New York State disaster ventilator allocation protocol

The New York State disaster ventilator allocation protocol was designed for a crisis in which hospitals could not provide life-sustaining treatment to everyone who needed it. It offered advance rules for distributing scarce ventilators, including criteria that could deny or withdraw treatment from some patients. Its later use as a model during Hurricane Sandy showed both the value of a shared decision process and the unresolved ethical stakes of rationing care.

Origins in pandemic planning

In 2008, New York planners published the protocol amid concern that a severe influenza pandemic could overwhelm hospitals. The document’s development was linked to the arrests of health professionals in connection with deaths at Memorial Medical Center during Hurricane Katrina, and to fears that another emergency might again force providers to make life-and-death choices without agreed criteria. New York State officials had been exercising scenarios involving H5N1 avian influenza, a strain that had caused deadly human outbreaks and that experts feared could mutate into a form that spread readily between people.

The planning problem was not merely a shortage of equipment. Ventilators were expensive, required highly trained operators, and consumed oxygen, another resource that could become scarce. Officials considered ways to stretch supply, such as canceling elective surgeries, and New York purchased and stockpiled additional ventilators. Those measures could help in a moderate pandemic but would fall far short in an outbreak on the scale of the 1918 influenza pandemic, which sickened more than a quarter of the population and killed an estimated 50 million people worldwide. The protocol was intended for the more severe scenario: a demand so great that hospitals would have to decide in advance how to distribute what remained.

Its stated objective was to save as many lives as possible within an ethical framework. Setting criteria before a crisis was also meant to spare exhausted clinicians from inventing rules at the bedside while under extreme stress. That rationale answered one problem revealed at Memorial, where staff had to make consequential decisions amid Hurricane Katrina’s breakdowns in evacuation, communications, and hospital infrastructure. It did not, however, remove the burden of deciding whose treatment would be withheld.

Scoring and limits on access

The New York plan drew partly on a 2006 resource-allocation proposal developed in Ontario during planning for a severe SARS outbreak. The Ontario approach adapted the Sequential Organ Failure Assessment, or SOFA, score—a tool used to track the progress of intensive-care patients—to help determine who would receive intensive-care treatment in an emergency. The New York planners adopted the score despite important limitations: it had not been designed to predict survival and had not been validated for assessing children. It was used because planners lacked a more appropriate alternative.

Under the protocol, patients with the highest scores—among the sickest—could be denied access to scarce ventilators. The plan also excluded some people with serious chronic conditions, including kidney failure, cancers that had spread and carried a poor prognosis, and severe, irreversible neurological conditions likely to be fatal. These criteria translated the general goal of saving the greatest number into decisions about particular patients, whose diagnoses and prospects could place them on opposite sides of an allocation threshold.

The document went beyond deciding who could initially receive a ventilator. It included procedures for removing patients who did not improve after a trial period, with or without their families’ permission. New York officials examined whether the governor could suspend a state law barring physicians from withdrawing life support without express consent from patients or their authorized health agents. The possibility of switching such extraordinary measures into effect quickly made the protocol a plan not only for distributing equipment, but also for changing the legal and clinical conditions under which treatment could be stopped.

Use as a model at Bellevue

Years later, as Hurricane Sandy threatened Bellevue Hospital, critical-care medical director Laura Evans confronted a related shortage. Hospital leaders believed that a generator in a neighboring building, served by a separate and better-protected fuel pump, could keep exactly six intensive-care outlets operating. There were 56 patients. Evans had studied New York’s ventilator-allocation guidelines and voluntarily repurposed their scoring approach to help decide which patients should have access to the outlets. The plan had been developed for a severe respiratory pandemic; it had not been tested for this kind of immediate hospital power emergency.

Evans convened an ad hoc committee rather than making the choices alone. Its members had no patients under consideration, so each patient’s primary clinician could continue advocating for that patient. The group included doctors, nurses, and ethicists, but no patient or family representatives. It used defined scoring guidelines to consider the patients together, and communicated its decisions and reasons to other staff. Some staff challenged the choices, but accepted them. Within about two hours, the team had made a list and moved patients between beds and outlets.

The circumstances sharpened the distinction between a planned allocation process and a decision made in the instant of collapse. As Evans and her colleagues completed their list, utility power failed; backup generators restored function after about seven seconds. The command center then warned that, apart from the six outlets, power might be lost within 60 to 90 minutes. Floodwater was filling the basement of the 25-story hospital, and staff prepared to ventilate patients manually with Ambu bags if needed. The episode does not establish that the protocol was formally activated or that any patient died as a result of the allocation. It shows instead how a framework devised for one anticipated emergency could be adapted to another, under the pressure of an actual threat to hospital power.

Fairness, consent, and accountability

Evans’s use of the guidelines reflected a concern with process as well as outcome. She wanted decisions that could be explained, applied consistently, and defended as fair and equitable. She knew the story of Anna Pou and the controversy surrounding Hurricane Katrina, and imagined that the decisions at Bellevue might later be examined by the public. A committee, written criteria, and communication with other staff offered a way to make responsibility less dependent on a single clinician’s judgment. But the absence of patients and families from the committee also marked a limit: the process sought fairness among patients without including their representatives in its deliberations.

The protocol’s most consequential unresolved question was whether a procedure that appears orderly is necessarily justified. Withdrawing treatment without consent raises serious concerns; doing so under a rule that may not actually save more lives would be harder still to defend. The score’s limits, the need to make forecasts about survival, and the authority contemplated for withdrawing life support all complicate the promise of neutral criteria. Advance planning can reduce improvisation, but it cannot by itself settle what counts as acceptable evidence, whose interests should be represented, or how much weight should be given to an individual patient’s consent.

The protocol thus stands between two kinds of failure. Without preparation, clinicians may face scarcity in isolation, as staff did amid the compounded emergencies at Memorial. With a plan, the decision may become more consistent and transparent, but also more formalized—and capable of authorizing treatment limits that patients or families did not agree to. Its importance lies not in resolving that tension, but in making visible the choices that ordinary medical practice can leave unspoken until a disaster forces them into view.