Who serves as the surrogate, and how that person should decide, is shaped by law. When a patient has not appointed a healthcare agent and no guardian has been appointed, most states have default surrogate (family consent) statutes that set a priority list of decision makers, typically beginning with a spouse, followed by adult children, parents, and siblings; some states also recognize domestic partners or close friends, while others do not [11,12]. Surrogates are expected to use substituted judgment, making the decision the patient would have made based on the patient's known wishes and values, and to rely on a best interest standard only when the patient's wishes cannot be determined [13,14]. In practice these standards often blend, and clinicians can help surrogates by asking what the patient would have wanted and valued, rather than what the surrogate wants for the patient [14]. Some patients have no advance directive and no available surrogate (so-called unrepresented or "unbefriended" patients). Approaches to decision making for these patients vary by state and institution and may include ethics committee review or court-appointed guardianship [15].

Behavioral health professionals play a valuable role in advance care planning, but they should remain within their scope of practice. Explaining the purpose of advance directives, helping patients clarify their values and priorities, and facilitating family communication are appropriate counseling functions. Advising a patient on how to execute legal documents or interpreting state law is legal advice and should be referred to an attorney or to the facility's designated staff. Requirements for a valid advance directive (e.g., witnesses, notarization, statutory forms) vary by state, and many states restrict who may serve as a witness. California, for example, prohibits the patient's healthcare provider, an employee of the healthcare provider, and operators or employees of certain care facilities from witnessing an advance health care directive [22]. Clinicians should know their state's requirements and their employer's policies before agreeing to witness a document. They should also decline to serve as a client's healthcare agent, which would create a conflict of interest and a multiple relationship [23,24]. Finally, POLST/MOLST forms are medical orders that must be signed by a physician or other authorized practitioner; behavioral health professionals can support the goals-of-care conversation that informs these orders but cannot complete them [25].

Psychiatric advance directives (PADs) allow people with serious mental illness to document treatment preferences (e.g., preferred and refused medications, preferred facilities, crisis contacts) and to appoint an agent to make mental health treatment decisions during periods of psychiatric incapacity. Many states have PAD-specific statutes, and in others, mental health decisions may be addressed through a general healthcare power of attorney [26]. In a randomized trial, a facilitated PAD intervention improved the working alliance between patients and clinicians and increased the likelihood that patients received mental health services consistent with their preferences [27]. For a patient with serious mental illness who also has a life-limiting illness, clinicians can help ensure that the psychiatric and end-of-life directives are consistent with each other and that the appointed agents know about both documents.

The ACA Code of Ethics includes a standard devoted specifically to end-of-life care for terminally ill clients (Standard A.9) [28]. It directs counselors to help clients obtain high-quality end-of-life care, exercise the highest degree of self-determination possible, and receive an assessment of their ability to make competent decisions from a mental health professional experienced in end-of-life care. It also recognizes that counselors may choose whether to work with terminally ill clients who wish to explore end-of-life options (providing appropriate referrals if they do not) and gives counselors the option to maintain confidentiality when a terminally ill client is considering hastening his or her own death, depending on applicable laws and the specific circumstances and after seeking consultation or supervision [28].

Decision-making capacity is a clinical determination of whether a patient can make a specific decision at a specific time; competence is a legal determination made by a court [37]. Capacity is decision-specific (e.g., a patient may be able to choose a healthcare agent but unable to weigh complex treatment options) and may fluctuate with delirium, medication effects, pain, fatigue, or the progression of dementia. Clinicians typically assess four abilities [37]:

Patients near the end of life may say that they want to die or wish their death would come sooner. Such statements are not equivalent to suicidal ideation and should be explored rather than dismissed or reflexively treated as a psychiatric emergency. An international consensus definition describes the wish to hasten death as a reaction to suffering in which the patient can see no way out other than to accelerate his or her death; it may be expressed spontaneously or after being asked and should be distinguished from the acceptance of impending death or a wish to die naturally, although sooner [38]. Depression and hopelessness are strong, independent predictors of a desire for hastened death among terminally ill patients, and both are treatable [39]. A careful assessment explores what the statement means to the patient and contributing factors such as uncontrolled symptoms, depression, fear of being a burden, loss of autonomy, and existential distress, as well as any intent, plan, or access to means.

Clinicians should also understand which end-of-life options are lawful in their jurisdiction. Patients with decisional capacity have the right to refuse or withdraw life-sustaining treatment, and some patients choose voluntarily stopping eating and drinking (VSED) to hasten death [40]. A growing number of U.S. jurisdictions permit medical aid in dying (MAID), in which a terminally ill adult with decisional capacity may obtain a prescription for medication that he or she self-administers to end his or her life. Oregon's Death with Dignity Act, the model for most subsequent laws, requires the attending or consulting physician to refer the patient for counseling by a psychiatrist or licensed psychologist if either physician believes the patient may be suffering from a psychiatric or psychological disorder or depression causing impaired judgment, and it provides that actions taken in accordance with the Act do not constitute suicide or assisted suicide under the law [41]. A study of Oregon patients who requested aid in dying found that some met criteria for depression, raising concern about whether depression is consistently identified before a prescription is written [42]. Laws differ across jurisdictions regarding eligibility, waiting periods, which professionals may perform the mental health evaluation, and terminology. Behavioral health professionals should know whether MAID is lawful in their state and how their state's duty-to-protect provisions apply when a client is pursuing a lawful end-of-life option, a question that is unsettled in many jurisdictions and warrants consultation with a supervisor, ethics committee, or attorney. Clinicians may decline to participate in MAID-related services on grounds of conscience but should provide a timely referral [28].

Confidentiality obligations continue as a patient's capacity declines and after the patient dies. Under the HIPAA Privacy Rule, a person with legal authority to make healthcare decisions for a patient (e.g., a healthcare agent or guardian) is generally treated as the patient's personal representative with respect to relevant health information, and after death, the executor or administrator of the estate serves in that role [43]. Providers may share information relevant to a family member's or friend's involvement in the patient's care or payment for care, including after death, unless doing so is inconsistent with the patient's previously expressed preference. A deceased person's health information remains protected for 50 years after death, and psychotherapy notes generally require a specific authorization before they may be disclosed [43]. State law and professional codes may be more protective than HIPAA; the NASW Code, for example, directs social workers to protect the confidentiality of deceased clients [23]. While patients still have capacity, clinicians should ask what information may be shared with family members and surrogates and document the patient's preferences.

The clinician also can take advantage of available resources, including community or religious leaders, family members, and language translators [57]. It is important to note that using professional interpreters for patients with limited English proficiency will help ensure quality care. Convenience and cost lead many clinicians to use "ad hoc" interpreters (e.g., family members, friends, bilingual staff members) instead of professional interpreters. However, professional interpreters are preferred for several reasons. Several states have laws about who can interpret medical information for a patient, so healthcare professionals should check with their state's health officials about the use of ad hoc interpreters [58]. Even when allowed by law, the use of a patient's family member or friend as an interpreter should be avoided, as the patient may not be as forthcoming with information and the family member or friend may not remain objective [58]. Children should especially be avoided as interpreters, as their understanding of medical language is limited, and they may filter information to protect their parents or other adult family members [58]. Individuals with limited English language skills have actually indicated a preference for professional interpreters rather than family members [59].

LGBTQ+ older adults are more likely than their heterosexual peers to live alone and to rely on partners and chosen family for support, and these individuals may not be recognized under default surrogate statutes [11,68]. Encouraging formal appointment of a healthcare agent helps ensure that the patient's chosen decision maker has legal authority.