family members in discussions, (2) impact of the ward processes on EOL discussions, (3) process of discussion, and (4) professional ability and reflections. The major concerns included family members and the word processes. The participants developed a certain process of discussion form clinical experience, and they also had reflections about their own ability and roles in discussion.
3.1. Multiple Roles of Family Members in Discussions
The participants spent a large part of each interview discussing family members, who were a major concern in DNR and EOL care discussions. Family members had multiple roles, including information receivers, collaborative decision-makers, blockers, and the persons to be caring for patients. As information receivers, family members needed the information about the patients’ prognosis and future care plans. As collaborative decision-makers, family members were involved in the discussion and decision-making process. When the patients’ physical condition and consciousness status did not allow them to make decisions, family members had to take responsibility for decision-making.
As blockers, family members impeded physicians’ approaching the patients and did not allow them to discuss EOL matters with patients. Family members who were not primary caregivers might not understand a patients’ physical conditions, future prognosis, or they did not understand the patients’ suffering or painful experiences. They worried about the negative impacts on patients, such as loss of fighting spirit, negative emotions, and refusal of future treatment. Facing a relative’s death was difficult, and they could not accept the fact that patients were near death; they would ask physicians and the team to take any measures that might prolong patients’ lives. As the persons to be caring for patients, some family members had strong emotional responses, such as anger or anxiety, and avoided the discussion. Their emotional response frequently meant denial or an inability to accept patients’ death.
K: “One of the most important considerations is family members’ attitudes. They may not want us to tell the patient. They have a lot of worries, such as losing fighting spirit or having negative feelings. Even when some family members agree with the ideas of DNR and hospice care, they will make decisions when patients are in coma.”
H: “Hospice care addresses ‘care for the whole family,’ and family members who cannot accept patients’ death are our target of care. Their resistance or denial have psychological meaning.”
Meanwhile, family dynamics influenced the discussion and decision. There were different opinions amongst family members about the EOL decisions, and conflicts between family members was one of the barriers that physicians had to address. One physician said (E), “we have to be aware of the family dynamics because most of time there is not just one family member who will make the final decision.”
3.2. Impact of the Ward Processes on EOL Discussions
There were different approaches in different wards, including “time urgency” and “familiarity with patients and family members.” Time urgency meant the decision had to be made in a short time. In ER and ICU, physicians needed the decisions immediately to give treatment. One physician said (I): “we do not have time to wait, we cannot give family members much time to think. Just five minutes, five minutes, and then we must take action.”
In contrast, physicians on the hospice palliative care teams may visit patients several times to solve the barriers. A physician on the hospice ward said (H): “I usually start by understanding patients’ symptoms and their main concerns, and explore their opinions on DNR and hospice care. If needed, I can come a second or third time.”
Familiarity with patients and family members meant previous experience with patients and family members, so that physicians knew patients’ treatment process and understood patients and family members’ characteristics. Physicians in the oncology department usually had a long period with patients in which they built relationships of trust. For physicians in the ER and ICU, however, it might be the first time they had seen the patients, and they had to make decisions based on their current information and family members’ opinions.
N: “We have never met the patients before, and the first time is when patients’ physical condition is very bad. After physical examinations and reading the medical chart, we inform family members and ask them to make a decision. We have no idea about the patients or their illness experience.”
In addition, workload was an important consideration. A majority of the participants mentioned that they had to prioritize their work. Most of the time physical care and treatment came first. If they needed to spend a lot of time in DNR discussion, they accepted and followed the first decision that patients and family members had made.
N: “We might not have so much time, and only provide care based on their first decision, whether for DNR or to continue CPR.”
3.3. Process of Discussion
There was a process of discussion. First, physicians noted what was needed for the discussion and prepared the relevant physical information and data. Second, they explored patients’ and family members’ willingness to discuss EOL issues and decided an appropriate time. Regarding the timing of discussion, patients often did not consider DNR issues to be relevant or found it hard to think about when their physical condition is good; but they also found it difficult to accept when close to death. Physicians in the oncology or internal medicine departments tended to raise the issue when their physical condition had started to decline, which ensured that patients and family members had time to think about it.
M: “When people are healthy, they may not be aware of the necessity of this issue; however, when their physical condition is very bad, they may not have the ability or time to think and discuss it. I used to open the discussion when the disease had progressed to the advanced stage.”
Third, they delivered a variety of information. The information package included the diagnosis and prognosis, future disease trajectory and life expectancy, the process of dying and symptoms, possible treatment plans, the limitations of medical, hospice, and palliative care, CPR and life-sustaining treatment, DNR and its consequences, the effects and side effects of CPR, and the goals of care and care planning. CPR events included endotracheal intubation, cardiac massage, artificial nutrition, and hydration. The contents of information were based on patients’ and family members’ needs and physical conditions.
Fourth, physicians had to surmount the barriers to discussion. Physicians tried to establish relationships of trust with patients and family members as the foundation of dialogue. In rational terms, they gave precise and definite information such as the probability or rate of survival in dying patients after receiving CPR. However, they had to deal with patients and family members’ emotional responses and explore the real concerns or worries behind the emotions.
O: “It involves not only rational thinking but also emotional feeling. We had to empathize with their feelings and difficulties. This may take time and we need help from psychosocial care professionals.”
Because of the significant roles of family members in EOL decisions, family conferences were frequently used to solve the conflicts between patients and family members. In addition, they needed psychologists and social workers’ help to solve psychosocial and family dynamic issues. It was not easy to accept the nearness of death and make an appropriate decision, and they needed time to think about the issues.
K: “Some family members did not realize patients’ physical conditions, or there were differences of opinion between family members, and then I would call family conferences where I could tell all the family members at the same time and give them a chance to discuss it.”
In the end, physicians in non-hospice wards followed the decision for DNR or continuing CPR. If patients and family members did not accept a DNR order, then they gave CPR at the patient’s death based on the law.
F: “After all, we should follow the law. If they did not accept DNR, then we had the responsibility to give CPR.”
3.4. Professional Ability and Reflections
Some of the participants had received communication training courses in school, but most of them stated that they learned communication skills from their clinical experience. They needed to think through the process of discussion, their good or bad points, and ways they could improve. Furthermore, the physicians expressed several reflections. First was the nature of a DNR and the best interests of patients. Physicians should not give futile medical treatments that only caused harm rather than cure to patients close to death. Patients had a right to make decisions for themselves. However, patients might not wish to face decisions related to death. One physician (B) stated: “Sometimes when we told patients that DNR is better for them, they could not accept it. When patients come to the hospital, what they want is to be cured. They did not think about death.”
The second issue was whether family members had the power to stop physicians telling the truth and discussing it with patients, and then they made a decision when patients lost consciousness.
A: “In general, we respect family members’ opinions; however, we also have duty by law to tell patients the truth, and ethically we should respect patients’ autonomy. It is definitely a barrier if family members do not want you to tell them. Do family members have a higher power than patients themselves?”
The third concerned the struggles between the law, medical professionals, and the best interests of patients. Several physicians stated that if patients did not agree to a DNR order, they had to perform CPR. However, as medical professionals, they knew it would not be successful or have any benefit, so they questioned whether they had to perform CPR procedures to the end.
I: “We knew CPR would be harmful and not succeed in saving the patient’s life, so why could we not just stop and let the patient die peacefully?”
Fourth was the balance between telling the truth clearly and keeping hope. Physicians would like to deliver clear information; however, that was difficult to do while still helping patients and family members keep hope, as well as to relieve the impact of negative emotions.
O: “We want patients to continue hoping, not for cure, but for the company, that we will always be with them. However, facing death can cause fear, anxiety, uncertainty, and other negative feelings, which makes telling them difficult.”
This study used a qualitative methodology to explore physicians’ experiences in DNR and EOL discussions. Physicians considered family members to have multiple roles in discussions. Discussion was a dynamic process, which time urgency and relationships with patients and family members influenced their discussion. Physicians self-reflected about their abilities and the meaning of discussions.
The discussion was usually held at a time of transition, and physicians delivered information on the limitations of curative treatment and offered a new direction of treatment [17
]. Similar to previous study [18
], the results of this study showed that family members were generally involved in EOL discussions and had a lot of impact on decision-making. In a family-oriented culture, family members may ask physicians to tell them the diagnosis or prognosis before the patients, and let the family decide whether to tell patients [19
]. Sometimes they may avoid telling patients the truth to protect them out of worries that the patient cannot handle the truth and would have an emotional reaction [21
]. However, these worries constituted a barrier for physicians [11
In a familial model of decision-making, physicians not only considered patients’ willingness but had to embed the decision in familial relationships [22
]. Some physicians told family members about imminent death and DNR/CPR options only when patients became terminal [23
]. Furthermore, physicians might need to inhibit family members’ influence on decision-making [24
] or strike a balance between family members’ wishes and the appropriateness of care [9
]. Family conferences were often used to improve communication and decision-making between patients and family members [14
ICU and ER had unique characteristics. Patients usually lack decision-making capacity due to their physical condition; family members bear high stress due to the complicated clinical situation, and physicians often do not have a prior relationship with patients [25
]. The specifics of the departments influenced physicians’ behaviors, such as how long they could wait, how many times to visit, whether they needed to get the decisions and make a response quickly, and whether they had a relationship of trust as a foundation to directly open the discussion. The relationships between healthcare professionals, patients, and family members were important as a foundation of discussion [26
As with clinical guidelines, there was a certain process of discussion [5
]. Physicians had to initiate the discussion when patients and family members seemed ready, explore what they already knew and what they wanted to know, and give them opportunities to discuss the future [26
]. Patients needed information on medical procedures, outcome probabilities, and the goals of care to make a decision [29
]. Sometimes physicians could not give clear information due to the uncertainty of prognosis [11
] and the survival rate of CPR [30
]. Just delivering information was insufficient [7
]; physicians had to be sensitive to and deal with emotional reactions, emphasizing what they could do [31
], as well as try to maintain the sense of hope [8
]. In the end, they took action based on the patients’ decisions. On the other hand, a trust relationship was as the foundation of discussion, and previous studies revealed physicians had to establish partnerships with patients and family members for discussing and decision-making [32
Physicians in different departments needed to cooperate. Physicians in hospice wards needed those in oncology to introduce them and open the discussion. However, previous studies showed that conflicts between professionals, such as priority of care, role overlap and blurring of responsibility, and lack of communication, could hinder collaboration [5
]. In addition, physicians needed teamwork for EOL communication with various kinds of healthcare professionals, such as nurses, psychologists, and social workers, who could be aware of patients’ physical or psychosocial needs and manage the difficulties in discussion [34
Similar to previous study, physicians had internal conflicts about roles and duties between what they had to do and what was best for patients [24
]. They had the sense of avoiding relevant discussions [34
] because of the nature of discussion and worry about harm on patients. On the other hand, they had positive growths, for example, learning skills, working as a team, gaining appreciation from patients, and self-growth through the discussion process [9
Regarding clinical implication, physicians have to develop communication skills and sensitivity [26
]. They can learn from experiences to develop their own discussion process. They also need educational trainings about how to establish trust relationships with patients and family members, solving the barriers, delivering information, and dealing with emotional reactions. The opinions of family members are one of the major concerns in discussion, and they can learn to conduct family conference. The hospital policy can provide support from a palliative care team and psychosocial professionals.
Some limitations should be acknowledged. First, all the participants were recruited from one medical center and shared the same policy of discussion. Second, these participants were willing to discuss DNR and EOL issues in their clinical work, and those who tended not to discuss them did not participate in the study. A future study can address the roles of family members in decision-making and the issue of educational training on specific barriers.