The literature has abundant examples of natural or human-generated disasters as well as the hospital preparedness and related training necessary to face those kinds of events. Less frequent are reports on in-hospital incidents and their consequences, in particular, lightning strikes. Most of these incidents are reported in local newspapers only. We describe here a lightning strike in the proximity of a university hospital that disabled all landlines and paging communications in 2007, the way the incident was managed, and the lessons to be learned.
Narrative
The Lausanne University Hospital is a tertiary medical center with about 1500 beds. It is located within a city of 140 000 inhabitants in the western part of Switzerland and serves a population area of 300 000. The hospital is a complex with 6 main units (Figure 1). In 2007, telecommunication within the hospital relied on copper (coax cables) landlines and a paging system, including the in-hospital Cardiac Arrest Team Call System. There was no internal mobile phone network, except for private ones (personal ownership), which were only partially effective because of the lack of internal antennas.
Figure 1 Map of the Bugnon Hospital Campus.
At that time the hospital already had a dedicated internal incident and disaster plan as well as a command-and-control organization staffed by team members who received alarms by the paging system. The hospital also had special landlines that were not part of the hospital network routing (redundant); these were based on a different central exchange and were effective between the hospital’s central call center, the hospital’s central command control center, the state’s command and control center, local firefighters, and police services.
On May 21, 2007, at 8:29 pm, a storm hit the city of Lausanne and its surrounding districts with 220 lightning strikes over a period of 40 minutes. One struck close to the hospital. The chronology of the incident is detailed in Table 1.
Table 1 Chronology of the Lightning Strike Incident
Discussion
The hospital was not directly struck by lightning, such as the event previously reported in Belgium.Reference Mortelmans, Van Springel and Van Boxstael 1 However, the electromagnetic wave after the strike seriously damaged all 3 telephone exchange centers (indirect impact), burning 25 sensitive connective cards; thus, 95% of the 8500 hospital phones were out of order. The wave did not affect the power supply or the computer network. The state emergency dispatch center, which used the same technical center for communication, had 2 lines that were not seriously damaged by the incident. Neither the hospital nor the telecommunications provider had sufficient replacement cards on hand to repair the hospital network. The time of the incident, during the evening, is typically when activity in the hospital is less intense.
One critical situation occurred in the pediatric ward, when a patient had a massive hemorrhage 4 days after a tonsillectomy. The nurse in charge of the patient sent all staff present in the ward to get help (on-call pediatrician, on-call anesthetist), to the blood bank, and to inform the operating room to get ready. The situation was immediately life threatening. There was an increased delay of 15 minutes to take the patient to the operating room but no adverse consequences for the patient.
What Are the Necessary Options and Requirements to Handle Such an Incident?
Most hospitals are equipped with lightning rods; however, these structures are helpful only in the case of a direct lightning impact or strike (3% of incidents). In our situation, the incident was caused by an electromagnetic wave after a distant strike (such as in 75% of incidents) and propagated through the multiplicity of copper phone-line connections between the hospital buildings and the central exchange center. For hospitals still using this technology, the necessary options and/or requirements to face such a risk are (1) keep the copper lines and equip every phone (or group of phones) with protection against overtension, which is very costly; (2) review the copper line network to avoid telephone line curls, though hospital networks are so dense that this may not be a realistic solution; (3) migrate the copper network to a computer network (with fiberglass optics), which would not be conductive and, thus, would be resistant to lightning strikes and electromagnetic waves; or (4) do nothing. The additional advantage of the migration from copper to the fiberglass option relies on the redundancy of the optical fibers between buildings and between floors.
Cost of the Incident
The cost of the incident was CHF 230 000 (about US $191 800) for the hardware and the technical services from the private telecommunications company. The cost for additional security personnel from a private company was CHF 13 000 (about US $10 800). Therefore, the total cost was CHF 243 000 (US $202 600), but this amount does not include the additional working hours of the hospital staff. In France, for example, lightning is responsible for 11% of all of the costs reimbursed by insurance companies to hospitals after a similar unlikely event.Reference Goffion 2 This incident confirms the need for lightning protection, which should be considered every time a new hospital is being built.
Although such an action is not an absolute form of protection, a lightning-rod network can transform a building into a Faraday cage, thus lowering the risk of damage in case of a direct strike. This installation should not be compromised by installing additional high antennae (radios, telecommunication). Fiberglass optics and mobile phone technology, today’s standards for new construction, drastically reduce the risk related to electromagnetic waves.
Interestingly, a nearby hospital was directly struck by lightning shortly before our incident. The strike did not affect telecommunications but caused a fire within the building that was quickly managed.
Lessons Learned
After the lightning-strike incident, the hospital decided to migrate the copper network to a computer network connected with fiberglass optics. As a residual risk persisted, a second decision was to deploy a redundant parallel telephone system: an internal low-power mobile phone network that replaced the paging system. The two networks (the fiberglass optic and the mobile phone network) are interoperative but are not interdependent. During an incident or during maintenance tasks, a switch from one network to the other is possible and effective.
Fire, loss of power, and the complete or partial loss of telecommunication are major risks for hospitals. Plans should exist for such events, as they do for multiple-case-casualties and nuclear, radiological, biological, chemical, and explosive (NRBCE) disasters, and staff should be fully trained for any of these internal “act of God” circumstances. In our hospital, a specific plan for a telecommunications breakdown did not exist at the time of the incident. Since then, the control command team, led by the head of the emergency department, has undergone multiple training sessions to assess the time needed to gather its members and to test its efficiency in multiple scenarios. The team has also functioned during real incidents, such as the arrival of multiple casualties. Since the telecommunications breakdown in 2007, we have been confronted many times with landline or mobile telecommunications failures, either during maintenance work or unannounced incidents. On every occasion, 1 of the 2 telecommunication systems was functional. As a redundant measure during those incidents, we also deployed more security personnel with radios to secure sensitive communications, in particular between rapid response teams, intensive care units, operating rooms, and the emergency department.
When dealing with disasters or major incidents, hospitals mainly focus their training on surge capacity and the arrival of multiple casualties. In-hospital incidents deserve as much attention and training. Testing the control command team operation under internal hospital incidents is easier to organize and not as costly as simulating the arrival of multiple casualties and should be done at least once a year.
Acknowledgments
The description of this case is based on hospital incident reports and interviews with key persons present in the hospital on that night. We thank Stephan Misteli and Jean-Luc Mayor from the information technology department; Laurent Meier from the security department; Mirjam Martinez, registered nurse from the pediatric department; and Dr Bertrand Yersin, head of the emergency department, all of whom were present the night of the incident, for their testimony.
