Meeting Physicians’ Responsibilities in Providing End-of-Life Care^* FREE TO VIEW

While prevention of premature death has always been a primary goal of medicine, provision of a comfortable and peaceful death has been widely acknowledged as an important end in itself only in the last several decades.¹ Even with widespread professional acknowledgment of the importance of palliative care, many patients die in moderate or severe pain,² physicians are often unaware of patients’ wishes regarding end-of-life care,³ and interventions are often inconsistent with patients’ preferences.² Providing high-quality end-of-life care is difficult and complex. Figure 1 shows the diverse skill set necessary for providing high-quality end-of-life care, as derived from the perspectives of patients with chronic and life-limiting illnesses, family members, physicians, and nurses.

In this review, we describe the responsibilities of pulmonary and critical care physicians in providing end-of-life care to patients and their families. Although many of these responsibilities are common to all physicians caring for patients with life-limiting illness, we focus on issues most relevant to pulmonary and critical care settings. We also focus on practical aspects of providing this care, such as prognostication and decision making about goals of care, approaches to communicating with patients and family, the importance of interdisciplinary collaboration and addressing conflicts, principles of withholding and withdrawing life sustaining measures, and the role of cultural competency in end-of-life care.

End-of-life care is frequently cited as an area of particular relevance to medical professionalism, both in concert with and in addition to medical ethics.⁴ Perhaps never more than in dealing with end-of-life issues will the physician’s culture, spiritual beliefs, and personal values influence decision making and therefore have the potential to compromise or test professionalism.⁵ The hospital setting, and particularly the ICU setting, compounds these challenges to physicians’ professionalism. The ICU setting, a strange, stressful, and often seemingly hostile environment to patients and families, increases vulnerability to the opinions, influences, or preferences of clinicians. Variable and conflicting approaches to care may be more pronounced and more common than in other health-care settings.⁶ In the context of these challenges, maintaining awareness of one’s personal values and simultaneously understanding and respecting the values of others becomes especially important. The key components of professionalism, including respect, integrity, honesty, and compassion, provide a foundation for the skills needed to provide competent and compassionate care at the end of life.

Approximately 20% of all deaths in the United States, or 540,000 deaths per year, occur in the ICU.⁸ The majority of ICU deaths involve decisions to withhold or withdraw life-supporting therapies,^{9101112131415161718} which require specific skills in end-of-life care. Outside of the ICU, pulmonary physicians and cardiologists also care for many patients with chronic and life-limiting diseases, such as COPD, pulmonary fibrosis, and heart failure. Some studies¹⁹²⁰ suggest that the quality of end-of-life care for patients with chronic lung or heart disease is poorer than for patients with cancer. For example, compared to patients with cancer, patients with COPD were more likely to die in the ICU, receiving mechanical ventilation, and with dyspnea.²¹²²²³ These differences occurred despite the treatment preference of most patients with COPD for comfort over prolonging life; in fact, one US study²¹ found that patients with cancer and patients with COPD were equally likely to prefer forgoing intubation and receiving cardiopulmonary resuscitation. A British study²³ also found that patients with COPD were much less likely to die at home and to receive palliative care services than patients with lung cancer. Health care for patients with chronic lung or heart disease is often initiated in response to acute exacerbations rather than being proactively based on a previously developed plan for managing disease.²²²⁴

In meeting the responsibilities of end-of-life care in the ICU or for patients with chronic lung or heart disease, several important challenges arise. ICU care frequently uses the most technologically advanced medical care to restore health and reverse injury or illness; transitioning to a palliative focus can be especially difficult in this setting.²⁵ The gravity and acuity of critical illness can lead to conflicts among family members and between patient, family members, and the medical team, which can complicate communication and decision making.²⁶²⁷ Navigating the transition from full life support to comfort measures only can be difficult; time is often short, and clinicians, patients, and family members have often not met each other prior to the critical illness. Challenges also exist for providing end-of-life care for outpatients with chronic lung and heart disease, such as determining when to consider and discuss end-of-life care, given the difficulty of accurately predicting prognosis for these patients.¹⁹²¹

Providing end-of-life care can exact a heavy emotional toll on clinicians, and this toll may be much greater in settings like the ICU that focus on prolonging life than in end-of-life settings like hospice.²⁸²⁹³⁰ Developing methods to cope with the stress of providing end-of-life care in a predominantly curative care setting is important for preventing physician burnout and may help ensure that patients and family members get the best care possible.³¹

Physicians have long recognized that there are circumstances in which patients cannot benefit from, or do not wish to endure the burdens of life-sustaining therapies. As medical experts and patient advocates, physicians have an important responsibility to make and share prognoses with patients and their families. In a study of ICU family conferences, some physicians do not discuss prognosis directly; and when they do, there is considerable variability in how this is done.³²³³ Physicians must work with patients and family members to reconcile the prognosis with a good-faith assessment of the patient’s goals of care and values regarding current and likely future quality of life. When physicians cannot determine goals and values from the patient, as is often the case in the ICU, they should work closely with the patient’s family in order to understand the patient’s values and preferences and to understand and advocate for the patient’s wishes. Finally, the physician should recommend a medical plan based on his/her best assessment of prognosis and a thorough consideration of goals of care, as established by the patient and their family. This means avoiding exerting unreasonable influence based on one’s own personal assessment of the patient’s likely quality of life after the current illness. Physicians ought not exert undue pressure on patients and families to accept a poor prognosis and the inevitable consequences of withholding or withdrawing life support in order to serve goals such as the physician’s or hospital’s financial or resource utilization goals.³⁴

Despite increasing agreement on ethical principles that guide end-of-life decision making in the ICU,³⁵³⁶ studies³⁷³⁸ show substantial variability in what clinicians consider appropriate goals of care, along with significant geographic variability in how this care is delivered. Much of this is due to variation in physicians’ attitudes and biases. Physicians need to be aware of their own biases, offer a range of best- to worst-case prognoses, acknowledge prognostic uncertainty, and recognize that others may not share their assessment of goals of care, especially patients and families from different cultures or socioeconomic status.³⁹⁴⁰

While patient autonomy has risen to be the foremost principle of contemporary Western medical ethics, the importance of autonomous choice to individual patients varies greatly, with some individuals preferring to delegate or share decision making with family members or clinicians.⁴¹⁴² In fact, some patients and family members are greatly relieved that clinicians are willing to take responsibility for decisions, such as withholding cardiopulmonary resuscitation, when it is not indicated.⁴³ Physicians must then determine where individual patients and families “fall” on this spectrum and adapt their communication accordingly. This requires that physicians ask questions of patients and family members about their preferred role in decision making in a way that allows patients and family members to defer some responsibility to physicians, if they wish, without making patients and, especially family members, feel that they are abdicating their own responsibilities.⁴³

Because a minority of critically ill patients can participate in discussions concerning their own end-of-life care, patients’ family members play a key role in most end-of-life care decisions in the ICU. These family members are likely to experience immense burdens, including symptoms of anxiety and depression⁴⁴⁴⁵ and financial burdens.⁴⁶ They are often unable to articulate their needs and concerns⁴⁷⁴⁸⁴⁹ and may be overwhelmed.⁵⁰ The physician has an important responsibility to prepare family members to make difficult decisions, helping them understand the patient’s prognosis, eliciting their sense of the patient’s values, and working in partnership to develop a sound and ethical treatment plan based on the patient’s prognosis, values, and preferences.⁵¹

Future quality of life is an extremely important subject of discussion regarding decision making. Interestingly, physicians are more likely to discuss prognosis in terms of quality of life than prognosis for survival during ICU family conferences about end-of-life care, even though there is much less outcome data concerning quality of life.³³ Most patients without severe illness say that if life-sustaining treatments would result in severely reduced quality of life, such as severe cognitive impairment or dependence on others, they would not want this care.⁵² Therefore, prognostication about future quality of life can have an important impact on decision making.⁵³ However, in studies⁵⁴⁵⁵ of patients’ current quality of life, there is generally no association between quality of life and treatment preferences. Physicians should therefore avoid making assumptions about patients’ treatment preferences based on current quality of life.

One important concept regarding quality of life and decision making about life-sustaining treatments is the concept of “states worse than death.” Many (although not all) patients can imagine a quality of life that would be less desirable than death, and discussions of such states can be a useful tool for identifying those health states for which life-sustaining therapies are not indicated.⁵⁶⁵⁷ Talking with patients prospectively in the outpatient setting to identify states worse than death can be useful for patients prospectively. These discussions may also be helpful for family members, who can refer to them when asked to make a substituted judgment for a patient in the critical care setting.

The concept of medical futility remains controversial, and there is no general professional consensus about the appropriate use of futility in medical decision making. Interventions are said to be futile if there is either no likelihood or an exceedingly low likelihood that the therapy will result in a medically beneficial outcome.⁵⁸⁵⁹ A number of authors^60616263 have argued that this principle is fundamentally flawed for a number of reasons, including the difficulties of prognostication and the risks of overriding patient autonomy. The American Medical Association recommends that when a patient or surrogate decision maker insists on a therapy that the physician believes is futile, a communication and negotiation process should be initiated to reconcile these differences, and the treatment should be provided until differences are reconciled.⁶⁴ More recently, the state of Texas enacted a law that provides for a process by which physicians can, over the objections of patients or surrogate decision makers, involve a committee to review and approve decisions to withhold or withdraw life-sustaining treatments that are judged to be futile.⁶⁵ However, this law has been challenged for several reasons, including the fact that the majority of patients for whom the process has been invoked were from racial and ethnic minority groups and underserved populations, raising questions about the fairness of such a law in our current health-care system.⁶⁶⁶⁷⁶⁸

In our view, invoking medical futility to withhold or withdraw life-sustaining treatments against the wishes of a patient or surrogate decision maker is ethically problematic and rarely necessary if clinician communication has been good and if building trust has been a focus of the clinical team. The desire to avoid the provision of nonbeneficial interventions embodied in the concept of medical futility, however, should be central to a framework of end-of-life decision making, provided benefit is defined with reference to the particular patient’s preferences and values.

One approach to such a framework is outlined in Figure 2 . First, the physician should elicit the patient’s preferences and values for therapies and outcomes from the patient or family. These preferences and values should be formulated into therapeutic goals. Once the goals are formulated, the physician should use published literature and clinical experience to determine whether treatment can achieve any of the goals. If so, the patient or family should be asked if the patient is willing to accept the burdens of the treatment. If, however, the primary goals are not achievable regardless of the treatments used, then treatments being provided or considered are no longer appropriate and should be withdrawn or not offered. The physician should review the basis for this determination to ensure that there are adequate data or clinical experience to support this determination, understanding that prior research has shown that some physicians make a determination of medical futility in settings where it does not apply and more commonly with patients from racial minorities.⁶⁹ Patients and/or their families should be informed of medical decisions to limit care made by clinicians. For some patients and family members who do not want an active role in medical decision making, it may be useful to approach this process as “informed assent” so that the patient or family member is not asked to make a decision that they would prefer not to make.⁴³ In most cases, patients or families will agree with a physician’s determination and will appreciate not being asked to choose to forego a treatment that is not indicated.⁷⁰ However, in some circumstances, patients and families will not agree with forgoing the treatment, we endorse a process of communication and negotiation as described by others.⁶⁴⁷¹

Communication successes and failures generate more gratitude and complaints than any other aspect of end-of-life care.²⁷⁷³ Patients with life-limiting illness and their families identify communication about end-of-life care as one of the most important skills for clinicians.⁷⁷⁴ Efforts to improve communication and support for family members in the ICU can reduce symptoms of anxiety, depression, and posttraumatic stress disorder among family members after the death of a critically ill loved one,⁴⁵ yet physicians frequently do not meet families’ communication needs.⁷⁵⁷⁶⁷⁷ The poor quality of formal training⁷⁸⁷⁹ and educational literature⁸⁰ has contributed to physicians not feeling competent to discuss end-of-life care.⁸¹ Physicians often are unaware of and unable to elicit patient and family concerns,⁸² and lack the skills and confidence to seek out and address family fears or the patient’s understanding of his/her illness and prognosis.⁸³ Several components of communication are associated with improved quality of end-of-life care and patient or family satisfaction with care and represent a good place for physicians to start improving physician communication skills (Table 1 ). In the ICU setting, interventions designed to improve communication within the team and with the family have been shown to improve end-of-life care.^45848586

While many experts⁹⁰⁹¹⁹² once believed that advance directives would allow patients to inform physicians about the care they want, advance directives have not lived up to these expectations.⁹³ A number of studies suggest that advance directives do not influence the treatments that patients receive⁹⁴⁹⁵⁹⁶ or change end-of-life decision making.⁹⁷⁹⁸⁹⁹ However, some studies¹⁰¹¹⁰¹ suggest that advance directives decrease levels of stress and increase ratings of satisfaction with care among family members of dying patients. Furthermore, when advance directives are included as one component of a more thorough process of communication about end-of-life care, often called advance-care planning, they permit patients, family, and physicians to think prospectively about the patients’ values and preferences. Advance-care planning, as a component of communication about end-of-life care that focuses on treatment preferences, may be especially useful for patients with progressive diseases, such as chronic lung disease, for whom specific decisions can be anticipated, such as whether to provide mechanical ventilation for acute respiratory failure. Therefore, advance directives should be discussed with all patients with moderate or severe chronic and progressive diseases as part of advance-care planning. Patients’ prior medical experiences, particularly with mechanical ventilation or with relatives or friends who have required life support, can facilitate patient/physician communication about treatment preferences and end-of-life care.¹⁰²

A common setting for discussions about palliative and end-of-life care in the ICU is the formal family conference.¹⁰³ As a setting that allows multiple family members and clinicians to participate in discussions and decisions, the family conference may improve quality of care, especially end-of-life care.^4386103 The recommendations that follow regarding the optimal conduct of such conferences are based on both empirical research and our own opinions.

The clinician leading the conference should ensure that appropriate members of the clinical staff are asked whether they should be present; these include the medical staff, nursing staff, primary care provider, social workers, and spiritual counselors involved with the patient or family. Palliative care consultants also provide an important perspective and skill set for these conferences and in caring for the patient and family after the conference.^84104105 A “preconference” meeting of clinicians can help develop team consensus and facilitate discussion of issues or conflicts that may occur within the team. During the family conference, clinicians should look for opportunities to support family members by eliciting patients’ preferences, answering family’s questions, acknowledging family emotions, clarifying the role of surrogate decision-making, expressing nonabandonment, ensuring patient comfort, and supporting family decisions.⁸⁸ A useful mnemonic for supporting the family is listed in Table 1.⁴⁵

Conflict occurs frequently in end-of-life decision making, especially in the ICU or acute care setting. Conflicts occur over therapeutic decisions as well as issues like communication styles, interpersonal interactions, and pain control.²⁷²⁸ In a prospective study²⁷ of ICU patients for whom withdrawal of life support was considered, conflict occurred between staff and family in 48% of cases, among staff in 48% of cases, and among family members in 24% of cases. Because of its prevalence, physicians have a responsibility to learn to identify and manage conflict within the medical team and with patients and their families. Conflict about decision making may also be constructive when it helps to identify differences in values or creates opportunities to resolve differences. Most conflicts and requests that appear on the surface to be “unreasonable” can be successfully resolved by well-conducted communication about goals, prognoses, and treatment options (including palliative care).¹⁰⁶¹⁰⁷ Early and formal communication within and between the team and the family oriented toward clarifying goals of care, prognosis, and principles of ethical decision making is a central component of successful interventions to improve end-of-life care in the ICU.^8586108109 Involvement of primary care physicians may be very helpful for understanding patient values and preferences and can also be an important tool for addressing conflicts.

Physicians also have a responsibility to work to address conflicts between themselves and other members of the clinical team. In the ICU and acute-care settings, nurses often support decisions to withhold or withdraw life support earlier than physicians.¹¹⁰ This can cause tremendous frustration for critical care nurses^111112113 and conflict between physicians and nurses,¹¹⁴ contributing to burnout and posttraumatic stress disorder symptoms.²⁸²⁹ Ensuring open lines of communication is essential. In observational studies,^115116117118 poor interdisciplinary communication and collaboration among ICU nurses and physicians is associated with increased patient mortality, length of stay, and readmission rates. Enhancing interdisciplinary collaboration is an important target for improving quality of end-of-life care.

Most patients who die in ICU and acute-care settings do so after a decision to limit life-sustaining treatments.^{9101112131415161718} Physicians have an important responsibility to improve the process by which life-sustaining treatments are withheld or withdrawn.³⁵ Although little empiric evidence is available to guide clinicians in the practical aspects of withdrawing life-sustaining treatments, ICU physicians should thoroughly understand the goals of withdrawing life-sustaining treatments (to remove all treatments no longer desired or indicated while ensuring patient comfort during the process) and should develop expertise in withdrawing life-sustaining treatments in order to minimize patient symptoms and support the family.¹¹⁹

Withdrawal of life-sustaining treatments is a clinical procedure; physicians must have the same preparation and expectation of its quality as for other procedures. The rationale for the decision to withdraw life support should be documented in the medical record. Several topics should be discussed with families, including explaining how interventions will be withdrawn, how the patient’s comfort will be ensured, the patient’s expected length of survival, and family or patient preferences about other aspects of end-of-life care.¹²⁰¹²¹ An explicit plan for performing the procedure and handling complications should be formulated. The patient should be in the appropriate setting with irrelevant monitoring removed; the process should be carefully documented, including reasons for increasing sedation or analgesia; and outcomes should be evaluated to improve the quality of future care.

Once a decision is made and a time is set to withdraw life-sustaining treatments, the course and timing of withdrawal should be determined by the potential for patient discomfort as treatment is stopped. Although time should be provided for family to say goodbye, the only rationale for tapering life-sustaining treatment is to allow time to meet the patient’s needs for symptom control. Vasopressors, antibiotics, nutrition, or most other critical care treatments can be discontinued immediately, without tapering. Mechanical ventilation is one of the few life-support treatments for which abrupt termination can lead to discomfort; consequently, physicians have a responsibility to develop an approach to terminal ventilator discontinuation that ensures patient comfort. A protocol that explicitly details an approach to withdrawal of life support in the ICU, including mechanical ventilation, has been associated with high ratings of clinician satisfaction and may help improve the quality of care, especially in settings where physicians are not familiar with withdrawal of life support or where there is significant practice variation.¹²² A sample protocol devised for withdrawal of mechanical ventilation is shown in Table 2 . Physicians also should inform families that, while death is expected after withdrawal of support, it may not be certain and the timing can vary.

Any protocol for withholding life-sustaining treatments should include an explicit protocol for sedation and analgesia during this procedure.¹²³ Such a protocol, carefully developed and implemented, has been associated with high levels of physician and nurse satisfaction, as well as with increased use of opiates and benzodiazepines for some patients without change in time from ventilator withdrawal to death.¹²² Furthermore, higher doses of opiates and benzodiazepines in the context of withdrawing mechanical ventilation has been shown to be associated with no change or an actual increase in time from withdrawal of mechanical ventilation to death, suggesting that these drugs can be used to provide for patient comfort without hastening death.¹²⁴

Family members are important as surrogate decision makers, but they also require care during the dying process as much as—and sometimes more than—the patients themselves. For many patients, support of their family is an extremely important physician skill concerning end-of-life care.⁷ Therefore, physicians have a responsibility to provide medical care that supports the family. Hospitals and especially ICUs often are not “family-friendly,” being designed to meet other priorities than family needs and comfort.¹²⁵ Families are often excluded from the bedside and ICU by visitation restrictions and by the preferences of clinicians during provision of care, including rounds and procedures.¹²⁶ Physicians must work to ensure that end-of-life care in the hospital and the ICU does not exclude family members and that care is comparable to what they would want for their own family member.¹²⁵

Physicians who practice in a multicultural environment have a responsibility to understand that family members from some cultures may have very different perspectives on the family’s role and on who should be involved in treatment decisions.³⁹¹²⁷ Individuals from some cultures may not endorse tenets of Western clinical ethics, such as the equivalence of withholding and withdrawing life support or the definition of brain death.¹²⁸ It is important to anticipate differences in perspectives and apply principles of culturally effective end-of-life care to these situations.¹²⁹¹³⁰

Several studies^{3738131132133} of patients’ attitudes toward end-of-life care identify values that vary by race, ethnicity, and geographic origin. On average, nonwhites are more likely to request life-sustaining therapy,^39134135136 and are less likely to have advanced directives or do-not-resuscitate orders^3137138 or accept hospice care.¹³⁹ Much of this difference is influenced by patients’ and families’ lack of trust in physicians and health-care institutions.^127134137140 Although understanding ethnic variations in preferences will not solve all problems with end-of-life care, a clearer understanding of what contributes to patients’ and families’ understanding, fears, and preferences improves communication and is a crucial step in providing better end-of-life care.

Physicians have a responsibility to avoid stereotyping patients or making assumptions about their attitudes based solely on race, ethnicity, religion, or other demographic characteristics.¹⁴¹ Existing recommendations can help physicians provide end-of-life care that accommodates needs based on religion and culture.¹²⁹¹⁴² The dying process is one of the most important events in which ritual aspects of religion and spirituality play a role. During discussions of end-of-life care, time should be spent discussing, understanding, and accommodating cultural and religious perspectives,¹³⁰ and reasonable efforts should be made to accommodate rituals associated with dying. Physicians should not assume that physician and hospital staff share the same values as patients and families of similar religious or ethnic background, and should not substitute statements from persons of similar background for a more thorough discussion with the patient and their family.

Working with patients’ and their families’ fears and distress at the end of life often places physicians in difficult situations. End-of-life care requires management of patient/family relationships, appropriate provision of hope, actively seeking to understand patients’ values and priorities under circumstances of loss and grief, negotiating goals in the setting of intense emotion and conflict, working amid uncertainties in prognosis, and managing emotional responses of staff to “difficult deaths.”¹⁴³ In addition, physicians may experience institutional or external social forces that influence them to pressure families to withdraw life support.³⁴ Burnout and posttraumatic stress disorder symptoms are especially common among critical care physicians and may interfere with job satisfaction, retention, and quality of care.²⁸ Multiple studies^{112144145146147} identify common stressors of caring for the dying and recommend management strategies to reduce their effects. Physicians should routinely perform emotional self-checks and may benefit from efforts to reduce personal emotional and stress burdens, such as restructuring the work environment to reduce stress, self-care activities like exercise and hobbies, and debriefing with colleagues, family and friends.¹⁴³

Increased burnout among critical care clinicians is associated with worse interdisciplinary relationships in the ICU.²⁸²⁹ Conversely, better nurse/physician communication is associated with enhanced professional relationships and learning for nurses and physicians, and decreased job stress for nurses.¹⁴⁸¹⁴⁹ Thus, enhanced interdisciplinary collaboration may be an important target for reducing burnout. The structure of the work environment, such as increased work hours and decreased days off, is associated with increased symptoms among clinicians.^2829150 Interest is growing in creating healthy work environments for ICU clinicians. The American Association of Critical-Care Nurses’ standards for establishing and sustaining healthy work environments advocate skilled communication, true collaboration, effective decision making, appropriate staffing, meaningful recognition, and authentic leadership.¹⁵¹

Physicians who care for dying patients and their families have a responsibility to provide the highest possible quality in end-of-life care. End-of-life care requires specific skills that traditionally have not been well taught in medical school or postgraduate training. More attention is needed to find effective methods to teach and evaluate these skills The needed skills are diverse and include cognitive skills like symptom management, affective skills such as providing emotional support, communication skills, and facilitating coordination and continuity of care. Physician factors, including physician’s culture, spiritual beliefs, and personal values, can influence end-of-life care and therefore can compromise or test a physician’s professionalism and diminish the quality of end-of-life care. Settings such as the ICU present significant challenges to high-quality of end-of-life care due to the focus on technologic life-sustaining therapies; physicians in these settings are also likely to care for patients with diagnoses that provide challenges, such as difficult prognostication and lack of access to palliative care. These challenges increase the importance of intensivist physicians acknowledging and meeting their responsibilities for end-of-life care. In the context of these challenges, maintaining awareness of one’s personal values and simultaneously understanding and respecting the values of others becomes especially important. The provision of a peaceful and dignified death permits us to fulfill our professional responsibility to patients and their families. When done well, it can be as satisfying as any other aspect of clinical practice.