How, When and Where to Discuss Do Not Resuscitate: A Prospective Study to Compare the Perceptions and Preferences of Patients, Caregivers, and Health Care Providers in a Multidisciplinary Lung Cancer Clinic

Background: Do Not Resuscitate (DNR) is a significant but challenging part of end-of-life discussions when dealing with incurable lung cancer patients. We have explored the perceptions and preferences of patients, their caregivers (CGs), and health care providers (HCPs) and the current practice and opinions on DNR discussions in a multidisciplinary lung cancer clinic. Materials and Methods: This is a prospective descriptive study with a mixed quantitative and qualitative methodology to capture perceptions of the participants. To obtain a rich description of participant responses to questionnaire items, we employed a ‘think aloud’ process that prompted participants to immediately verbalize their thoughts when responding to questionnaire items. We used content analysis and constant comparison techniques to identify, code and categorize primary themes in the captured data. Results: Ten patients with advanced-stage lung cancer; nine CGs from the lung clinic and ten HCPs from the Thoracic Disease Site Group (DSG) were enrolled in the study. Most patients had only a limited understanding of DNR. Most CGs had a fair to good understanding of DNR. Most HCPs perceived their patients to have understood DNR most of the time. When patients were interviewed, a theme of “anticipated discussion” about DNR was identified. Patients and CGs expressed having faith in the system and responsible physicians as to when to discuss DNR. HCPs embraced a clinician preference-based decision-making approach to engaging in DNR discussions. They desired more resources, more knowledge, more structure and more time to discuss DNR. Most HCPs felt that it would be worth conducting a prospective clinical trial to determine the best time to discuss DNR. Conclusions: This pilot study provides a unique mixed quantitative and qualitative understanding of the perceptions of patients with lung cancer and their CGs and HCPs regarding DNR discussion. Our findings will help further the development of evidence-based guidelines and a broad prospective study that would have important implications for policies and practices around DNR discussions in order to reduce the emotional pain of dying patients, their CGs and HCPs.


Introduction
Most patients with lung cancer are found to have incurable diseases shortly after their diagnosis. More than one million people in the world die of this disease every year [1]. While patients struggle to reconcile themselves with the reality of impending death, the families, nurses and oncologists involved in their care are faced with the challenge of finding the appropriate timing and environment to initiate End of Life (EOL) care discussions. Do Not Resuscitate (DNR) is the most significant and critical part of such discussions. One of the most commonly cited reports involving more than 9,000 patients with life-threatening diseases revealed that only 47% of physicians knew the DNR status of their patients and 46% of DNR orders were written within two days of death [2]. Given the current global awareness and cultural shift towards patient autonomy, patients desire to know their prognosis and be a part of the decision-making process: a caring and sensitive physician is an integral part of such communication [3][4][5]. When DNR is not addressed in advance, terminally ill hospitalized patients may undergo unwanted interventions [6].
Cancer Care Manitoba (CCMB), Canada is a tertiary care center with a well-established multidisciplinary Thoracic Disease Site Group (DSG) that sees approximately 800 newly-diagnosed lung cancer patients per year. Upon being assessed as incurable, patients are often enrolled in the palliative care program at CCMB. This is a comprehensive program intended to address the multidimensional needs of patients and their families. One of the most important pre-requisites for patients to enroll in this program is to have their DNR status determined and signed by an authorized cancer care provider which, in most cases, is either an oncologist, a physician designated by the oncologist, or a nurse practitioner with special training and experience in the field of oncology. Based on the evidence from an extensive literature review and our clinical experiences, this discussion can at times cause significant distress to the patient, their family and cancer care providers involved in this discussion [7]. Members of CCMB Thoracic DSG ,including oncologists, nurses, palliative care physicians and psychosocial counsellors, identified a need to explore the current practice and opinions around DNR discussion and future development of advance care planning guidelines, which forms the basis of this study.
We aim to present perceptions and preferences of patients with incurable lung cancer, their family caregivers (CGs), and their health care providers (HCP) involved in the care of lung cancer patients, followed by another paper where we present more in-depth analysis of perceptions and preferences of patients and their CGs in response to DNR discussions.

Materials And Methods
The study was approved by the Research Ethics Board University of Manitoba (approval #H2011:227). Informed consent was obtained for this study.

Patients and CGs
We aimed to recruit from the lung cancer clinic at CCMB, a purposive sample of ten patients and ten CGs. The eligibility criteria for patients included individuals who: received a pathological diagnosis of non-small cell and small cell carcinoma, had incurable disease, and were not to receive further curative therapy. In addition, patients may have either accepted or declined a DNR order following a discussion of their DNR status with their HCP. Patients were able to attend a CCMB clinic or be seen on a hospital ward to complete the interview and have a CG who was a willing study participant. The inclusion criteria for CGs included individuals who were: nominated by the patient as the primary adult individual who assisted them with his or her care, involved in the DNR status decision-making with the patient, and able to attend a CCMB clinic or be seen on a hospital ward to complete the interview.
Members of the Thoracic DSG identified potential patient participants who met the inclusion criteria. A designated research nurse then invited eligible patients and CGs to participate in the study. If the patient and the CG agreed to participate, then the research nurse scheduled separate face-to-face interviews with them at CCMB or on a hospital ward. Informed consent was obtained by the research nurse from all the participants prior to conducting the interviews.

HCPs
We recruited ten HCPs who met the eligibility criteria described as any physician, a nurse, or a psychosocial counselor who was a member of the CCMB Thoracic DSG. The designated research nurse invited eligible HCPs to participate in the study. Informed consent was obtained by the research nurse from all participants prior to conducting the interviews.

Design and administration of the study
We employed a mixed quantitative and qualitative methodology to capture perceptions of the participants. In order to adhere to a consistent interviewing technique and maintain the quality of audio-recorded interviews with study participants, we provided training on the interview technique to the research nurse who was designated to conduct all participant interviews.

Quantitative Component
Patients, CGs, and HCPs were asked to complete respective pen-and-paper versions of an investigatordeveloped questionnaire. Socio-demographic and medical information of patient participants were collected from patients and verified with CCMB medical records (see Appendices 1-3).

Qualitative Component
To obtain a rich description of participant responses to questionnaire items, we employed a 'think aloud' process where the research nurse prompted participants to immediately verbalize their thoughts when responding to questionnaire items on DNR status and DNR discussions. These prompts explored their level of understanding, personal wishes, triggers, timing, appropriate setting, emotional experiences, appropriate health care provider, and the extent of family dialogue on DNR decisions (see Appendices 1-3). The research nurse was guided by an open interview script with prompts. Interviews with participants were audiorecorded and transcribed for interpretation and analysis of the key themes. The in-depth responses provided by respondents precluded reporting all data in one manuscript. Hence, in this article, only portions of qualitative responses to interview questions are presented. The second manuscript will provide greater detailed information from the patient and family perspective based on their experiences, beliefs, and preferences.

Data analysis
Descriptive statistics were used to describe the demographic and medical characteristics and subgroup responses to the questionnaire items.

Qualitative Component
The audiotapes of participant 'think aloud' responses to questionnaire items and semi-structured interview questions were transcribed by a trained medical transcriptionist. The investigators then examined the transcripts and used content analysis and constant comparison techniques to identify, code and categorize primary themes in the data [8].

Themes and Coding
Three members of the investigative team (NA, WH and ML) extracted possible themes from transcribed data from three groups of participants. All themes were subsequently discussed and consensus was obtained for the extracted themes. A coding template was developed for each of the themes and served the basis of ongoing analyses and interpretation of the remaining transcripts by the third author. Regular meetings occurred between the first and third authors to discuss the addition or revision of themes as data analyses and interpretation ensued.

Participant characteristics
Ten patients with advanced stage lung cancer, nine CGs and ten HCPs from the Thoracic DSG were enrolled in the study. Due to logistical reasons, one CG could not be interviewed and was excluded from the analysis. Patients, CGs, and HCPs demographics are outlined in Tables 1-3. The majority of patients were male, 67.5 years of age (median), affiliated to some religion and had pathologically confirmed non-small cell lung cancer, an identified DNR status, no brain metastases, and had received no whole brain radiation prior to being interviewed. Most patient participants were interviewed in the outpatient clinic at CCMB ( Table 1). The majority of caregivers were female, retired, 67.5 years of age (median), the spouse who lived with the patient, varied in religious and ethnic backgrounds, reported having less than or equal to a high school education, and earned either $40,000 to 80,000 or less than $20,000 in annual income ( Table 2). The majority of HCPs were female and were either a nurse or a medical or radiation oncologist ( Table 3).

Religious Affiliation
Affiliated to Some Religion 9 Not Affiliated to Any Religion 1

Histological Diagnosis of Lung Cancer
Non-Small Cell 9 Small Cell 1

Received 2
Not Received 8

Outpatient Clinic 8
In Hospital 2

Frequency of Contact with Patient
Living with Patient 7 Visiting Weekly 2

Understanding of DNR
The following describes quantitative and qualitative responses by patients, CGs, and HCPs on how patients and CGs understand DNR status ( Table 4). Exemplars of study participants' qualitative responses provide some insight as to the rationale for responding as they did. Some responses were common to all three groups and some are exclusively from HCPs.  When patients, CGs, and HCPs were prompted to explain their responses to their respective questions noted in Table 4, the theme of an anticipated discussion was found. HCPs were also prompted to explain how they knew that their patients understood or did not understand DNR when discussed. Some HCPs explained that, due to their direct explanation and clarification of what to expect, they felt satisfied that their patients and CGs understood DNR. The following statement provided by one HCP demonstrated his or her belief in the key role that the HCP has to play in taking a direct approach by providing a clear explanation of outcomes of treatment and next logical plan of care: "Oh I directly ask them, I don't go around it… Sometimes I have spoken about it even before that." Another HCP similarity stated: "And I feel like I would assume that by the end of our conversation they would understand it because I would normally check out what they understand, right".
The above statement indicates how he or she valued clear, direct communication and validation of whether the patient and CG understood the implications of DNR. Still another HCP described an appreciation for gauging patients' understanding of DNR based on the questions they ask and their prompts for more information about DNR: "Because of the questions they ask for clarification."

Appropriate time to discuss DNR
Patients, CGs, and HCPs were asked 'when' was the most appropriate time to discuss DNR. Their quantitative responses are outlined in Table 5.
Appropriate time to discuss DNR Patients (n = 10) CG (n = 9) At time of initial diagnosis 3 2 On follow up visit after previously being told of poor prognosis 1 1 On hospital/palliative care admission 1 1 Only when death is imminent 1 Other 1 1 When discussing referral to palliative care program 1 1 When poor prognosis is first heard 2 3 Appropriate time to discuss DNR HCP (n = 10) Discussing referral to palliative care 2 On follow up visit after previously being told of poor prognosis 1 Other (at any time, patients may have opinions even when well) 1 Other (depends) 1 When first told of poor prognosis 5 To further explain their responses to the timing of having a DNR discussion when prompted during interview, the following theme of "Having faith in the doctor as to when to discuss DNR" was captured. Most patients expressed 'faith' in their doctor's discretion. Essentially, patients described that the physician was responsible to identify the appropriate time and opportunity to initiate the DNR discussion with the patient. One patient stated: "I think once they say or Dr. X [oncologist] says "I can't do anymore, we're going to refer you on to a Palliative Care doctor and nurse and so on," that sort of tells me that you know the DNR should be discussed now." Similarly, most CGs described having a faith in the health care system and the physician's discretion to engage in a timely discussion about DNR. For example, one CG stated, "Well because again, we put our faith in the health system." Having this expressed faith in the health care system appeared to be based on the skilled knowledge of HCPs that, however, was not directly elucidated by patients and CG participants.
The requisite skills and knowledge were described by one HCP who shared his or her experience with challenges and sensitivities related to the need to respect patients' hope and psychological status, which guided the timing of a DNR discussion: When HCPs were asked if they felt comfortable discussing DNR, all ten participants were affirmative in feeling at ease in leading a discussion with patients about DNR. Based on HCPs' responses on 'timing,' it appeared that individual HCPs embraced a clinician preference-based decision-making approach, thus making it more comfortable for them to engage in DNR discussions with their patients. (I.e., either basing the DNR discussion on signals they identified from the patient's unique situation or preferences, unfolding treatment phases that moved from curative to palliative treatment or the patient's explicit wishes.) In the following, we exclusively present responses by HCPs when asked about aspects of DNR practices, policy, and the workplace or environment that required improvement and further study under the major theme of "Revisiting current practice." Seven out of 10 HCPs stated that they thought that current practice and the policy of DNR in their workplace required revisiting. To further explain their responses under "Revisiting current practice," several sub-themes identified specific elements that require further attention, including the need for "more resources," "more knowledge" and "more structure" in the clinic setting.

"More resources"
The sub-theme of more resources captures recommendations shared by some HCPs about workplace elements that are needed to bolster their ability and/or confidence to engage in DNR discussions. For instance, the need for more or protected time to engage in difficult DNR discussions was a key resource expressed by six out of 10 HCPS. As one HCP (oncologist) stated: Besides the need for more time, several HCPs also felt that a dedicated, private room was needed for confidential and safe expression of feelings by patients. Of note, all ten HCPs stated that the clinic was the best setting to discuss DNR. The hospital setting was perceived as the worst place due to the lack of privacy that does not foster sensitive interaction about DNR. Other HCPs explained that the patient's situation and other factors such as psychological status of the patient, also dictate the setting for optimal, timely dialogue and decision-making about DNR. Furthermore, all ten HCPs described that the physician, the patient, and single or multiple family members should be present during the DNR discussion to attain a shared understanding of decisions about DNR. Again, this provides additional rationale for HCPs' expressed need for more private, comfortable and dedicated spaces for uninterrupted DNR discussions in busy clinic settings.

"More structure"
The sub-theme of more structure addresses HCPs' expressions of an organized, evidence-based approach to guide difficult DNR discussions. For instance, six out of 10 HCPs felt that a script would be useful as a template or a guide to engage in a DNR discussion with patients. One HCP stated, "But very often when you have a script and you see how it's put in there, then you adapt that and you can use your own words. So having a script might be helpful in the same way. Not that the practitioners are going to use those words necessarily but it gives them a bit of a template and they can substitute their own." Other HCPs described the need for more formalized involvement of other health care disciplines to meet the patients' multi-faceted needs (e.g., physical, emotional, spiritual, financial, etc.) when discussing DNR or in follow-up to address concerns with a social worker, palliative care physician or nurse practitioner: "More time and more personnel and nurse practitioners … it's not that we are trying to hide away from it, but I could initiate it, but sometimes I feel that we leave for the next patient; you don't know how they are doing … So if somebody could take that on and explain everything how that would be when they go home, they'll feel much better."

"More knowledge"
The sub-theme of more knowledge addresses HCPs' expressions of their lack of confidence in taking an appropriate, personalized approach to engage in DNR discussions. There was one HCP, however, who indicated that oncologists are sufficiently trained to engaging in difficult DNR discussions: "I mean, we all have trained as oncologists. If you need to teach us something, I'm talking on a physician's point of view, I shouldn't be here".
Still others indicated the need for more training and evidence-based information packages to engage in difficult discussions about breaking bad news. Being skilled and confident to 'personalize' the DNR discussion was important to this HCP: In addition, the patient's age, psychological impact of the prognosis and implication of DNR, rapid deterioration in the patient's condition, religion, and language barriers are factors described by HCPs as adding complexity to their DNR discussions with patients and families.
Six out of 10 HCPs felt that it would be worth conducting a prospective clinical trial to determine the best time to discuss DNR. One HCP explained: "I think one of the things that happens when people are doing research is regardless of the outcome of the research, in the process of the research being done, people are thinking about it." The above statement indicates that a trial may initiate reflection and dialogue about the issue or an appropriate time. On the other hand, four out of 10 HCPs did not feel a clinical trial was a worthwhile endeavor to answer questions about timing of DNR discussions. For instance, one HCP offered the following cautionary statement: "Medicine is an art and a science," suggesting that clinical trials are not a panacea to determining the right time to hold a DNR discussion.

Discussion
Deciding on the patient's DNR status is vital to the care of advanced stage lung cancer patients. The patient's DNR status influences how oncologists, family physicians, oncology nurses and social workers manage clinical and psychosocial aspects of a patient with an incurable cancer. Evidence-based recommendations concerning how to discuss DNR with patients and their families are lacking. The need to conduct this pilot study arose due to the lack of a well-defined policy to discuss DNR in an academic center where specialized HCPs provide comprehensive cancer care in an established Thoracic Disease Group. In this paper, the authors focused on describing preferences and perceptions of HCPs to discuss DNR.
A shared understanding of the concept of DNR by the patient, the CG, and their HCPs is crucial in the decision-making process. In our study, when patients were interviewed, they all expressed a limited degree of understanding of DNR. Most of the CGs had reported an adequate understanding of DNR. Almost all participating HCPs perceived their patients to understand DNR "most of the time." A cross-sectional survey involving five Canadian teaching hospitals with 477 patients and 160 CGs revealed that most patients were not aware of the nature of the process of the Cardio Pulmonary resuscitation (CPR). Subset analysis indicated that cancer patients were more likely "unaware" (53.6%) of the key components of CPR as compared to patients with other medical conditions (38.7%). Only a very small number of patients (< 3%) had an adequate knowledge of CPR outcomes [9]. On the contrary, in another Canadian study published in 2012 that involved 429 patients (none were diagnosed with cancer) who presented to their primary care physicians, 84% of 386 respondents were aware of the term DNR. When these respondents were evaluated further, 83% of the 'DNR aware' patients correctly identified the definition of DNR [10]. On the other hand, this study's quantitative and qualitative responses indicated that most of the patients with lung cancer lacked a full understanding of the DNR concept. It is difficult to understand why patients lacked a greater understanding of DNR without further prompting which is a limitation of our study.
Interestingly, HCPs perceived their patients to have understood DNR most of the time.
Studies indicate oncologists have communication difficulties and hold misperceptions and misunderstandings when discussing EOL issues with their patients [7,[11][12][13][14]. Baile stated that, "This is a mistaken belief that patient communication is an innate skill of minor importance when compared with the technical aspects of care [11]." Further research is needed to more fully examine the rationale for discrepancies between the physician and patient/CG level of understanding of DNR.
Choosing the appropriate time and environment to discuss EOL issues, specifically about DNR, is a challenging and complex task. In this study, the quantitative responses of patients and CGs were diverse ( Table 5). However, both patients and CG expressed their faith in the health care system in making this determination. Both groups felt that the physician is responsible for choosing the appropriate time or opportunity to initiate the DNR discussion. This is in concordance with published literature, as most patients believe physicians should be the one to initiate such discussions [15][16]. When HCPs were asked the same question, they offered varied responses: "when first told of poor prognosis," and "at the time of referral to palliative care," as the time points for DNR discussion. Further examination of qualitative responses, however, indicated that most of the HCPs described "before palliative care" as the best time to discuss DNR. A study from Australia with a similar group of participants (19 palliative care patients, 24 CGs, and 22 HPs) found the following themes when asked about who and when to initiate discussions about prognosis and end of life issues: "wait for the patient or caregiver to raise the topic," "health professionals (HPs) to offer all palliative care patients and their caregivers the opportunity to discuss the future," "HPs to initiate the discussion when the patient/family needs to know," and "HPs to initiate the discussion when the patient/family seem ready [17]." These themes are similar to the themes expressed by the patients and CGs in our study.
Regarding the need for a script or a structured information sheet to discuss DNR, HCPs provided conflicting responses. Some HCPs, specifically oncologists, felt that oncologists received sufficient training to engage in DNR discussions without a structured script. Interestingly, other studies have shown that although physicians may be sufficiently trained to manage medical issues, they are least trained to communicate with dying patients and their families [18][19]. In his paper, "A Physician's Guide to Talking About End-of-Life Care," Balaban narrated a typical scenario that involves a DNR discussion led by physicians and other authorized HCPs [20]. In this scenario, very sick patients are asked questions about Cardio Pulmonary Resuscitation (CPR) in the most mechanical fashion. Patients are asked if their heart stops, do they wish to be revived with electrical shocks, chest compression, intubation, and connection to a ventilator? Such a script appears to be inadequate in many respects. EOL discussions should not be limited to DNR. HCPs need to address the psychosocial aspects of the dying patient and his family. Balaban provided a comprehensive four-step guide for physicians to discuss EOL care [20]. A review of the literature indicated that, in most situations, physicians describe the mechanics of CPR to patients with terminal cancer without discussing the overall goals of care. Physicians act on an urgent need to discuss and document the DNR status without discussing issues that are significant to the dying patient [21][22]. These issues include physical and psychosocial concerns, including the patient's fears of pain, indignity, uncertainty, and abandonment, which are a vital part of any EOL discussion [23][24][25]. von Gunten, through real-time clinical experiences, has pointed out serious gaps in the process of DNR discussions with cancer patients and provided a comprehensive multi-step approach to discussing DNR [26].
Finally, all HCP participants in our study indicated that the outpatient clinic is the most appropriate place to discuss DNR. Some HCPs considered the hospital as the worst place for such discussions. As evidenced by HCP responses in this study, there is also a desire to revisit the current institutional policy and practice to discuss DNR and possibly conduct a prospective trial to build a knowledge base that describes the most appropriate time, environment, or resources, and structure of DNR discussions.

Conclusions
Our pilot study provides a unique mixed quantitative and qualitative understanding of the perceptions of patients with lung cancer alongside their CGs and HCPs regarding DNR discussions. Our study findings lend evidence to the need for further development of evidence-based guidelines that address issues of concern raised by patients, their families and HCPs about current practices and policies regarding DNR.
The next steps in our research include a broad prospective study including patients with lung and other terminal cancers where EOL issues are frequently discussed and where the timing of DNR discussions could be evaluated empirically. Findings from such a large study would have important implications for policies and practices around DNR discussions that will benefit not only those who suffer emotional pain during the last days of their life, but also the team of dedicated oncologists, family physicians, palliative care physicians, nurses and psychosocial counselors who genuinely desire to facilitate comfort and a pain free death to the terminal cancer patients.