An uninterrupted power supply is a mission-critical resource for hospitals and operating rooms. Planning for power failure is imperative. If a health care facility has a power failure, the post hoc assessment of the organization's response can be of immeasurable value toward improving preparedness for similar events in the future. These principles are illustrated by a partial power failure in our institution that affected our operating rooms.
During the middle of the day, the emergency power system failed at an academic hospital facility. The nonemergency power system was unaffected. The primary cause of the power failure was construction. During testing and commissioning of old transfer switches onto a new generator, a phase loss relay was removed and replaced. This device senses whether or not adequate current is being delivered from a primary power source. If it detects a problem, it activates a sequence of events that convert the system to backup power. Later in the day, unobserved, this relay fell out of its mounting socket causing a break in the circuit. The critical branch transfer switch, which transfers the connection from primary power to the backup generators, redirected the power input to a generator that was being serviced instead on one that was active. The result was a failure of the emergency (red outlet) system (Fig. 1).
Some initial confusion in communication caused resources to be directed toward one ward. It took some minutes before the electricians recognized that multiple sites were out, including the operating rooms, and the power generators were the most likely location of the problem. When they arrived at the correct site, they reinserted the phase loss relay, which immediately triggered the resumption of input from the primary power supply. The problem went uncorrected for approximately 15 minutes.
The loss of emergency power was widespread throughout the main hospital, including the inpatient operating rooms. Eight of 15 operating rooms had ongoing cases at the time of the event including a craniotomy, Whipple procedure, and kidney transplant, but no cardiac surgical procedures. The responses of the anesthesia personnel varied. Most providers maintained ventilation by switching to manual ventilation through the anesthesia machine. One patient had a laryngeal mask airway in place and was ventilating spontaneously. Two providers continued mechanical ventilation, relying on the backup power from the anesthesia machine.
Anesthesia maintenance was adjusted as necessary. One provider was already administering total IV anesthesia for a malignant hyperthermia-susceptible patient. Two providers changed from desflurane (one to sevoflurane and one to isoflurane), correctly believing that the power to the desflurane vaporizer was unreliable. One provider switched to a propofol infusion and gave 1 mg midazolam. The other providers maintained anesthesia with isoflurane or sevoflurane. There were no known cases of awareness.
All of the anesthesia machines, Aestiva/5 (GE Healthcare, Madison, WI), converted to backup battery power. According to the online product specifications, the backup time of the machine under typical operating conditions is 45 minutes when the battery is fully charged.1 An e-mail communication (May 2006) from a GE representative stated that one could expect a battery life of 30 minutes if using mechanical ventilation but substantially longer if hand-ventilating with display of tidal volume, respiratory rate, etc.
The outlets on the back of the Aestiva/5, which were used to power the anesthesia monitors, are not supplied by the backup battery. The battery backup for the desflurane vaporizer (Tech 6, GE Healthcare) only powers the no-output alarm and low-battery alarm. Loss of electrical power to the Aestiva/5 also terminates the output from the desflurane vaporizer.
Monitoring was interrupted for all but one provider, whose anesthesia machine was inappropriately plugged into a nonemergency outlet. Five providers requested and quickly received a portable monitoring unit. The other 2 providers monitored their patients with a hand on the pulse. They believed that their patients were stable and did not want to use portable monitoring resources that might be needed elsewhere.
All operating rooms, except for the magnetic resonance imaging operating room, had flashlights. All personnel who attempted to use the flashlights found that they were functional. One provider had natural light streaming through the windows. Another provider used her laryngoscope for emergency lighting. Surgical field lighting, which is on the emergency power grid, was lost. Surgical cautery was inoperable. Ambient room lights, half on the nonemergency power system, continued to function normally. Emergency battery backup lighting was not properly deployed in all locations. In one room, the ambient lights had been turned off for surgical purposes, and the clinicians were completely in the dark. All rooms now have portable backup lights triggered by failure of the emergency outlets. The oxygen pipeline supply was uninterrupted, as was the vacuum system. The telephone system was intact. In most cases, surgery paused during the power failure.
Automated drug supply cabinets were inoperable for an extended time during and after the power failure. When exposed to a prolonged outage, they do not automatically reboot and become available for service. Although these cabinets are essential after hours and on the weekends, the operating rooms have a satellite pharmacy staffed with pharmacists. These pharmacists directly supply the majority of drugs needed by anesthesia providers and nurses. They also have a key to the automated drug supply cabinets and were able to access additional drugs despite the power failure. Patient care floor nurses had to fax orders to the pharmacy, resulting in delays for those patients. There was no adverse impact for patients in the operating rooms, recovery room, or perioperative holding.
An institutional-wide post hoc evaluation of the event was made and a number of areas for improvement were identified. Poor communication was the most notable problem. Information about the planned test of the generators was not communicated appropriately. The dispatch office was located remotely from the site of testing, impeding free communication. Once the problem was identified, there was inadequate communication to stakeholders. The paging system became backlogged, further slowing communication.
Anesthesia Services maintains 15 portable monitoring units, a sufficient number for all of the operating rooms to each have one if needed in an emergency. However, the units are often scattered throughout the hospital after monitored transports are completed. It is unknown how many were physically in the operating room environment during the power failure, but there was no reason for the 2 sites to decline offers of portable devices.
The time a power failure occurs can affect an organization's response. Fortunately, the operating rooms at midday during the week were fully staffed with an adequate number of anesthesia and nursing personnel to respond to the crisis.
Several protocol changes were made after the evaluation was complete. Although the root cause of this power failure was new construction, the very unique incident of the switch falling out of its mounting socket was remedied with “keepers” or seatbelts to secure the switches to prevent accidental disconnects. A protocol was established to improve communication with batch paging of key personnel for emergencies. All future utility testing or planned service interruptions require personnel from the dispatch control center to be present at the site of the testing and increased staffing for the duration of the testing. A maintenance plan for all uninterrupted power supplies throughout the hospital was initiated.
The Joint Commission on Accreditation at Healthcare Organizations has set standards for emergency electrical power systems.2–4 They have also provided risk reduction strategies for ensuring power during new construction, a time of particular vulnerability. Their recommendations for clinical contingency plans for provisions5 include rapid deployment of battery-powered equipment (monitoring, suction units); ensuring critical equipment is plugged into emergency power outlets; providing a backup method for accessing electronic patient care data; ensuring access to automated drug supply cabinets; and providing access to emergency supplies such as 2-way radios, flashlights, and extension cords.
Yasny and Soffer6 have provided recommendations for the management of a patient in the case of a power failure. The ABCs are, of course, paramount. In addition, one should call for help, utilize emergency equipment (flashlight, battery-powered monitoring, and airway equipment), and ensure delivery of anesthesia to the patient. Some reports describe accounts of a case-specific response during cardiopulmonary bypass.7,8 Other reports focus on specific causes of the failure such as water shorting a circuit9 or construction severing the main power line and the failure of one emergency generator causing an overload and failure of the second generator.10
Power loss to only the emergency power outlets has been reported.7 Essential equipment should ordinarily be plugged into emergency power outlets. However, if some equipment in the operating room is continuing to function in the face of power failure to essential equipment, it would be worthwhile to consider plugging that equipment into nonemergency power outlets.
It is important for the care team to discuss whether it is best to continue or abort the current procedure. Key data pertinent to this discussion are the degree to which the power is adversely affected and the anticipated time of the return of power. This information, however, may be unknown to the care team for an extended period of time. Further considerations include patient stability, the anticipated duration of the remaining surgery, and the feasibility of ending the surgery partially completed and returning at a future time. A pause in the surgery is usually indicated while assessing the impact of the failure. If the power has failed throughout the hospital, a decision must be made as to whether a transfer will make the patient safer, considering the risk of transport and the level of safety at the destination. Pagan et al.11 described a power failure from an electrical fire, which required evacuation of the hospital.
An uninterrupted power supply is a mission-critical resource for hospitals and operating rooms. Planning for a power failure is imperative. This includes an inventory of systems that would be affected. Table 1 lists systems relevant to the operating room that may fail during a power outage. Finally, if a health care facility has a power failure, the post hoc assessment of the organization's response can be of immeasurable value in preventing future occurrences.
1. GE Healthcare. Product Specifications Aestiva/5 Anesthesia Machine, 2006
2. NFPA 70, National Electric Code. Quincy, MA: NFPA, 2002
3. NFPA 99, Standard on Healthcare Facilities. Quincy, MA: NFPA, 2002
4. NFPA 110, Standard for Emergency and Standby Power Systems. Quincy, MA: NFPA, 2002
5. JCAHO. Preventing adverse events caused by emergency electrical power system failures. Sentinel Event Alert 2006;37:1–3
6. Yasny J, Soffer R. A case of power failure in the operating room. Anesth Prog 2005;52:65–9
7. Hargrove M, Ramish BC, O'Donnell A, Aherne T. Electrical failure during cardiopulmonary bypass. Perfusion 2002;17:369–72
8. Troianos CA. Complete electrical failure during cardiopulmonary bypass. Anesthesiology 1995;82:298–302
9. Tye JC, Chamley D. Complete power failure. Anaesthesia 2000: 55;1133–4
10. Nixon MC, Ghurye M. Electrical failure in theater—a consequence of complacency? Anaesthesia 1997;52:88–9
11. Pagan A, Curty R, Rosemarie RN, Rodriguez MI, Pryor F. Emergency-total power outage in the OR. AORN J 2001;74: 514–5