Patient triage and evacuation prioritization at Memorial Medical Center was an improvised system for deciding who would leave first as the hospital’s remaining patients were gathered for evacuation during Hurricane Katrina. Staff divided patients into three numbered groups, placing those judged most able to travel at the front and those deemed very ill, including patients with DNR orders, at the back. The system helped organize a desperate movement of people, but it also made a number on a chart consequential: it could determine who was moved, who waited, and whether a patient or family understood the reason for the delay.

Background: A hospital already losing its capacity

The ranking took shape amid a wider evacuation failure. The city’s mandatory evacuation order did not provide a way out for everyone: residents without cars depended on transportation that officials knew was inadequate, while traffic trapped ambulances trying to move patients. Some ambulances redirected patients to the Superdome rather than out of the city. Memorial itself had known that flooding could disable critical parts of its emergency power distribution system. Eric Yancovich’s calculations indicated that less than four feet of water on Magnolia Street could send water over the loading dock; he warned that the Medical Center could lose power and that patients might have to be moved to the Clara Wing. The anticipated improvements to vulnerable electrical equipment had not been approved because of limited funds.

By Wednesday morning, conditions inside Memorial had further narrowed the options. The hospital was hot and dark; water service had stopped, toilets were backed up, and communications were severely limited. Information traveled through staff members speaking to one another on staircases. The incident command team met at about seven in the morning and initially planned to use emergency triage armbands: green for patients who could walk, yellow for those needing assistance, red for those entirely dependent on care, and black for patients with DNR orders. A nurse was sent to retrieve and distribute the bands. The eventual system, however, was different. Staff had difficulty finding enough official armbands, and the number of patients still awaiting evacuation exceeded one hundred.

The numbered groups and the movement downstairs

That morning, doctors and nurses settled on three numbered categories. Patients in relatively good health who could sit up or walk became 1s and were prioritized first. Patients who were sicker and needed more assistance became 2s. Patients assigned 3s were to go last; the group included people judged very ill and, following a decision doctors had made on Tuesday, patients with DNR orders. Anna Pou took on a coordinating role in the movement, which staff understood as a response to the judgment that they could not save everyone under the circumstances.

The ranking was applied as patients were brought downstairs. Nurses opened charts and read diagnoses by flashlight, using the light sparingly to preserve batteries. A nurse wrote a number with a Marks-A-Lot pen on a sheet of paper and taped it to the patient’s clothing; some numbers were written directly on hospital gowns. Roughly three dozen patients from Memorial and LifeCare were classified as 1s and guided to the emergency-room ramp, where they waited for airboats. Karen Wynn and LifeCare nursing director Gina Isbell oversaw their care there. Around seventy patients were estimated to be 2s, generally placed along the corridor leading toward a passage through the machine-room wall and the route to the parking garage and helipad. About eighteen 3s were gathered in a corner of the second-floor lobby, near a Hibernia Bank ATM and a planter of striped green dieffenbachia.

Patients waiting for transport still received limited assistance: staff changed diapers, fanned them, and offered sips of water to those able to drink. But once patients had been moved from their rooms on Wednesday, most other medical interventions were curtailed. The numbered labels made the process legible to the staff organizing it, yet at least one doctor regarded assigning a person’s prospects by number as efficient but distasteful. The label was not simply a record of a patient’s condition; it also governed the order in which that person would be carried toward a possible way out.

DNR status, information, and individual choices

The categorization was made by medical professionals alone. No system was established to explain the decisions to the patients and families most affected, and some information was deliberately withheld. On the fifth floor of the Clara Wing, transporters stopped moving a bed-bound patient after seeing her DNR order and maneuvered other patients around her. Nurses were told not to tell her daughter that the order was the reason. The daughter had agreed to the DNR when her oxygen-dependent mother was admitted, believing it meant that she would not be resuscitated if her heart stopped—not that she would be left behind during evacuation.

When a man with a walkie-talkie told the floor’s staff they had to leave, the nurses refused to go without the patient. Michelle Pitre-Ryals explained to the daughter that she could ask for the DNR order to be discontinued. The daughter did so, and the patient was moved downstairs, then out through the parking garage toward the helipad. She survived that immediate ordeal. Her case showed that a DNR order and a decision not to rescue were not the same thing, even though the evacuation practice had connected them. It also showed how a family member’s understanding of an order could change when its consequences became clear.

On LifeCare’s seventh floor, Angela McManus learned that her mother, Wilda, was being deferred because she had a DNR order. Angela asked staff to rescind it, but they told her no doctors were available to do so. Wilda remained upstairs. Doctors in the staging area on the second floor told LifeCare staff to stop sending patients down because there was no more room. Thus the ranking operated alongside a separate physical bottleneck: even patients whose families wanted a change in status could not necessarily obtain one, and the staging area’s capacity constrained who could reach the evacuation route.

The risks and aftermath of improvised triage

The system’s urgency did not remove the uncertainty inherent in predicting who could wait. Triage assessments can vary between experienced rescuers, and a patient’s condition can change after an initial ranking. Once assigned, a category can be difficult to revisit, even when more help or new information becomes available. Some disaster protocols therefore place patients considered unlikely to survive after those needing immediate treatment, but before patients with significant injuries who can wait. At Memorial, staff had little if any training in triage systems and no particular protocol to guide them. Their three groups were developed in a crisis, rather than selected through a process that involved patients, families, or the wider public.

The evacuation’s emotional effects persisted. Gina Isbell later described sleeping poorly, having nightmares about patients, and feeling guilty about the people she had chosen to work the storm with and the patients who died. Karen Wynn, preparing for a possible grand jury appearance, found that discussions of the events could make them feel newly present and leave her struggling for weeks afterward. These accounts do not settle the judgments made during the evacuation; they show that the sorting process and its consequences remained part of the staff’s lives after the immediate emergency ended.

The episode belongs to a larger account of disaster response in which the ability to evacuate depended on far more than a patient’s diagnosis: transportation, power, communications, space, and the decisions of people controlling access to each stage all mattered. The numbered system offered order when those supports were failing. Its limits were equally revealing: a rapid ranking could not make an uncertain prognosis reliable, ensure that patients knew what their status meant, or guarantee that someone categorized to wait would still be found and carried to safety.