(Posted April 7, 2020; updated April 15, 2020)

These guidelines set out some basic moral principles and examples of what a sound triage protocol might include, as well as what it should avoid, in outline format. They are not exhaustive and remain subject to revision. Please contact the author with any suggestions at jdicamillo@ncbcenter.org or 215-871-2001.

1. Triage protocols can be ethically appropriate when a genuine crisis situation arises, the demand for resources (e.g., space, staff, and equipment) surpasses availability, and other reasonable efforts to increase supply fail to meet the need. They must be built on a proper, principled moral framework.

a. Triage protocols should be temporary, activated only in crisis scenarios and deactivated as soon as feasible (as soon as the crisis situation has passed).

b. The principle of moral (prudential) certitude helps ground the ethics of triage: absolute certitude of outcomes is not needed. Rather, the best reasonable expectation about clinical outcomes (using best clinical data and treatments available in a reasonable timeframe) is morally sufficient. Extensive testing and lengthy or multiple trial periods for patients with unpromising or unclear prognoses—which may be required in noncrisis situations—may not be a moral obligation when a sound triage protocol is appropriately activated.

i. The ethical requirement is due diligence in light of the actual circumstances, which involve unusual limitations on time and resources. When more time and greater resources are available, moral certitude requirements would be increased.

c. The principle of stewardship of limited resources helps ground the ethics of triage: limited resource availability means that prudent and charitable use of resources is different in crisis and noncrisis situations. Patients who may otherwise be given more time and multiple attempts at a difficult recovery would now be receiving limited resources at the expense of other patients with an urgent, serious need.

d. The principles of solidarity and subsidiarity help ground the ethics of triage: there is a duty to help one another, to sacrifice for one another, and to address issues at the most localized level that can achieve the good of individuals and the common good. In a triage situation, escalating certain aspects of decisions about care delivery for input from a higher level (e.g., individual clinicians or care teams interfacing with a facility- or system-based triage team) may be necessary to safeguard the common good and offer appropriate care to all those in need.

e. The principle of respect for the dignity of human persons and the principle of the common good help ground the ethics of triage: there must be no net shift from a clinical health approach (i.e., individual driven) to a population health approach (i.e., public health driven). Rather, the dignity of the human person must continue to be foundational even as the role of the common good comes into sharper focus. It must be recalled that the common good cannot be achieved by disregarding the dignity of the individual, as utilitarian ethics do. The common good, properly understood, upholds the dignity of each individual.

2. No patient is to be abandoned. This should be made clear in policy goals and emphasized throughout. Many patients who might otherwise receive intensive care support will be unable to receive it in a true triage situation.

a. All patients who do not receive intensive care will be offered appropriate care for their needs, particularly palliative care or hospice.

b. All patients who do not receive intensive care will continue to receive basic human care, including food and water—even by medically assisted means when required—to the extent these (1) are available, (2) achieve their purpose (e.g., nourishment and hydration), (3) are tolerated, and (4) do not cause serious harms or complications.

c. Euthanasia and physician-assisted suicide remain immoral, and no patient will be influenced to pursue either of these actions.

3. Triage priority levels for the use of limited resources.

a. Clinical criteria should be the primary and most fundamental criteria for determining triage priority levels and allocating scarce resources.

i. They should be focused on short-term survival goals. That is, they should avoid considerations that reach beyond the immediate crisis situation. Examples of short-term criteria include

1. Short-term mortality risk despite the use of critical care resources, and

2. Short-term readmission risk (i.e., it is reasonable to give lower priority to those whose clinical situation would make them likely to be readmitted before the crisis situation has passed).

ii. Examples of clinical criteria for assessing short-term mortality risk could include the Sequential Organ Failure Assessment (SOFA) score, the Acute Physiologic Assessment and Chronic Health Evaluation (APACHE II) score, the Laboratory Acute Physiology Score (LAPS2), and others.

b. Long-term survival should not be a factor in triage priority. Long-term survival is often discussed in terms of life years saved or life stage considerations, which are either utilitarian or value-laden assessments that extend beyond the crisis situation that necessitates the triage protocol.

c. Nonclinical criteria should be secondary to clinical criteria. That is, they may be used only when clinical criteria for multiple patients are roughly equivalent.

i. Special categories of individuals may receive higher triage priority, based on considerations of justice or charity, which is not equivalent to “social value” or “utility.” The following are some possible examples.

1. Justice and the common good allow for protection of the system itself, including triage level prioritization for individuals whose health care delivery role (1) provides an active, essential contribution to ensuring the effectiveness or continued functioning of the care delivery system for the duration of the crisis and (2) may put them at higher risk of illness. These may include but are not limited to

a. physicians and physician assistants,

b. nurses and nurse practitioners,

c. inpatient pharmacists,

d. first responders,

e. medical equipment technicians, and

f. medical support staff.

2. Pregnant women with their unborn children can be prioritized over other patients in similar clinical situations because they are two patients at the same time, not just one, both of whom are particularly vulnerable.

3. Sole caretakers of minors or other dependents might be prioritized in light of the vulnerable persons in their direct care.

ii. “Tie breakers” may be used when clinical and top nonclinical criteria have been exhausted.

1. First come, first served. Health care workers should not remove a person who is already receiving care if there is a tiebreaker situation between this patient and someone else who is waiting. In this sense, the one who arrived first and began treatment is first come, first served.

2. Randomization (i.e., lottery) may be used to select patients who will receive available resources if all else is equal. This also could apply when choosing among possible new candidates for intensive care or ventilator use who all arrive within a similar time frame.

d. Periodic reassessment of triage priority levels is morally required, since clinical criteria can change rapidly in a triage situation. This could be every forty-eight hours, for example, but timing might vary on the basis of patient volume and staff resources.

e. Triage score calculations are a standard part of triage protocols, but they might not be necessary in all cases. For example, if urgency makes calculations for all patients impracticable, clinically determined exclusion criteria that reliably predict short-term mortality may suffice for determining which patients are not candidates for critical care resources.

4. Triage committees are not required but can be ethically sound in the context of a triage protocol. They can help ensure consistency of triage level assignment and application of triage protocols, supporting health care workers and patients. They may be more useful in large facilities with a greater triage need, where consistency would be more difficult to achieve.

a. Ethics personnel should be part of the triage committee.

b. There should be provisions for an appeals process for doctors, patients, surrogates, and families, who may be able to provide additional information or identify errors in the triage level assignment.

i. Clinical judgments by medical providers should be heavily weighted.

ii. Ethics personnel should be involved in any appeals process.

5. “Reallocation” of limited resources from current patients to incoming patients may be morally appropriate when using clinically grounded triage priority levels.

a. There should be no unilateral withdrawal of resources bypassing patient consent unless extreme circumstances warrant. (See informed consent below.)

b. From a moral standpoint, proportionality of treatment is what matters: does it offer a reasonable hope of benefit without imposing an excessive burden?

i. There is no moral weight to the distinction between withholding a treatment that has not been started and withdrawing a treatment that has already begun. The morally relevant consideration in both cases is whether the treatment in question is proportionate or disproportionate.

ii. In a triage situation, the burden to the community, which is part of assessments of proportionality in the Catholic moral tradition, takes on a greater weight. Continued use of resources for some patients showing no likelihood of recovery can compromise the ability to provide any care to other patients with a likelihood of survival in the short term. This is not usually a significant factor in noncrisis situations.

6. Informed consent is always preferred but not always necessary under a triage protocol if the clinical situation or emergency circumstances do not allow for it. Informed consent policies should be stated and mechanisms put in place in the protocol to obtain it.

a. Voluntary decisions by patients, families, or surrogates to decline disproportionate treatments or to place do-not-resuscitate (DNR) orders should be enabled and promoted wherever possible even when the triage priority may already indicate that the patient is not a candidate for initiation or continuation of certain types of treatment.

b. Physician decisions in the absence of patient informed consent, possibly in conjunction with input from a triage committee or officer, can be morally appropriate to move patients to appropriate care levels and place medical orders on the basis of well-informed moral and clinical judgments when obtaining fully informed consent is unreasonable (e.g., when the patient unable to communicate, it is not possible to contact the family or surrogate, there is not enough time, and so on). A mechanism should be established to reduce possible subjectivity and human error in cases where medical orders will be issued without consent or against expressed wishes (e.g., having a second physician review and sign off in these cases).

c. Communication is always a priority even when the patient’s care is determined without fully informed consent. Patients and families should not be summarily disregarded and simply ordered around, but informed about the realities of the situation and given the best understanding possible (under the circumstances) of what is happening to them and around them and why, even when they may not be able to choose differently.

i. Informational documents alerting patients, families, and surrogates to the reality that resources may not be available are advisable.

ii. Information on specific treatment outcomes for patients in a triage situation is morally appropriate, perhaps in document form and certainly in communication with the clinical team. For example

1. CPR may have lower likelihood of success. Time for medical staff to respond to a “code” may be increased because of limited and overworked staff, the need to change into appropriate personal protective equipment before initiating resuscitation, and so on.

2. CPR may be more likely to leave patients who survive in worse condition than it would during a noncrisis situation. Increased time to respond to a “code” may mean that, even if successful, patients will have suffered greater brain damage than typical before resuscitation.

d. Do not place blanket or universal DNR orders. Rather, (1) use override mechanisms as a last resort on the basis of judgments by physicians or input from triage officer based on clinical factors (e.g., triage priority level), or (2) implement DNR orders without consent only when it is unreasonable to obtain informed consent.

7. Alternative treatment paths should be proposed to patients, palliative care in particular, when their triage level (based on clinical factors) and available resources indicate that they are not candidates for intensive care or ventilator use.

a. Significant resources and efforts should be going not only to managing volume in the intensive care unit but also to managing palliative and hospice care volume, since many patients not able to receive intensive care will be quickly transitioning to these kinds of care under an active triage protocol.

8. Pastoral care, including spiritual and sacramental support for Catholics, should be emphasized and available, particularly for patients who are expected to die and are not receiving intensive care.

a. A thorough policy to protect pastoral care personnel, health care workers, and priests offering sacraments should be developed and implemented in Catholic health care facilities.

b. Particular attention should be given to the human and spiritual harms of isolation, offering creative and medically sound solutions to providing human connection (e.g., families) and spiritual resources (e.g., sacraments for Catholics and other appropriate pastoral support for Christians and all others).

Disclaimer: This document does not offer any medical or legal advice; rather it speaks to moral principles. It does not take into account all particular legal realities that may affect the issues addressed. Nothing said herein should be understood as an interpretation or claim regarding the law or legal requirements and risks. This document also makes no claims about the best clinical tests, methods, considerations, or calculations for determining “short-term mortality” and other clinically-grounded concepts.