Physicians’ Experiences of Do-Not-Resuscitate (DNR) Orders in Hematology and Oncology Care – A Qualitative Study

Objective: In oncology and hematology care, a patient can have such a poor prognosis that cardiopulmonary resuscitation (CPR) for cardiac arrest is not considered feasible. The responsible physician can then decide to apply a do-not-resuscitate (DNR) order, meaning that neither basic (heart compressions and ventilation) nor advanced (defibrillator or medicines) CPR should be performed. Previous research has found disagreement in relation to DNR decisions among physicians and nurses working in this field. The aim of the present study was to explore the experiences and perceptions of physicians with respect to DNR orders within hematology and oncology care, in order to better understand their specific roles in this decision-making process, with focus on the ethical aspects of the decisions. Methods: Individual interviews were performed with 16 physicians working in hematology and oncology departments at seven hospitals in Sweden. Data were analyzed using systematic text condensation (STC). Results: Two main parts of the data were discerned. The first described the physicians’ different roles in the team in relation to decisions on DNR, as decision maker, patient advocate and mediator for relatives and team member. In the second part the physicians shared their experiences of making ethically challenging decisions on DNR. The experiences ranged from feeling as though they were making themselves God, to regarding a DNR order as one regular medical decision among others. The physicians made decisions on CPR or DNR with the patient's best interests in mind, primarily on medical grounds. However, they also considered ethical aspects of the decision, weighing beneficence versus maleficence. Conclusion: The results indicate that DNR decisions in oncology and hematology care are not only taken on medical grounds, such as the prognosis of the disease and the expected survival after CPR, but also with reference to ethical values, such as patient autonomy and the quality of life after resuscitation.


Introduction
Within oncology and hematology care, patients can sometimes be considered to have such a poor prognosis that cardiopulmonary resuscitation (CPR) for cardiac arrest would not be feasible, as the patient would not survive it or would survive with poor function and quality of life. The responsible physician can then apply a do-notresuscitate (DNR) order, meaning that neither basic (heart compressions and ventilation) nor advanced (defibrillator or medicines) CPR should be performed. If there is no DNR order when a patient suffers a cardiac arrest, CPR must start within 60 seconds, according to Swedish guidelines [1]. In 2013, the Swedish Society of Medicine, the Swedish Society of Nursing and the Swedish CPR Council jointly published guidelines for CPR and DNR in hospitals [2]. These guidelines emphasize patient autonomy and consent, information to the family, and consultation with other physicians and the medical team. They also mention ethical considerations in relation to DNR decisions, for example risks and benefits for the patient in a DNR decision [2].
Previous studies have shown that the decision for DNR is often made late in the patient's care, sometimes on the day of death [3,4]. According to Sanderson and co-workers [5], many physicians are of the opinion that this decision should be made earlier. Uncertainties concerning the prognosis or the patient's desire for continued treatment have been reported as complicating issues in the decisionmaking process [6].
Studies have also reported divergent opinions on whether the patient should be informed of DNR orders [7,8]. Löfmark and Nilstun [9] have shown that physicians sometimes avoid discussing DNR with their patients. Various reasons for this have been reported. For example, the physician may lack self-confidence [10], does not want to remove hope from the patient, might be uncertain of the prognosis, have poor contact with the patient's family, or lack the necessary time [5].
DNR orders can be interpreted differently [11,12] and written policies are often not available in hospital wards [3,13,14]. They can also be documented in different ways [11]. According to Brown and co-workers inadequate documentation may lead to patients receiving unwanted CPR [15].
Previous studies have shown that nurses often consider DNR decisions to be taken too late in the patient's palliative care [16]. It has also been reported that for nurses to provide good care, clear and welldocumented DNR orders are needed [17].
DNR decisions involve obvious ethical dilemmas. On the one hand CPR could reverse premature death, but on the other it can prolong illness and suffering and consume big health care resources. Further, patient autonomy should be respected, but is not always easy to obtain in these situations.
DNR decisions are made frequently in the field of hematology and oncology, but studies on this are scarce. In our previous study among nurses in hematology and oncology wards, we found uncertainty of the procedure for DNR decisions and disagreements between physicians and nurses. We also found examples of ethical dilemmas, such as the wish for doing well to the patient (beneficence) and reduce suffering (non-maleficence) versus respecting patient autonomy [17].
Therefore, the aim of this study was to explore physicians' experiences and perceptions of DNR orders within the area of hematology and oncology care, in order to achieve a deeper understanding of their specific roles in the decision-making process, with focus on the ethical aspects of these decisions.

Participants
The study was conducted in the hematology and oncology departments at seven hospitals in Sweden. Between one and three physicians participated from each hospital. Data were collected during the autumn of 2012 and spring of 2013.
The heads of the departments gave permission for the study, and introduced it to the eligible physicians. Those physicians who were interested in participation then contacted the first author. After receiving written and oral information, the physicians agreed to participate by signing a consent form.
The participating physicians were eight men and eight women (n=16). Their median age was 46.5 (28-64) years, their median work experience as a physician was 21.5 (3-39) years, and their median work experience at the department was 11.5 years (10 months -35 years).
All informants were currently working with hematology and/or oncology patients. Some departments also included other specialties, such as surgery, medicine and palliative care. All participants were scheduled to work shifts in the emergency room.
The represented specialist training included hematology, oncology, internal medicine, surgery, geriatrics and palliative care. The characteristics of the participants are described in Table 1.

Interviews
Individual interviews lasting between 32 and 67 minutes were performed by the first author. All interviews took place at the participants' hospitals, in rooms close to the wards or in the physician's office. An interview guide with semi-structured questions was developed and used.
Probing questions were asked when needed for deeper understanding. The questions were based on a manual developed for a previous study [17]. Because new topics evolved during the first interviews, three questions were added to the final interview guide. The interview guide is presented in Table 2.

Main question Examples of supplemental probing questions
Can you describe a situation you experienced when a DNR order was made? What factors contributed to the decision?
Who participated in the decision?
Was it a difficult decision to make?
Have you experienced problems associated with DNR decisions?
How do you feel about including patients and/or relatives in the discussion about DNR?
Do you have a clear discussion with patients on shift of treatment from curative to palliative care?
Can you describe a situation when you informed a patient/relative about a DNR decision?
How should the information be given?
Is the information usually a surprise to the patient/relatives, or do they usually already know?
Is the information usually given too early, at the right time, or too late?

Analysis
The interviews were recorded and transcribed verbatim and analyzed using systematic text condensation (STC) according to Malterud [18], including the following steps: i) reading the material to obtain an overview and bracketing previous preconceptions; ii) identifying meaning units representing the informant's experiences of DNR and coding these units; iii) condensing the content of each code group, and iv) summarizing the content of each code group to generalize descriptions of experience with DNR. Quotes, consisting of meaning units, were identified to illustrate the content. The initial analysis was performed by the first author. The co-authors listened to random selections of the recorded interviews, and read selections of transcripts. The final version of the result was made as a consensus by all three authors.

Ethical Considerations
The research followed international guidelines for empirical research, as outlined in the Helsinki Declaration [19]. The heads of the relevant departments gave permission for the study. The informants were included after written informed consent. The information stressed that participation was voluntary, that the participants could terminate their participation at any time and that data would be handled confidentially. In accordance with national regulations, an ethical review was not required for this study [20,21].

Results
Two main parts of the collected data were discerned ( Figure 1). The first part describes the physicians' different roles in the team in relation to decisions on DNR, namely decision maker, patient advocate and mediator for relatives and team member. The second part describes experiences of ethically challenging decisions on DNR in oncology/ hematology departments, including transition to palliative care, differences between oncology and hematology, definitions of DNR and important values. The findings are further elaborated below and illustrated by relevant quotes from the physicians. The ellipsis [...] indicates excluded non-significant statements.

Part I: The physicians' different roles in relation to decisions on DNR
Decision maker: All physicians in this study mentioned that they were responsible for all medical decisions, including a DNR order. They had experience of making DNR decisions, sometimes on their own and sometimes in consultation with colleagues.
"Sometimes it's just me who is involved, sometimes it's the patient too, and sometimes relatives. Sometimes we decide without involving the patient, if the patient is not able to understand the information. Sometimes I involve my senior specialist [...] if I feel uncertain or want support for the decision" (Physician 9).
The initiative to discuss DNR and other treatment limitations sometimes came from the physicians, but often from nurses, who according to some informants had closer contact with the patients and therefore could see deterioration earlier.
"The question can often come from the nurses in the ward, because they have a feeling for the patient's condition, long before we have been bed-side" (Physician 3).
Some physicians did not describe DNR as a big decision, but saw it as a small part in a larger context, that is, to decide what would or would not be done in a situation of treatment limitation, which in itself was a difficult decision.
It will probably be as ethically difficult as you make it... If you make it with thoughts of "making yourself God" or deciding on people's lives and so on, then it can become very large and difficult. [...] I think we have really difficult ethical considerations throughout the day, but DNR is not the most difficult" (Physician 7).
However, some physicians said that decisions on DNR were difficult and that they felt like making themselves God, deciding over life and death.
"I think many of us decide [on] CPR… just because we think it's so hard to say no.
[…] Because it is like making yourself God. I [have to] decide whether this person should be resuscitated or not" (Physician 1).
The decision could be medically easy, but ethically difficult, especially for young patients and in complex family situations, according to the informants. They also thought it was difficult when patients had several diagnoses, such as chronic or congenital diseases, to take into account.
Concerning involving patients and/or relatives in the decisions, all physicians answered that they could be involved, but that providing of information was most important. They emphasized the importance of making it clear that, according to Swedish law, it is the physician who makes the decision on DNR.
"It's important that it is [made] clear to [...] patient and family that I make the decision. Because it would be very unfortunate if some relatives believed that they have decided on DNR. I think that would be a disaster. Because then, ... when the patient dies, they can get the idea that they have "killed" their relative, and that must never happen" (Physician 3).
Some physicians mentioned that it was important to realize that the decisions were made from the available information at the moment, and that things could be seen differently in retrospect. In those cases, a patient might receive CPR, although DNR might have been a better choice, had more information been available. Some informants also mentioned that it might be positive rather than frightening for relatives to be in the room when CPR was performed.
"It can be difficult but, on the other hand, relatives need to be involved in CPR. It can have a positive impact, in that they see that we have done everything (Physician 13).
Patient advocate and mediator for relatives: All physicians in the study described their contact with patients and relatives as varying depending on the situation. Some physicians mentioned that they had obligations firstly to the patient and that their loyalties were, and should be, there. Several informants mentioned that it was not always important to inform the patient about the DNR decision, as they might not perceive what was happening anyway. Therefore, information was sometimes considered more important for the relatives. However, in terms of level of care, i.e. concerning the transition from curative to palliative care, the patient's participation was important, according to the informants.
Some physicians mentioned the term DNR in front of the patient, but some never did, as they thought it would create worries or take away the patient's hope. Instead, they informed the patient about what they would do and focus on when cure was not possible. Those physicians felt that there was a co-understanding with the patient and the relatives that resuscitation would not be undertaken, although the DNR decision had not been communicated.
"I tell them what the planning of the treatment looks like and that we have no plans to move them to the intensive care unit (ICU) if they are getting worse, because we don't believe that they will benefit from that. That may be enough, and then just write DNR in the form" (Physician 12).
However, some physicians saw informing patients and relatives as important, and wanted it to be a calm moment with time for questions. Most participants considered themselves as being good at providing information about DNR, but some mentioned that it was difficult and sometimes burdensome. According to one informant, the department had a policy not to inform if it was not necessary, which was in accordance with the opinion of the informant.
Several physicians brought up the question of DNR in association with discussions about the shift of treatment from curative to palliative care. Sometimes that could take place early, when the physician knew from the beginning that the disease was incurable. None of the physicians thought that information on DNR decisions was given too early. Most thought it was given at the right time but some said information was sometimes given too late. Most participants thought information did not come as a surprise to the patient, but described how it could sometimes be difficult for relatives to understand and accept the decision.
Team member: All informants expressed the need to document DNR orders, but routines differed between departments. Some used paper records and some computer records. Difficulties in making sure everyone in the team was informed of the decision were expressed. For physicians working in palliative care units, it was particularly important to document CPR orders, so that the team would not omit CPR if it was needed.
"I can imagine that a nurse who starts her shift thinks it's important to know whether there is CPR or DNR on different patients; whether she should press the alarm or not" (Physician 15).
The physicians sometimes had the task of supervising nurses and less experienced colleagues. This was often done through discussions about patients and medical decisions.
"I think we teach each other, we educate each other, by having these discussions. We discuss with the doctors and with the nurses ... and that goes on all the time" (Physician 5).
Some participants also thought that information to patients and relatives should be given in collaboration with the team. Examples were also given of disagreements between physicians, as well as between wards or different specialists. In such cases the attending physician must make the final decision, according to the informants.
If anaesthesiologists were to decide, all of our patients would have a DNR decision. [...] Because, if you have a patient with a severe blood disease who suddenly needs intensive care, then they have read in their books that the chance for haematological patients at ICU is so small, that it's pointless" (Physician 3).
Disagreements could also occur when nurses thought that the patient should have a DNR order, but the physician did not. Some physicians mentioned that nurses could see things differently; they were closer to the patients and saw their conditions also in a nonmedical way. At the same time, physicians could have different knowledge about the disease prognosis and available treatment and therefore want to proceed with treatment and refuse DNR.
As a nurse you don't take responsibility for playing God and deciding that we should do nothing. Then it's a lot easier to say: 'These stupid doctors that just go on. ' [...]But as a physician you see patients at revisits and meet those who have actually survived. [...] The staff on the ward might forget that you had someone who was close to death and then actually was cured" (Physician 13).
In order to overcome disagreements between physicians and nurses, one physician requested more questions from the nurses.

Part II: Experiences of ethically challenging decisions about DNR
The transition to palliative care: An obvious dilemma for the informants was the transition from curative to palliative care. All informants emphasized that this decision required medical competence first and foremost, apart from ethical skills. Concerning patient autonomy, most departments had routines for some kind of discussion with the patient when the shift from curative to palliative care was made; this was mostly one discussion, but in some cases several.
Some physicians expressed a fear that nursing care might be poorer after a DNR decision. Others, on the other hand, mentioned that it might improve if clear information on the shift from curative to palliative care was provided. "I think it can be much better [...] if a decision on palliative care is taken, which often includes DNR and no intensive care. Then you can plan for the patient in a better way, establish that now it's like this, the patient understands it, relatives understand it ... we contact other caregivers ... plan for home care. So, I believe the nursing gets much better" (Physician 6).
Differences between oncology and hematology: According to the informants there were differences in how decisions were made between oncology and hematology departments, which was seen as ethically troubling. For oncology patients, although the physicians sometimes knew at an early stage that there was no possible cure, the patients might still have a long time left to live and could feel physically well during the palliative phase. Hematology patients, on the other hand, were often treated intensively for a long time. Although severely ill from the treatment, they could still be considered curable, with good chances of extended survival. For these patients, a shift to palliative care could occur at a very late stage, when all treatment possibilities were exhausted, and the end could then come quickly. Definitions of DNR: Another ethically challenging aspect was the definition of DNR. All physicians in the study agreed that the meaning of DNR was to not carry out compressions or ventilation, i.e. to refrain from basic CPR. However, some of the informants also mentioned refrain from intensive care and use of ventilator as part of a DNR decision. Some had experienced that the definitions differed between colleagues and were therefore careful in their information and documentation.
"For me it's just that we should not start ... chest compressions and should not take patients to the ICU when they have no circulation and are unconscious. So the palliative approach is not affected at all. But many mix that up" (Physician 3).
Some mentioned the possibility that a colleague or nurses on the ward might have different interpretations of DNR, and that the patient could therefore receive poorer care, or not receive the intended treatment, due to misunderstandings of DNR.
"Some believe that when we decide DNR, they should not get any help at all. I think that's very unfortunate" (Physician 3).

Important values:
The physicians also referred to important ethical values when describing DNR decisions. According to the informants, DNR decisions were based on medical grounds, such as the prognosis of the underlying disease and the expected survival after CPR. Age was also important, since high age could increase physical weakness. However, quality of life after resuscitation and weighing maleficence against beneficence were also mentioned as important aspects in the decision making. Most physicians mentioned that they made decisions on DNR based on what they thought was in the best interest of the patient.
"The goal is to do what I can ... what is good ... for the patient. If you have that perspective you're doing the best you can, and then you can't do more" (Physician 12).
Some also mentioned the importance of being able to ensure that the patient had a peaceful death.
"If someone is about to die, I want to assure them that they can go in peace, with no fuss. [...] We know that these people are about to die ... and since we know that, we need to be clear and tell them that when they go, they will go in peace, without someone pressing an alarm button and starting chest compressions" (Physician 15).

Discussion
All physicians in the study made decisions on DNR regularly. Their experience varied from feeling as though they were making themselves God to regarding DNR as one decision among others. The most ethically challenging decision, according to the informants, was not the DNR order in itself, but the transition from curative to palliative care.
Three different roles in relation to DNR decisions were found, namely decision-maker, patient advocate and mediator for relatives, and team member. As decision makers, the physicians made decisions about CPR or DNR with the patients' best interests in mind. They could include patients and families in the discussions, but were unanimous about letting them know that the decision was the physician's. That this is an important aspect has been reported in a previous study [22], and is also underlined in the Swedish national guidelines for palliative care [23].
The informants also gave ethical reasons for how decisions on DNR were made. If the potential outcome of CPR was unclear, the consequences were assumed to be better if CPR was initiated, especially for the relatives. Thereby, relatives could see that the medical team did all they could to try to save life and this might also help relatives in their grieving process.
As advocates for the patients and mediators for relatives, the physicians' decisions on whether to inform patients and/or families about the DNR decision were based on cautious judgments of what they thought patients and families could cope with. Some physicians included relatives in the discussion, whereas some spoke only with the patient. While offering information, most physicians chose to talk about what they would do to ease pain, for example, instead of mentioning the word DNR, which they thought could have negative associations. The physicians expressed no difficulties in informing patients or relatives, but described how they sometimes did not inform them for ethical reasons, such as if they thought the information would create worry and grief in the patient or if they thought it would take away the patient's hope for a good quality of life at the end of life.
There is an ongoing discussion in the literature about whether information on DNR decisions should be given [7,8], and it has been reported that physicians sometimes avoid talking to patients regarding these decisions [5,10,11]. The results of our study indicate that such decisions are based on ethical aspects concerning individual factors pertaining to the patient and that general rules for this would be difficult to establish. Such a position is also supported in the current Swedish guidelines for DNR [2].
As team members, the physicians saw themselves as responsible for providing information on DNR decisions to the team. They also acted as supervisors, for other physicians and for nurses. In this role, they were also involved in solving disagreements regarding DNR decisions. The differing roles of the health care team around severely ill patients have been described earlier [24].
Participants in this study confirmed that disagreements between nurses and physicians regarding DNR orders could occur, as has been reported in our previous study [17], but they also gave suggestions for how this could be avoided. For example, it was suggested that nurses should be present when patients or relatives were informed and that nurses should ask for ethical and medical grounds for DNR decisions.
The second part of the analysis concerned the physicians' experiences of ethically challenging decisions on DNR in hematology and oncology care. They indicated that there were differences in how end-of-life decisions were made in these settings, which could create ethical tensions. In oncology, the palliative phase could be extended, while in hematology the patient could be severely ill from the treatment but still be curable, and hence the palliative phase could be short. This could make it difficult to foresee the prognosis, and could be the reason why DNR orders were sometimes given late in the treatment of some patients. Uncertainty in prognosis has previously been reported as a reason for difficulties in deciding on end-of-life care [6].
Diverging understandings of the content of DNR orders was also found to be ethically troubling for the informants. They were aware of the importance of clear definitions and saw DNR as no heart compressions and no ventilation. Some of them also added not using the ICU and ventilator to the DNR order. There were, however, different understandings between some wards, and the physicians thus saw it as important to be thorough in their documentation of DNR or CPR. Differences in the interpretation of DNR between colleagues and wards have been reported earlier [12].

Methodological Considerations
Several methodological aspects were considered throughout the study. Participants from nine departments at seven hospitals in Sweden participated, and recruitment continued until saturation was achieved [25,26]. The interview guide was developed from a guide used in a former study with nurses, where it functioned well [17]. During the first interview, three questions were added in response to aspects brought up by the participants, and were included in the manual and used in the following interviews.
All participating physicians worked with oncology and/or hematology patients and had extensive experience of the subject matter of the interviews and were able to supply rich answers to the questions. However, there is a risk of selection bias, as physicians who choose to participate might already have given the subject a lot of thought.
The fact that the first author is a nurse with experience of hematology and oncology might have affected the interviews both negatively and positively. On the one hand, there is a risk that some things may have been taken for granted, with a subsequent risk that important aspects were missed. On the other hand, the interviewer was familiar with the environment and had no difficulties in understanding the participants' medical language. As a consequence, the interviewer tried to be aware of her preconceptions both during the interviews and in the analysis process [26]. The transcribed material was thoroughly analyzed by all three authors, who identified and formulated the parts of the result together. The co-authors are experienced in using various qualitative methods of analysis.

Conclusion
Physicians working in the field of hematology and oncology frequently make DNR decisions. The results indicate that these decisions are not only taken on medical grounds, such as the prognosis of the disease and the expected survival after CPR, but also with reference to ethical values, such as patient autonomy and the quality of life after resuscitation. Diverging opinions on the difficulty of the decision were found, from feeling as though making themselves God, to regarding DNR as one regular medical decision among others.