Skip Navigation U.S. Department of Health and Human Services
Agency for Healthcare Research Quality
Archive print banner
Hospital Evacuation Decision Guide

Public Health Emergency Preparedness

This resource was part of AHRQ's Public Health Emergency Preparedness program, which was discontinued on June 30, 2011, in a realignment of Federal efforts.

This information is for reference purposes only. It was current when produced and may now be outdated. Archive material is no longer maintained, and some links may not work. Persons with disabilities having difficulty accessing this information should contact us at: Let us know the nature of the problem, the Web address of what you want, and your contact information.

Please go to for current information.

Chapter 2. Pre-Disaster Self-Assessment

Significant and detailed planning is required in order to support sound decisionmaking regarding hospital evacuation. Hospitals are required to have plans in place detailing evacuation, but those plans differ tremendously in their level of detail and often lack construct and context for decisionmaking. Two additional tasks, which together comprise a Pre-Disaster Self-Assessment, should be completed as part of the planning process.

The first task is completing a Pre-Disaster Assessment of Critical Infrastructure (go to Table 4). As discussed in Chapter 3, a key consideration in deciding whether to issue a pre-event evacuation order is to assess vulnerabilities and determine the anticipated impact of the impending event on a hospital and its surrounding community. This impact on critical infrastructure is paramount, with water and power being most important.

The second Self-Assessment task involves estimating the time required to evacuate patients from the hospital. An Evacuation Time Self-Assessment Worksheet (Table 5) is included in this chapter to assist with this task. For Advanced Warning Events, it is critical to have estimated how long it will take to move patients out of the building and relocate them to other hospitals, ideally under a number of different sets of assumptions regarding patients, building conditions, and transportation resources.

Hospital staff should complete this self-assessment as part of their disaster and emergency preparation planning, update it when needed, and not wait until an actual event occurs that necessitates an evacuation decision.

Return to Contents

Critical Infrastructure Self-Assessment Worksheet

The Pre-Disaster Assessment of Critical Infrastructure Worksheet (Table 4) is divided into eight sections: municipal water, steam, electricity, natural gas, boilers/chillers, powered life support equipment, information technology and telecommunications, and security. (Hospital staff, obviously critical resources, are considered in the Evacuation Time Self-Assessment Worksheet, later in this chapter.) The focus on environmental systems (HVAC), water, and electricity is appropriate, given that they are critical for hospital operations, and their loss for an extended period invariably triggers a need for evacuation. The Worksheet can be used in conjunction with the National Infrastructure Protection Plan (NIPP), which is a management guide for protecting critical infrastructure and key resources. A health care and public health sector-specific plan that details how the NIPP can be applied to health care settings is due to be released in 2010.17

Decision teams should know how long their hospital can shelter-in-place if critical infrastructure are damaged. For example, how long could the hospital maintain a safe temperature without city water during the summer months, and how long could essential power be maintained with only the current on-site fuel supply? The Pre-Disaster Assessment of Critical Infrastructure Worksheet (Table 4) is designed to help decision teams consider the vulnerabilities of their critical infrastructure and their hospital's ability to shelter-in-place, which in turn may guide investment decisions for mitigating vulnerabilities.

If critical infrastructure has not sustained damage, the hospital's ability to shelter-in-place will be affected by the extent to which staffing levels can be maintained, and whether the supply of critical consumable resources—such as food, blood, and medications—can meet the needs of patients and staff, drawing on existing caches within the hospital and regular and backup supply channels. Maintaining safe levels of staffing and consumable resources should be addressed in a hospital's plan for sheltering-in-place. If there is no such plan, the ability to shelter-in-place for more than a few days may be degraded.

Municipal Water

Considerable anecdotal evidence, as well as published reports, indicate that loss of water will lead to hospital evacuation if not promptly restored. Loss of the municipal water supply also jeopardizes hospital sprinkler systems and, in some hospitals, heating systems. A hospital pre-disaster self-assessment should recognize the presence/absence of backup water supply lines (in the event that the main line fails) and any on-site water reserves, such as a storage tanks or wells.

The Critical Role of Water

Children's Hospital of New Orleans withstood Hurricane Katrina and the resulting flood, and had sufficient potable water, generator fuel, staff, and supplies to shelter-in-place for many days. The hospital sits atop a levy, which floodwaters did not reach, and was essentially unscathed by the storm or the flood. When city water failed, however, it became impossible to fill the cooling tower, and the hospital lost air conditioning, forcing an evacuation in the heat of the Louisiana summer.7

Also in the aftermath of Hurricane Katrina, Kindred Hospital in New Orleans had power, but the hospital's water supply was cut off, which meant air conditioners could not operate. The loss of water necessitated evacuation of the hospital, including 50 ventilator-dependent patients.6


Some municipalities use large steam production plants to create steam and pipe it underground to buildings in local areas, including hospitals. For example, hospitals in Boston's Longwood Medical Area, which includes three tertiary care hospitals as well as a medical school, research labs, and ambulatory care areas, are heated with steam produced by the same off-site utility companyvi. These steam production plants are critical infrastructure, as are the pipes that deliver steam to area hospitals. Loss of water to the steam production plants, inability to generate steam, or inability to pipe it underground to hospitals would jeopardize heat and lead to hospital evacuation within 1-2 days during a typical northeastern winter.18 Some hospitals also use the incoming steam to generate electricity and for such hospitals, loss of steam would also mean loss of some of their electrical capacity.19 A hospital self-assessment should therefore include recognition of reliance on steam that is generated off-site and piped in for heating purposes, electricity generation, or both.

Natural Gas

For hospitals that use natural gas for heat and/or hot water, damage to gas mains lasting more than 1-2 days (especially in the winter) could lead to an evacuation. A hospital self-assessment should therefore recognize reliance on natural gas, whether there is more than one gas line feeding the hospital, and whether gas from just one intact gas line could meet the most critical needs of the hospital.


Electricity plays a large role in evacuation decisions. Prolonged loss of electricity can lead to HVAC loss, which can necessitate evacuation. In addition to controlling ambient temperature, electricity is essential for many medical technologies (e.g., monitors, CT scanners, dialysis machines, ventilators, incubators) as well as other critical functions. Patients dependent on electricity-driven life support equipment would require evacuation soon after power failure. Ventilator battery packs, for example, last only 2-3 hours, and the accompanying suction devices generally have no battery packs.22 The number of hours that a hospital can function without municipal electricity, or adequate fuel for backup generators, may be critical factors in an evacuation decision.

Backup Generators

Although patients and most staff and family members were evacuated within 4 days of the Hurricane Katrina strike due to lack of city water and air conditioning, the Veterans Administration Medical Center in New Orleans was able to keep limited electrical power available, as its backup generators did not fail and remained several inches above the flood water. A coffer dam was erected around the fuel fill connection for the underground fuel storage tank to permit refueling a week later, and power was maintained to support the security and engineering staff that remained on duty until utility power was restored 3 weeks after the storm.20

In the same post-hurricane flood, Charity Hospital's (Louisiana) backup generators failed, leaving them to rely on inadequate, portable generators (without direct fuel lines) to power life-support equipment such as ventilators.9

At Tulane University Hospital (Louisiana), backup generators failed, jeopardizing tissue and blood samples from patients enrolled in clinical trials of experimental cancer treatments.21

Most hospitals have backup generators, although the adequacy of these generators should be carefully assessed. For example, backup generators require fuel, and hospitals vary considerably in their on-site fuel storage capacity, whether there is a direct feed from the fuel tank to the generators, and whether it would be possible to refill the fuel storage tank—which is often underground—during a flood or after an earthquake.

A hospital pre-disaster self-assessment should include the number and size of backup generators and an estimate of the length of time these generators can sustain electrically powered life-support equipment and HVAC. A self-assessment should also consider the fuel storage capacity on site and any potential refueling issues.


Some hospitals use boilers to generate hot water; others use them for heating purposes, as well. Most hospitals also have chillers for air conditioning (with or without cooling towers). Redundancy in these types of critical infrastructure is rare, and their loss could necessitate an evacuation, depending on weather conditions. A hospital pre-disaster self-assessment should therefore recognize vulnerabilities due to the loss of boilers or chillers, irrespective of the loss of electricity, water, or steam.


Mount Auburn Hospital in Cambridge, Massachusetts, began to evacuate patients within hours of boiler failure in December 2005.23 An explosion that destroyed both the boiler and the chiller at a New Jersey also hospital prompted evacuation.24

Powered Life Support Equipment

Some powered life support equipment (e.g., ventilators) may have backup battery packs in the event of an electricity failure. The life of these batteries is generally 2-3 hours; patients dependent on such equipment may therefore need to be evacuated more quickly than others. A self-assessment should include an inventory of the powered life-support equipment in use on an average day, how many of these have backup battery packs, and how many hours these batteries will last (the latter being a critical factor in deciding how quickly such patients must be evacuated).

Health Information Technology

Loss of key health information technology (IT) and telecommunications systems will, at a minimum, significantly reduce a hospital's ability to deliver health care services efficiently. For example, if a hospital's computerized provider order entry (CPOE) system suddenly goes offline, substantial delays in order completion can be expected, as well as increased risk of errors. In other instances, service delivery may cease altogether because automated systems cannot be quickly replaced by manual systems. For example, hospitals that rely on decentralized pharmacy and automated dispensing units may no longer have redundant systems for safely filling patient medication orders.25

If patients must be evacuated, paper records are relatively easy to send with them, but many hospitals are moving away from paper records in favor of electronic medical records (EMRs). Some EMR systems can rapidly create and print a paper discharge summary to accompany each evacuated patient; other systems cannot, and the time required to create a useful discharge summary (current medications, allergies, orders, brief history) may substantially delay evacuation.


Landline telephones in New Orleans failed when lines were damaged and when backup generators for phone switches were flooded or ran out of fuel following Hurricane Katrina. The local wireless cell phone networks failed as well because the data systems that exchange information for routing and billing calls also lost power. Because information for each account holder and cell phone device is generally stored on computers in the "home" market where the account is held, every customer with accounts originating in New Orleans lost wireless service. Individuals who had recently moved to New Orleans and still had a wireless service contract in another city, or who arrived from elsewhere to provide assistance, were able to use their cell phones (as long as the batteries lasted) for outgoing calls9 because their "home" accounts in other cities ran on computers that were unaffected by the storm.26 Incoming calls, however, could not be completed because the local computers that identify and locate the intended recipient of an incoming call were offline. Satellite phone systems may work in circumstances like those in New Orleans, for calls placed to locations outside the city. Satellite phones obviously cannot reach landline or wireless phones that are out of service due to the disaster, and thus are of limited use for local communication.

Institutions and individuals with computers equipped for Voice Over Internet Protocol (VOIP), and who had power and Internet access after Hurricane Katrina, retained telephonic communication. In VOIP the audio (voice) signal is converted to digital packets of information that travel from one Internet address to another on the same VOIP system. VOIP networks were deployed by FEMA, and by fire and other emergency responders, in the days and weeks following Hurricane Katrina.27

Finally, lack of radio interoperability between hospitals and emergency responders, as well as between different teams of emergency responders (fire, military, police, EMS) may also jeopardize essential communication. (The Federal Government does not mandate how a State or local community organizes incident response activities or communications.28) Some hospitals in New Orleans resorted to local HAM radio operators to relay essential messages within and beyond the city.29 Two-way radio interoperability among emergency responders has improved more in some cities than in others since 2005.

The self-assessment focuses on the extent to which computer servers and essential data are backed up or managed offsite, whether redundant hardware and software systems exist, whether manual, paper-based systems can be quickly reintroduced, and whether the hospital has backup telephonic communication that does not rely on local service providers.


Security Concerns During Evacuation

One day after Hurricane Katrina made landfall, Kindred Hospital New Orleans lost its water supply, and the hospital administrator decided to evacuate the hospital. The evacuation was delayed because of civil unrest and looting in the neighboring area, as many of the ambulances sent to pick up patients were forced to turn back before reaching the hospital because of security concerns. Additional private security staff sent by Kindred corporate headquarters to protect the hospital were also delayed in their arrival.30

During a disaster, additional security staff are often needed to: 1) keep unauthorized people out of the hospital, 2) guard transport vehicles as they move to and from the hospital evacuating patients, or 3) maintain order inside the hospital. While the hospital is empty, security staff may be needed to safeguard the property and the costly medical equipment and supplies left behind. Some hospitals rely on a contracted service for primary or backup security; others employ their own security staff and augment this force when needed.30 Night and weekend shifts may have fewer security staff, making an evacuation at those times potentially less secure.

In some disaster situations, municipal and State police departments may be unable to deploy officers to hospitals because their officers, and the National Guard, are at capacity responding to other community needs. When civil unrest has been extreme, decision teams have occasionally resorted to hiring external security forces to escort evacuation vehicles. A pre-disaster self-assessment should address the availability of security staff (especially if the entire city is evacuating), backup plans if municipal/State law enforcement are insufficient, and how best to augment security staff around the clock, throughout an evacuation.

Return to Contents

Estimating Evacuation Time

As part of pre-disaster planning, a decision team should estimate the time required to safely evacuate all patients. This is especially important for Advanced Warning Events, as will be discussed in Chapter 3.

More specifically, hospitals should estimate time requirements for two components of evacuation time. The first is the time to empty the building—that is, the time required to move patients from their location inside the hospital (e.g., their room) to a staging area from which they can be loaded into ambulances and other vehicles for transport to another hospital. The staging area might be the lobby of the hospital, the emergency department (which has ambulance bays), or a parking lot across the street from the hospital. Movement of patients from their hospital rooms to the staging area and beyond depends in part on factors internal to the hospital, such as whether hospital IT systems can generate a patient discharge summary, whether elevators are operational, whether staff have participated in evacuation drills, and how quickly additional staff can arrive to help with the evacuation.

Urgent Evacuation - No Advanced Notice

A fire in one tower of Mt. Sinai Hospital (New York) necessitated the rapid movement of patients to other parts of the hospital. There was no time for a central evacuation decision or order; staff on each unit moved patients horizontally and then vertically, as the tower filled with smoke. The evacuated tower included an ICU and operating rooms; surgical staff moved patients mid-procedure and finished operating in the emergency department of the adjacent tower.14

After the 1994 Northridge (California) earthquake, six hospitals were evacuated immediately. Ambulatory patients were escorted from the building to a nearby open area. Hospital staff next moved patients in wheelchairs and, lastly, those in beds. At one of the hospitals, the strategy was to remove as many patients as possible, as quickly as possible, because it was not clear how long the building would remain standing.3

It is important to distinguish between an orderly and planned evacuation, in which there is time to move patients in a manner that maximizes safety for all patients and staff, and a "drop everything and go" evacuation, in which patients and staff are in immediate danger and must exit the unit and/or hospital as quickly as possible. In the latter case, optimal procedures for safely moving patients may be abandoned in favor of the fastest possible egress.

The second component of evacuation time is the time to transport patients—that is, the time required to transport patients from the staging area to receiving hospitals or other care sites. The time required to transport patients from the staging area to receiving care sites depends primarily on factors external to the hospital, such as transportation resources availability, road conditions, and the locations of hospitals that can accept and properly care for patients.

Key factors that affect evacuation time are listed below and in the accompanying Evacuation Time Self-Assessment Worksheet (Table 5). They include:

  • Number of patients and mix of patient acuity.
  • Available staff.
  • Available exit routes within the hospital.
  • Patient transportation requirements.
  • Available transportation resources (vehicles, as well as the necessary staff, equipment, and supplies that must be in the vehicles).
  • Entry and egress points at the hospital.
  • Road and traffic conditions.
  • Location of receiving care sites.

Number of Patients and Patient Acuity Mix

The total number of patients in the hospital who need assistance to evacuate safely will typically be substantially fewer than the total patient census. Some patients will be medically stable and likely to self-evacuate or evacuate with family members.vii,31

Census Reduction

In the aftermath of the Three Mile Island (Pennsylvania) incident, area hospitals, anticipating possible evacuation orders, discharged many ambulatory and stable patients, reducing the area-wide patient census from 80 to 50 percent of capacity.2

Of 575 hospitalized patients in Memorial Hermann Hospital and Memorial Hermann Children's Hospital (Texas), 169 were discharged prior to when Tropical Storm Allison made landfall, leaving 406 who sheltered in place during the storm (and who were subsequently evacuated with the assistance of hospital staff).4

During the Red River (North Dakota) flooding in 2009, patient census at Innovis Health was reduced from approximately 85 to 60 percent, through a combination of early discharges and halting transfers of trauma patients from hospitals more than 50 miles away.16

Other patients will be ambulatory and can walk out of the hospital with assistance, while still others will require wheelchairs. Some will require sophisticated equipment and handling if they are to survive the evacuation, and a few very ill patients will be unlikely to survive if moved. The Evacuation Time Self-Assessment should record the typical number of patients in the hospital who will require assistance from hospital staff to evacuate, by patient type or acuity, because different levels of assistance and types of hospital staff are required by different types of patients.

At a minimum, patients in the following hospital units should be considered:

  • Adult ICU.
  • Pediatric ICU.
  • Neonatal ICU.
  • Adult medical/surgical ("floor") units.
  • Pediatric medical/surgical ("floor") units.
  • Psychiatric unit.
  • Burn Unit or Burn ICU.
  • Other specialty care units.

In addition, within these units special consideration should be given to bariatric patients, patients requiring dialysis, patients in negative pressure/isolation rooms, and patients from correctional (prison) facilities.

Available Staff

Staff are required to move patients out of the hospital and may be needed to accompany patients during transport to a receiving care site. Depending on the type of disaster, there will likely be staff shortages. It is helpful to try to pre-estimate the attrition rate of a hospital's workforce during a disaster, as many employees may themselves become victims of the disaster, or may have family responsibilities that interfere with their ability to staff the hospital (e.g. evacuating dependent children). Decision teams at some hospitals that face annual hurricane threats require essential staff to remain in the hospital throughout a declared disaster so that they are available to assist in an evacuation—this is a condition of employment and staff must make alternative arrangements for their dependents.32 Hospitals that have no such personnel policies may find their staff substantially depleted during a community-wide evacuation. A pre-disaster self-assessment should therefore consider the personnel policies in place and the staff deficits that could occur in different types of disasters that may involve community-wide evacuation orders.

Staff Shortages During Disasters

Memorial Hermann Hospital (Texas), which was evacuated after Tropical Storm Allison made landfall, was isolated by impassible roads. As a result, staff were delayed in reaching the hospital to assist with the evacuation and could not arrive for work until 9:00 a.m. the day of the evacuation, leaving staff already present at the hospital unable to leave.33

Following the Three Mile Island (Pennsylvania) incident, area hospitals experienced severe staffing shortages, as hospital staff—especially those with small children—self-evacuated the community to protect their own safety. By the fourth day of the incident at the Three Mile Island nuclear facility, some hospitals had to combine units due to staffing shortages, and one hospital had only 6 of 70 physicians available to work.2

Hospitals also typically have significantly fewer staff on hand during night and weekend shifts, which would greatly affect the ability to quickly move patients out of the hospital in an urgent evacuation. Some hospitals rely more heavily than others on staff from temporary agencies, or 'traveling' staff who contract for short assignments (especially nurses and technicians). Such staff may not be as readily available as full-time hospital employees during an emergency. Volunteers, visitors, and family members may be available to assist in evacuating some patients. Volunteers must be assigned appropriate tasks, as trained medical staff are required to move and transport most patients with intensive care needs.

Volunteers Assist with Hospital Evacuation

Hundreds of volunteers, including Boy Scouts, were used to help move patients out of Memorial Hermann Hospital (Texas) after flooding caused by Tropical Storm Allison.34

Moving NICU and PICU Patients

Children's Hospital of New Orleans (CHNO) reached out to other children's hospitals across the country for evacuation assistance. These hospitals responded by sending planes for air evacuation and teams trained in emergency transfer of fragile children and infants. Other children's hospitals also coordinated the transfer of each NICU and PICU patient to appropriate receiving hospitals. Administrators from CHNO believe that the evacuation was successful because all ICU patients were transported by experienced teams, and were received by hospitals prepared to meet their needs.7

Available Egress Routes From Within the Hospital

While unlikely to be a problem during an "orderly and planned" evacuation, egress from a hospital may be severely constrained during a "drop everything and go" evacuation. Stairwells or exits may be obscured by smoke or unavailable because of fire. Stairwells may be dark if backup power has failed. Elevators can also be out of service, lengthening the time required to move all patients out of the hospital.

Out of Service Elevators

Because elevators were not operating, patients at Memorial Hermann Hospital in Houston, Texas, were carried down 10 flights of stairs on backboards without overhead lighting or air-conditioning. Up to five infants were secured to one backboard. Several adults were needed to carry out each adult patient. Evacuation was temporarily halted when staff and volunteers were exhausted, so as to avoid injuries.4

Rapid evacuation of patients from hospitals following the Northridge (California) earthquake included carrying patients on backboards and on canvas slings, after elevators failed.35

During post-hurricane flooding in New Orleans, water entered the elevator shaft "wells" at the VA Medical Center in New Orleans, making the elevators unusable even though power was not interrupted.36

Patient Transportation Requirements

The Self-Assessment Worksheet should include an estimate of the number (or percentage) of patients who will require transportation resources to be safely transported to other facilities. Transportation resources include not only the vehicle, but also the required accompanying staff, equipment, and supplies. Vehicle types typically include buses (if patients are ambulatory or need limited assistance), wheelchair vans (if they are capable of sitting up), BLS ambulances, and ALS ambulances. Some patients who are dependent on powered life support equipment may only be safely evacuated in an ALS or critical care vehicle with appropriate staff on board. For example, extra corporeal membrane oxygenation (ECMO) units typically have little or no battery backup, and patients requiring this equipment must be transported immediately when power fails in a vehicle equipped to power the ECMO unit.37 Pre-disaster planning and coordination with ambulance providers and local, county, and State EMS agencies is essential so that hospital staff know what types of patients can be transported safety in the ambulances used in their area.

Typically, a detailed census review will be required for a precise count of the number of patients that could be transported via bus, wheelchair van, BLS ambulance, or ALS ambulance because there is no formula based on patient diagnosis for mapping patient types (adult ICU, NICU, PICU, adult floor, pediatric floor, psychiatric unit, other specialty units) to vehicle types. As reference points, patient vehicle requirements assumed in planning and tabletop hospital evacuation exercises in New York City and Los Angeles may be instructive.38 Using patient data from six hospitals in New York City, a planning exercise estimated the following percentages: 14 percent of patients would require transport via ALS ambulance, 13 percent via BLS ambulance, 40 percent via wheelchair van, and 33 percent via bus. The corresponding figures for Los Angeles, based on patient characteristics at three hospitals, were: 40 percent of patients would require transport via ALS ambulance, 20 percent via BLS ambulance, 20 percent via wheelchair van, and 20 percent via bus.

Other Types of Vehicles Used to Evacuate Patients Following Hurricane Katrina

Many hospitals have no rooftop helipads, and those that do can typically accommodate only one helicopter at a time—able to carry, at most, two patients and associated staff and equipment.37 ICU patients from Tulane University Hospital (Louisiana) were evacuated by helicopter.

Patients may be moved to an airfield for evacuation by rotor or fixed wing aircraft. Many patients from Children's Hospital of New Orleans were moved first to a nearby airport, and then to other children's hospitals around the country.7

Boats may be needed to evacuate patients during floods. Patients were moved by boat from Charity Hospital (Louisiana), which had no helipad, to Tulane University Hospital across the street, and were then moved up to the helipad on the roof of the parking garage.40

Tall, 18-wheel trucks were able to roll through the floodwaters after Hurricane Katrina; many bedbound and ventilator-dependent patients from Charity Hospital, who had to be moved on stretchers, were evacuated by truck.9

Of course, many other types of vehicles can be used to evacuate patients, particularly when post-event conditions limit access to the hospital. Depending on the circumstances, patients may have to be transported in vehicles (or accompanied by hospital staff) that do not conform to ordinary standards of care.39

Available Transportation Resources

Transportation resources include not only the vehicle, but also the staff, equipment, and supplies that must accompany the patient in the vehicle. The number of vehicles of each type that are available to transport evacuated patients (assuming they are properly staffed and equipped) is a critical determinant in how long it will take to move all patients to receiving care sites. For example, if 100 patients require an ALS ambulance, and only five such ambulances are available, then each ALS ambulance will need to make 20 round-trips during the evacuation. In the New York City hospital evacuation planning exercise cited above, it was assumed that a relatively high percentage of ambulances could be devoted to the evacuation (40 percent of the city's fleet of private-sector ALS ambulances) because it was an Advance Warning Event. By contrast, because the Los Angeles planning exercise was a no-notice evacuation representing an earthquake scenario, a much smaller percentage of ALS ambulances were assumed to be available for the evacuation (only 5 percent of the city's ALS ambulance fleet).38

Even with pre-existing contracts with ambulance and other transportation providers, there is no guarantee that vehicles will actually be available, particularly if multiple hospitals are evacuating simultaneously, in which case the "competition" for ambulance and other transportation resources will likely be significant. Vendors who had contracted to transport patients from Charity Hospital did not deliver on their contractual obligations, stating that they lacked both vehicles and drivers. During widespread disasters affecting an entire metropolitan area, all the medical facilities rely on local ambulance companies—and many rely on the same ambulance companies. It is important to understand whether a hospital has an exclusive contract with one or more transportation providers, or whether they will be dependent on ambulance companies that also serve many other facilities. Even if a hospital has an exclusive contract, a backup plan is required because these resources may not be available when an entire community is trying to evacuate simultaneously.

Actions by government officials may also affect the availability of contracted ambulance services. A hospital administrator called his ambulance contractor the day before Hurricane Katrina's landfall to move 12 ventilator-dependent patients to Lake Charles, but he was told that the mayor had taken control of all ambulances and the traffic was so bad that they would not likely get back and forth before the storm hit.41

FEMA has entered into a national contract for emergency ambulance services.42 The contract currently covers two regions (the Atlantic Coast and the Gulf Coast), providing up to 300 ambulances, 25 air ambulances, and paratransit vehicles to transport 3,500 persons per region. Requests for additional ambulances can also be made from other States through the Emergency Management Assistance Compact (EMAC).43 These additional ambulances coming from outside a disaster area may be more helpful in some hospital evacuations (e.g., prior to hurricane landfall) than in others (e.g., an earthquake). Because the ambulances may not be able to reach the scene quickly enough to assist in an immediate no-notice evacuation, hospitals must plan to be able to support temporary patient care outside of their physical facility until appropriate transportation capability arrives. Decision teams should also consider whether these ambulances and accompanying personnel will be equipped to transport neonates, children, and other special needs populations. The national ambulance contract includes all patients, regardless of age, condition, or special needs. Medical special needs patients are not covered in the U.S. Department of Defense's (DoD's) Memorandum of Agreement for aero-evacuation of patients via the National Disaster Medical System (NDMS), although the DoD is committed to assisting with the evacuation of all patients.44

Entry and Egress Points

A hospital evacuation requires road access to ramp-equipped hospital exits. Some hospitals are located in densely-built urban areas, or have only a few exits with ramps. With these types of capacity constraints, an orderly and safe evacuation could take days, especially if the rest of the city (and all of its hospitals) are also trying to evacuate or if there is poor coordination between the hospital and transportation providers. Poor coordination can lead to numerous vehicles waiting for patients to transport, or numerous patients waiting at the hospital exits for vehicles.

A pre-disaster self-assessment should estimate the number of vehicles that can be loaded at ramp-equipped hospital exits, whether there is a single loading area or if there are multiple loading areas accessible from different streets, and how long it might take to evacuate all wheelchair and bedbound patients through these exits. Some hospitals have an interim plan to bring patients to a location (perhaps outdoors) where they could then be loaded into vehicles more quickly—in effect a two-stage evacuation.43 Again, a self-assessment could ascertain how long it would take to move patients to this alternative location, and then how long it would take to load them all into vehicles for the second stage of evacuation.

Location of Care Sites Receiving Evacuated Patients

A hospital evacuation will be planned differently depending on whether the entire area is being evacuated or just one hospital. If just one hospital is evacuating (e.g., due to a fire inside that building), patients can be more easily dispersed among nearby hospitals. In most metro areas, this transport would be for a distance of less than 10 miles, and ambulances could cycle back and forth moving patients. An evacuating suburban or rural hospital may have to send patients farther away to appropriate receiving care sites. A self-assessment should determine how close the nearest appropriate hospitals are and what numbers of various types of patients—especially ICU and other specialty-care patients—those nearby hospitals can absorb, on an average day. Local or regional bed availability systems, if available, can assist with this task.

In a disaster that causes a widespread evacuation of health care facilities, transport destinations may include other States. Traffic-choked highways and lack of refueling stations could also slow the evacuation and prevent ambulances from cycling back for repeated evacuation trips. Evacuation plans therefore should anticipate the possible necessity of including aeromedical services in their patient transport and distribution systems. Again, pre-disaster planning and coordination with State EMS and emergency management agencies is critical for understanding available resources and for facilitating coordination with Federal agencies, should Federal assets be needed for the evacuation.

Whenever possible, patients must be transported to another health care facility with available services and staff appropriate to their medical needs. For example, a New Orleans hospital triaged patients by type, destination, and mode of transportation. ICU patients were evacuated to the triage area at the airport; psychiatric patients were medicated and sent by bus to a psychiatric hospital in Alexandria, Louisiana.41 There was no other hospital in Louisiana that could care for all of Children's Hospital of New Orleans' PICU patients, so these patients were transported out of State. Similarly, following the Three Mile Island (Pennsylvania) incident, neonates on life support were flown to Children's Hospital in Philadelphia, more than 100 miles away.

Approaches to Estimating Evacuation Time

Given variability in all the factors discussed above, there may be a wide variety of estimated hospital evacuation times with different sets of assumptions. At a practical level, decision teams should focus on perhaps two or three of the most likely scenarios for their setting (e.g., earthquakes in California, hurricanes in Louisiana) and attempt to estimate the most probable evacuation times. One such scenario is a planned and orderly evacuation of the typical mix of patients, with systems operating normally (e.g., all elevators are functional), and assuming that the hospital is the only one in the area that is evacuating. A variation of this scenario is to assume that other hospitals in the area are also evacuating.

Actual and Estimated Evacuation Times

When confronted by immediate, life-threatening circumstances (e.g., fire in the building), hospitals can be evacuated very quickly. Mt. Sinai Hospital in New York City evacuated hundreds of patients from one hospital tower to another in less than 1 hour.14 Pomerado Hospital in San Diego moved its 77 patients out of the building in 2 hours and 20 minutes.13 Following the Northridge (California) earthquake, six hospitals evacuated all their patients and transferred them to other facilities within 24 hours.3 Similarly, the evacuation of Children's Hospital of New Orleans was completed less than 24 hours after the evacuation decision was made.7 In May 2009, all 152 inpatients at Children's Hospital of Pittsburgh's Oakland (Pennsylvania) facility, which was in the process of being closed, were transported in less than 7 hours to a new hospital 2.5 miles away.45

Very large hospitals may take longer to evacuate. The evacuation of 406 patients from Memorial Hermann Hospital and Memorial Hermann Children's Hospital (Texas) required 31 hours.4 In addition, planners who have not faced an urgent evacuation may be unprepared for the speed with which patients can (or must) be moved.

Listed below are four possible approaches to estimating the time required to evacuate patients.

  • Experiences of other hospitals. The easiest—but least rigorous—approach is to estimate evacuation time based on how long it took other hospitals to evacuate and how closely those circumstances resemble the assumptions in the relevant planning scenario. Given that a host of factors affect evacuation times, this approach is not recommended. Still, data from other hospitals may offer some guidance.
  • Exercises. The results of evacuation tabletop exercises, adjusted and confirmed by data from drills, can inform evacuation time estimates.
  • Computer models. Simulations and other types of computer models are available for hospital planners. The AHRQ Mass Evacuation Transportation Model,38 for example, estimates the time required to transport patients from one or more evacuating hospitals to one or more receiving care sites. (It does not consider the internal characteristics of evacuating hospitals that would affect the time to move patients from their rooms to a staging area location.)
  • "Back of the envelope" calculations. An alternative approach for estimating the transport time is to estimate the number of round-trips required for each vehicle participating in the evacuation and the average round trip cycle time from staging area to receiving care site.

vi Boston's steam utility is Trigen, owned by Veolia Energy. Accessed December 2009.
vii An expert panel convened under an AHRQ-funded project studying the risk tolerance of early discharge concluded that patients whose risk of a consequential medical event ("unexpected death, irreversible impairment, or reduction in function within 72 hours of hospital discharge for which an in-hospital critical intervention would be initiated to stabilize or ameliorate the medical disorder") during the next 72 hours was 4 percent or less can be safely discharged.

Return to Contents
Proceed to Next Section


The information on this page is archived and provided for reference purposes only.


AHRQ Advancing Excellence in Health Care