Easing the Burden of Surrogate Decision Making: The Role of a Do-Not-Escalate-Treatment Order

Introduction

We present a case illustrating the common problem of a surrogate decision maker who is psychologically distressed over the medical team's recommendation to withdraw life-sustaining treatment. We suggest how a do-not-escalate-treatment (DNET) order can be helpful in such situations when the usual approaches to withholding or withdrawing care are not acceptable to the surrogate. We define a DNET order, explain when it might be useful, and discuss how it can facilitate a humane, negotiated resolution of differences when the surrogate does not accept immediate withdrawal of life-sustaining treatments but the medical team views such ongoing support as inappropriate. The DNET order can allow surrogates to maintain some hope for successful treatment, yet may also eventually lead to orders for “comfort care only.” While we focus on surrogate decision making, a DNET order may also benefit patients facing similar dilemmas near the end of life.

Case Description

Ann was a 56-year-old woman with metastatic liposarcoma who underwent a thoracic debulking surgery. She had a long postoperative course complicated by intractable delirium. Later she became ventilator dependent and could not be weaned.

The patient had designated her husband, Jack, as her health care agent. But the two had not engaged in any discussion prior to her surgery about what she would consider an acceptable quality of life, her wishes about prolonged ventilation, or any other end-of-life preferences. Jack was aware of his responsibility to use substituted judgment to help the medical team understand his wife's goals and values. Jack believed that Ann wanted to fight for her life by continuing all treatments. Jack felt that the team's recommendations—comfort care only, Do Not Resuscitate (DNR), and terminal extubation—were morally unacceptable and would be equivalent to killing his wife.

Over time, the medical team worked to build a consensus with Jack by advancing his understanding of the medical situation, furthering his acceptance of the prognosis, and exploring what he understood about Ann's previously expressed values about end-of-life choices. The team also continued to gently recommend approaches for Ann's care that might ease the burden of decision making for Jack.

Before long, Jack appreciated that Ann's cancer was not curable and that she could never be extubated. He could state that he knew she was not going to improve. He worried that she was uncomfortable, but still felt very strongly that removing the respirator would be killing her. His beliefs were explored in depth with the palliative care team and chaplaincy. Still, Jack did not feel comfortable with terminal extubation. In place of terminal extubation, a DNET order was recommended.

Definition of a Do-Not-Escalate-Treatment Order

A DNET order is similar to any order to withhold treatment for a dying patient, but it does not specify withholding any specific treatment. Rather, the surrogate decision maker (or patient) and the medical team plan not to increase current life-sustaining treatments and not to start new ones. Some common examples of following DNET orders are not increasing or adding vasopressors when the blood pressure falls, not adding antibiotics in the face of an infection, or not intubating a patient for respiratory failure. Existing medical treatments such as vasopressors, antibiotics, or oxygen are continued at a fixed dose while ventilator or BiPap settings are typically not altered. Notably, no treatment is withdrawn.

As with similar orders for what is not to be done (e.g., DNR), DNET needs to be accompanied by a clear plan of what will be done, particularly to avoid suffering. Aggressive comfort-oriented treatments such as sedation and analgesia remain in the care plan, and may be adjusted appropriately. This caveat would extend to such interventions as changing ventilation settings to ease breathing. What constitutes “escalation” can vary from clinician to clinician^1,2 and with the course of an illness, so implementation of the DNET order requires ongoing monitoring of the treatment plan and may involve discussion with the surrogate. Is this therapeutic option an escalation of life-sustaining efforts or a needed comfort measure?

The DNET order is only proposed when the clinical team believes that the patient is going to die soon, regardless of medical intervention, and that additional life-sustaining treatments are medically nonbeneficial or even harmful. The order may be considered an acceptable alternative to immediately withdrawing life-sustaining treatment when the patient's or surrogate's personal, religious, or ethical beliefs mean such actions are unacceptable. ³ In our experience, DNET orders may facilitate a transition to comfort care over time. Because the patient is clinically decompensating, death typically occurs fairly soon after the implementation of a DNET order, although, compared to withdrawing life-sustaining treatment, the DNET order may prolong the patient's final hours and days.

In the only observational study we could find specifically about DNET, the authors defined “no escalation of care” (NEOC) orders, which were reported retrospectively in 30% of deaths in a tertiary academic medical center's medical intensive care unit—a remarkably frequent order. ⁴ NEOC typically involved withholding of hemodialysis, vasopressors, and blood transfusions—the same interventions that are usually stopped first in sequential withdrawal of life-sustaining treatments ⁵ —while other interventions, such as mechanical ventilation, feeding, fluids, and antibiotics, were usually continued. The time from NEOC to death averaged 0.8 days (range, 0–5 d). The authors believe that guidelines are needed for the use of NEOC. We prefer the term DNET, since “care” should never be withheld.

The DNET order has similarities to a treatment trial, such as a time-limited trial in which intensive, life-sustaining treatments are pursued but then reconsidered after a defined time period, and based on prespecified clinical milestones. In one study of ICU family conferences, treatment trials were presented as an option in 15% of meetings. These conferences occurred, on average, 10 days after ICU admission and the time frame for reevaluation ranged from a few days to two weeks. ⁶ In contrast, with a DNET order, clinicians and family members are sheltered from the process of regularly reevaluating and renegotiating treatment goals when the patient deteriorates, since not escalating treatment is explicitly addressed initially and much of the decision making has been done.

Why a DNET Order?

Even when expert care is delivered in a shared decision making model in which clinicians make recommendations based on the patient's goals and values, surrogates are burdened by decisions to withdraw medical treatments. Patients and surrogates struggle with ambivalence about end-of-life choices, often wanting both to avoid inappropriate prolongation of life ⁷ and to extend life as long as possible. Surrogates, in particular, worry about making mistakes and actually being responsible for the death, ⁸ and many feel guilty over their decisions. ⁹ Withdrawal of life-sustaining support may violate their cultural or religious norms, while withholding is acceptable. American bioethicists generally assert that withholding and withdrawing life-sustaining treatments near the end of life are acceptable and ethically equivalent options, ¹⁰ but surrogates may have the psychological intuition that one or both are morally wrong.^11–13

By offering patients and families an alternative to immediate withdrawal of treatment, a DNET order may relieve some of the stress and burden of decision making by diffusing decision making responsibility. The guiding principle of no escalation is established by the surrogate (or patient), but the physician typically holds the knowledge of what treatments are to be foregone. Families are protected from making detailed medical decisions about treatments that offer no medical benefit, and typically are not informed when later decisions are made to forego possible life-prolonging options for treatment. Compared to the decision to withdraw a specific treatment, a decision not to escalate a future treatment that might (or might not) be needed is somewhat vague. The balance of medical action (continued treatment) and nonaction (no new treatments) softens the impact of recognizing or causing the ultimate outcome, tempering the sense of personal responsibility for decisions. ¹⁴ Still, surrogates should also be given flexibility to modify a “blanket,” all-encompassing DNET order to accommodate personal, religious, or ethical beliefs that make certain options to withhold treatment unacceptable. ³

A DNET order may also decrease the stress and burden of immediate decision making by providing a more acceptable, gradual transition of the care plan that allows for both psychological adaptation over time ¹⁵ and even hope. The DNET order symbolizes formal acknowledgment by the family and medical team that survival is improbable; but because no treatment is withdrawn, patients and family members can continue to hope that the current regimen will succeed. The period of transition between the DNET order and death allows the medical team to work towards a consensus with family and it may promote a form of decision making that the surrogate can best tolerate.^16,17 A later move to more common procedures of withholding or withdrawing life-sustaining measures and to Comfort Care may occur. Indeed, in the medical intensive care study described above, 55% of NEOC patients had all life support withdrawn prior to death. ⁴

Finally, the extra time for the dying process may help families grieve. A DNET order provides an extended opportunity for patients and families to process the loss before the death. Families who experience a longer duration in life-support withdrawal report higher rates of satisfaction with care. ¹⁸ This extra time may also help improve bereavement outcomes. Families who identify that they have a difficult time accepting the illness have higher rates of complicated grief, ¹⁹ whereas a longer duration of caring may be protective. ²⁰

The Effect of a DNET Order on Medical Teams

Despite the hope that the surrogates' decisions reflect patient values, goals, and preferences, they make choices based not simply on substituted judgment but also on many other factors including their own wishes and interests, their cultural, religious, and/or spiritual beliefs, and their desire to promote family consensus.^21,22 The surrogate's personal difficulty coping with the removal of life support can also strongly influence the end-of-life decision making process. ²³ When a DNET order is thought to prolong the time spent on ultimately ineffective treatment, medical teams may worry that the surrogate's plan of care does not reflect what is best for the patient or that it inappropriately consumes scarce resources. In particular, medical teams report discomfort with the intensity of life support plans for ventilated patients when they believe it inappropriately prolongs the dying process. ²⁴ How can we deal with the tension between expecting the surrogate to adapt to medical realities and having the medical plan adapt to the surrogate's and, ideally, the patient's and family's values?

When confronted with conflict over how to transition to end-of-life care, the expert consensus is to respect surrogate interpretations of patient values and to take steps to decrease surrogate distress during the decision making process. ²⁵ Such recommendations recognize that patient preferences can change over time and that patients grant surrogates leeway in the decision making process.^26,27 Badgering the surrogate (or patient) to make a different decision only distresses them further, diminishes trust, and may increase decision making difficulties and lead to overt conflict. Some hospitals, including our own, allow physicians to set treatment limits independent of the surrogate's choices if those choices are nonbeneficial or harmful. ²⁸ Physicians are not obliged to offer resuscitation to dying patients ²⁹ or to provide or offer medical interventions they feel are inappropriate. In Ann's case, the team preferred a DNET order over a unilateral DNR and other treatment limitations, because this order allowed a consensus to be achieved between the team and surrogate, while providing some reassurance that dying would not be unduly prolonged.

Recent editorial responses to the NEOC article ⁴ discuss the appropriateness of DNET orders and explore the circumstances where DNET orders may not be the best choice.^1,3 Authors warn clinicians about implementation challenges and stress the need to carefully review any new orders, considering the benefits and burdens for the patient, noting that measures aimed at comfort may also prolong life. Authors also express concern about the potential overuse of DNET orders because they are easier or less time-consuming than negotiating a more explicit care plan, and emphasize that DNET orders should be used reluctantly rather than frequently, since other orders are available to the clinician that may not prolong dying. Referring to an article on “stuttered withdrawal,” ¹⁸ in which this approach improved family satisfaction, the authors suggest that “families need time to adjust to the realization that their loved one is dying.” ¹

Conclusion of Case

Although saddened, Jack consented to the DNET order. As a result, no new treatments were initiated, and ongoing treatments such as pressors were not increased. Comfort-oriented treatments were maximized. Ann was deeply sedated, and restraints and monitors were removed.

Once the DNET order was in place, Jack visibly relaxed. While he had been conspicuously distressed and largely absent from the bedside for the weeks leading up to this decision, he was now regularly able to spend time with Ann. The medical team, including the nursing staff, supported the DNET plan, tolerated the delay of death, and felt they were able to make Ann comfortable. She remained on the ventilator and died peacefully four days later.

Conclusions

The decision to withhold or withdraw treatment near the end of life can be stressful and burdensome to patients and surrogates. When distress is high or seems to be blocking what seem like medically appropriate choices, teams should recognize surrogate discomfort with proposed treatment, explore concerns, provide support, assure comfort, and consider recommending more acceptable options such as a DNET order. In our case, a DNET order reduced the burden of decision making for a surrogate and enabled the team to create a plan of care that was acceptable to all.