The Bellevue Hospital Power and Evacuation Crisis During Sandy tested whether a hospital could preserve critical care and move patients when floodwater threatened its power supply. At Bellevue Hospital Center, Dr. Laura Evans confronted a stark mismatch: 56 adult ICU patients depended on electrical equipment, while the hospital believed it might be able to provide only six working outlets. A triage list was prepared, but improvised fuel transport ultimately kept the backup generators running, allowing all of Evans’s patients to remain on power as the hospital evacuated.

Flood Risk and the Six Outlets

The danger was not simply that the storm might interrupt ordinary electricity. Hospital leaders warned Evans that a storm surge from the East River could flood the surrounding land and disable pumps used to deliver oxygen and vacuum suction, as well as the pumps moving fuel from underground tanks to the backup generators. If city power failed and those fuel pumps stopped, reserves in the generators’ small attached tanks might last only two hours. Nearly every patient in Evans’s 56-bed adult ICU relied on electricity; some were on life support, received intravenous drugs to regulate heart rate or blood pressure, or depended on intra-aortic balloon pumps to maintain circulation.

The command team identified a possible fallback: a generator in a neighboring building, connected to a separate and better-protected fuel pump, might continue operating. But it could supply exactly six ICU outlets. Hospital leaders asked Evans to identify which patients should receive them. The immediate crisis was therefore also a problem of allocation: the possible loss of power would make decisions about individual patients unavoidable, even though the hospital had no certainty that its limited alternative supply would be needed.

A Triage Process Prepared

Evans had access to a planning document unavailable to clinicians at Memorial Medical Center during Hurricane Katrina: New York’s ventilator-allocation guidance, published in 2008 after concerns that a pandemic could force providers to make life-and-death choices. She repurposed its illness-severity scoring approach to help decide which patients would receive the six outlets. The guidance had been devised for a severe respiratory pandemic, not tested as a rule for distributing power during a hospital flood.

Evans assembled an ad hoc committee rather than making the choices alone. Its members were doctors, nurses, and ethicists who did not have patients under consideration. The separation was intended to let primary physicians continue advocating for their own patients while the committee compared cases across the ICU using shared scoring guidelines. Patients and their families, however, were not represented. The committee’s decisions and their reasons were communicated to other staff; some colleagues challenged the choices but accepted them. Within about two hours, the team had compiled a list and arranged patients in relation to the scarce outlets.

The process was also shaped by Evans’s awareness of the earlier Katrina controversy involving Dr. Anna Pou. Evans wanted a method she could later explain and defend, applied consistently and as fairly as possible. Yet the plan left a serious communication problem unresolved: she and her colleagues could not imagine how to tell patients and relatives who would not receive an outlet, and they postponed doing so. Evans later concluded that families had a right to know. The team also did not consider asking whether a selected patient or family might voluntarily relinquish an outlet for someone else.

Power Failure and Improvised Support

As the list was being completed, the lights flickered out. City utility power had failed, and about seven frightening seconds passed before the backup generators restored power. The command center then warned that, apart from the six designated outlets, power might be lost within 60 to 90 minutes. Floodwater was filling the basement of the 25-story hospital and entering elevator pits with enough force to dislodge elevator doors.

Staff prepared to sustain patients without relying wholly on machines. Two health professionals were stationed at the bedside of every patient dependent on a ventilator, ready to move air manually with Ambu bags. Bright-orange extension cords connected to backup power ran through the patient corridors. Nurses counted drops on intravenous pumps so that, if the equipment failed, they could administer vasopressors manually by drops per minute. These preparations did not remove the danger, but they gave the staff a way to respond if power was interrupted before patients could be moved.

The feared cutoff was prevented by a practical workaround. Volunteers formed a chain and carried fuel up thirteen flights of stairs to feed the generators manually after the fuel pumps failed. Their effort kept backup power from going out and spared Evans’s patients the immediate consequence of the six-outlet limit. During the evacuation, all of her patients remained on backup power. The crisis thus turned not on the triage list alone, but on whether people could keep the equipment running while the hospital’s ordinary systems were threatened by water.

Evacuation and the Patients Kept Until Last

With the elevators out of service, evacuation of the large hospital took days. Staff faced particular concern about two patients left until the end. One weighed about 600 pounds, making movement without elevators exceptionally difficult. The other was critically ill and so fragile that his doctors feared he might die while being carried. Their circumstances made clear that evacuation was not complete merely because most patients had been moved; the final transfers required continued effort and attention to patients whose needs complicated the route out.

After other patients had been rescued, soldiers carried fuel up thirteen flights of stairs for several more days, until the elevators could operate. Doctors performed heart surgery on the fragile patient inside the disabled hospital. In the end, both he and the approximately 600-pound patient were safely transferred to another hospital. The evacuation therefore ended not with the abandonment of the hardest cases, but with their rescue after the hospital had already moved the others.

Katrina’s Legacy and the Limits of Planning

The episode connects an immediate New York emergency to lessons drawn from Katrina. New York’s ventilator-allocation planning had been influenced by the arrests of health professionals at Memorial and by fears that a future outbreak could again force providers to choose between patients. Evans knew that history and sought a process that could be explained to colleagues and, later, to the public. But a scoring system did not by itself resolve whether patients and families should be included, informed, or invited to participate in decisions about scarce support.

The Bellevue crisis also showed how an anticipated triage decision can be overtaken by improvised action. The six outlets framed a potential choice between patients; the volunteers’ fuel chain kept that choice from becoming necessary in the expected form. Even then, manual ventilation, bedside staffing, and a prolonged evacuation remained essential. The outcome depended on both preparation and the people who adapted when the backup plan itself was threatened. Bellevue’s patients were maintained on power and the two most difficult transfers were completed safely, but the experience left the unresolved ethical questions visible: who should decide, what should be communicated, and how fair rules can be made meaningful under pressure.