The primary goal of advance care planning is to promote and respect the autonomy of a patient when choices regarding life-sustaining treatments are required [2]. It is a generally accepted standard that end-of-life (EOL) decision-making should be based upon patients’ values, beliefs, and preferences [3]. A fundamental aspect of patient autonomy, which often has been overlooked, is respecting patients’ preferred decision-making styles [4]. The literature on cultural competency shows that patients’ preferred EOL decision-making styles can vary [5–8]. Race and ethnicity can affect patients’ decision-making style, as well as their beliefs and preferences, thus influencing EOL decision-making [9–13].

Advance care planning has been recognized as a complex process and not a one-time conversation or simple completion of a form. Eliciting and understanding a patient’s self-described decision-making style is crucial for clinicians who wish to assist patients in the difficult task of EOL decision-making. Figure 3.1 depicts a proposed typology of how patients approach decisions about EOL care and offers strategies to clinicians on how best to facilitate advance care planning with their patients [14]. It should be pointed out that the “avoiders” are the most heterogenous patient group with motives that could vary from “I don’t know/don’t care” to “I really don’t want to think/talk about this” to “Why should I think about this when it is only in God’s hands anyway.”

Shared decision–making is a collaborative process that allows patients, surrogates, and clinicians to make healthcare decisions together by taking into account the best scientific evidence available as well as patients’ values and preferences [15]. In this process, clinicians need to communicate complex medical information in lay language. This does not mean that physicians are purely sources of information only – fear of being perceived as paternalistic should not dissuade them from making a recommendation if appropriate. However, recommendations should be given upon invitation and in a sensitive way [16]. To improve these types of interactions, providers should seek out opportunities to learn or improve the communication skills needed to conduct difficult EOL care conversations [17].

Many efforts have been made to improve advance care planning. For example, the PREPARE study aims to better prepare patients and their surrogates in communicating goals and to make better informed medical decisions through use of a website with embedded, culturally inclusive video scenarios [18]. It gives patients in varying stages of behavioral change (knowledge, contemplation, self-efficacy, readiness) a resource regarding how to identify a proxy and how to elicit goals and values for care and discusses options for patients to consider such as giving leeway/flexibility for surrogate decisions or not. While formal completion of a living will is an outcome measure in this trial, the process of multiple communication events between patient and surrogate(s), patient and clinician, and surrogate(s) and clinician is the more important focus of the PREPARE intervention.

If a patient is decisionally impaired, surrogates will have to make decisions for others. The classic way to invite substituted judgment is to ask: “What do you think the patient would choose if he/she could make this decision?” This question is not always an easy one to answer, even if a detailed living will is available. Additionally, studies have shown that even when surrogates think they know what a patient would have wanted, patients would actually have chosen differently [19–25].

When surrogates feel unable to make substituted judgments because they have not previously had conversations with the patient about his/her wishes, the best interest standard should be used to guide decisions. The underlying ethical value for this approach is the patient’s welfare and beneficence. Acceptable clinical outcomes can be defined from the patient’s perspective (based on the ethical principle of respect for autonomy) or a clinical perspective (based on the ethical principle of beneficence). An autonomy-based acceptable outcome assumes an expected functional status resulting in a quality of life the patient is willing to accept [26]. A beneficence-based acceptable outcome usually entails that the patient retains at least some interactive capacity without overwhelming disease-related or iatrogenic morbidity, disability, pain, or suffering [27]. Some have suggested using a “best judgment standard,” combining both imputed substituted judgment and best interest, based on love for the patient [28].

Surrogates do not operate in a vacuum but struggle with cognitive, emotional, spiritual, and moral issues. They have different levels of health literacy, and often just understanding the complexity of medical details in a patient’s care is extremely challenging. Frequent, clear communication from clinicians in lay language is important but difficult when multiple consultants focusing on specific organ systems, as opposed to considering the situation holistically, are involved in the patient’s care. Surrogates do not wish to be held responsible for a loved one’s death by other family members and need to balance preventing undue suffering with the desire to pursue any chance of recovery in addition to the need to preserve family well-being [29]. When surrogates have been informed of a poor prognosis, they do not always accept the information and can be overoptimistic. This does not necessarily stem from poor comprehension of medical facts but may be due to different beliefs, e.g., that the patient may have unique strengths, religious beliefs that the patient will recover, or the need to maintain hope to benefit the patient [30, 31].

Coping strategies that surrogates may use to assist in decision-making include recalling conversations with loved ones; sharing decision-making with other family members; cultural, spiritual, and religious practices; storytelling; and delaying or deferring decision-making which leads to life-sustaining treatment by default. Clinicians need to pay close attention to surrogates’ emotional and spiritual needs and may need to facilitate when there is conflict in the family about medical decision-making. This requires excellent communication skills that are not traditionally taught during medical training. However, studies have shown that they can be taught to clinicians successfully and providers should seek out these opportunities within their institutions and medical organizations [17, 32].

Patients who have not made or communicated their decisions clearly are not always aware of the burden placed on surrogate decision-makers, even though this psychosocial and moral burden can be substantial and reach measurable levels of depression and posttraumatic stress [33–35]. When surrogates are uncertain about a patient’s wishes, they can feel overwhelmed leading them to ask for “everything” to be done resulting in life-sustaining treatment by default even when such care might be futile [36]. If surrogates believe they are simply reporting a patient’s decisions, because they are confident that the patient already made all important decisions regarding their medical wishes, they feel much less burdened in their role [37]. Similarly, the presence of an advance directive or a DNR order has been associated with decreased odds of significantly learned helplessness, defined as decreased motivation, difficulty in determining causality, and depression [38].

Advance care planning can reduce the burden of surrogate decision-making, and clinicians have a duty to point out to patients that they can lessen the burden on their proxies and empower them by facilitating conversations about EOL care [39]. Once general values and goals have been identified, discussions can move on to particular life-sustaining interventions and how pursuing or not pursuing these will fit in with the patient’s goals. Clinicians should try to support the surrogates’ role and their interpretation of patients’ values as well as attempt to decrease their level of stress during the decision-making process [36].

Ethical conflicts about surrogates’ interpretations of patients’ wishes can occur [40]. Strategies to manage such conflicts can include suggesting a time–limited trial, i.e., an agreement to use certain medical treatments for a defined period of time to see if the patient improves or reaches a critical predefined clinical milestone [41]. This can also prepare the surrogate for a discussion of a possible shift to comfort care. Discussing do–not–escalate treatment orders is another strategy that can reduce distress and burden when withdrawing life-sustaining treatments is not acceptable to the surrogate [42]. Occasionally, involvement of an ethics consult will be required.

A multitude of studies have described racial/ethnic disparities in EOL care and EOL decision-making [43]. Clinicians should be aware of their nonverbal communication and try to consciously increase the amount of positive rapport-building nonverbal cues with African-American patients [44]. An advance care planning intervention in hemodialysis patients, SPIRIT (Sharing Patients’ Illness Representations to Increase Trust),was shown to be particularly effective in African-Americans and helped surrogates to strengthen relationships with the patient, helped them to feel prepared, and gave them a sense of peace during and after EOL decision-making [45].

Unfortunately, there are patients, often among the elderly, who for various reasons do not have a default surrogate decision-maker to fall back on. In the absence of any family or friends, they have been called “adult orphans” or “unbefriended elders.” In one study, 16 % of ICU patients and 3 % of nursing home residents had no available surrogate decision-maker [46]. Several strategies have been suggested to provide surrogate decision-making for these individuals including expanding the default surrogate list to include ethics committees, treating physicians, or professionally trained and certified health fiduciaries [47–51]. It is clear that seeking guardianship is unlikely to solve the problem when immediate treatment decisions have to be made. Additionally, guardians have no preexisting relationship to the patient, are financially compensated for their role, and often reluctant to limit life-sustaining treatment [48]. A patient preference predictor may be a possible tool to supplement the shared decision-making process – it predicts which treatment a patient would want based on which treatments similar patients would want in similar circumstances. Based on a survey of over 1100 patients, a clear majority (79 %) would like integrating the patient preference predictor into their care; thus, some have suggested its use for “unbefriended” patients [51]. The best approach for EOL decision-making in patients at risk for being “unbefriended” is prevention; clinicians should make special efforts in promoting advance care planning with such patients. A study of 262 homeless patients, a group at high risk for becoming “unbefriended,” has shown that such efforts can be effective: 38 % of them completed advance directives in the intervention group (versus 13 % in the self-guided group) [52].

Physician Orders for Life-Sustaining Treatment (POLST), also known as Medical Orders for Life-Sustaining Treatment (MOLST), started in Oregon in 1991 as “POLST Paradigm Initiative,” a task force of healthcare professionals and ethicists from the Center for Ethics in Health Care at Oregon Health & Science University (OHSU). The “paradigm” is supposed to involve a process of EOL care planning that culminates in the completion of a POLST form, a standardized, highly visible, vividly colored, portable, single-page medical order set. The POLST form is an order set that synthesizes discussions between a patient and/or his/her surrogate decision-maker and a clinician (a physician or, in many states, also a physician assistant or a nurse practitioner). This order set is legally valid across the different possible care settings a patient may encounter within a single state. It can be used by a variety of providers in multiple settings including paramedics, fire departments, police, hospitals, and nursing homes to provide the care a patient desires. POLST advocates suggest using the “surprise question” – “Would I be surprised if this person died in the next year?” If the answer is “No,” discussing a POLST form is advised. While POLST forms were developed for terminally ill or chronically debilitated frail elderly patients, in many states their use is not restricted to these groups. In 2004, OHSU’s Center for Ethics in Health Care convened a task force of representatives from participating states to spread the POLST paradigm nationally. This National POLST Paradigm Task Force (NPPTF) created standards for endorsement of individual states’ POLST programs; currently 19 states have NPPTF-endorsed state POLST Programs [53]. Three of these states are designated as having “mature” programs (Oregon, West Virginia, California) because POLST is used by 50 % or more of hospitals, nursing homes, and hospices in each region of the state. Figure 3.2 shows the status of POLST Paradigm Programs throughout the USA [53].

POLST differ from out-of-hospital DNR orders, which only apply when the patient is in cardiopulmonary arrest, as it can include orders not just about CPR but regarding hospitalization and feeding tube placement. POLST forms can be used by patients for advance care planning; however, they are substantially different from advance directives: (1) Whereas advance directives can only be completed by patients who possess decision-making capacity, POLST forms can be completed with surrogates alone; (2) POLST forms are legally binding orders, and advance directives – unless they are very specific and indicate that instructions should be followed “strictly” – are expressions of preferences for desired and undesired treatments in the future that often need interpretation by surrogates together with the clinical team before they get “translated” into medical orders, e.g., “full code” or “DNR” orders; and (3) POLST forms are orders that follow patients into all care settings as opposed to DNR orders that are valid only for a limited amount of time during an acute hospitalization. The fact that POLST are binding orders and transferable into other care settings is both their appeal and a target for criticism, especially since the informed consent process leading to POLST completion can be unclear. Additionally, not all states require signatures by the patient or the designated surrogate.

Several studies have shown that POLST forms are generally leading to treatments that match the orders on the form, though less research has explored if POLST orders are truly congruent with patients’ wishes [54–57]. The TRIAD (The Realistic Interpretation of Advance Directives) studies showed misunderstandings among prehospital providers’ interpretation of POLST, raising patient safety concerns regarding around avoiding under- and overtreatment [58]. The Centers for Medicaid and Medicare Services’ recent decision to compensate physician efforts to have end-of-life care discussions with their patients will likely spur future research into the content of such discussions and the quality of the informed consent process leading up to completion of advance directives, POLST, and other advance care planning tools [59].