Hospitals, and their EDs in particular, serve an important function in the community as wardens of public health in times of crisis and disaster[23]. IHCDs can directly cause physical harm and the sudden closure of a hospital or ED may force the diversion of patients to distant hospitals where patient records are lacking [24]. We have shown that hospital and departmental closures have occurred 134 times in the Netherlands with an increasing trend across a 20-year study period. Natural disasters rarely occur in the Netherlands and did not cause an IHCD in our dataset, with the exception of 4 incidents caused by inclement weather. Cascading events of multiple sequential failures occur 29.1% of the time where ultimately a hospital or department becomes disabled. EDs are often involved both directly and indirectly with 71% of IHCDs incurring an ED closure.
ICT Failures are an Increasing Trend:
The 20-year-period studied showed an average of 6.7 IHCDs per year. Alarmingly, we have shown that the incidence of these events is increasing and are in large part explained by the increase in ICT failures. This should not come as a surprise considering that the healthcare sector has seen increases in the use of digital technologies in communication, diagnostics, and treatment applications[25]. The frequent occurrence exposes a weakness in hospital preparedness and business continuity that must be addressed. Malfunctioning of backup servers, impossibility of quickly reverting to paper dossiers at an instant’s notice and general lack of training were problems mentioned in the news articles. While our search did not uncover any reports of cyberattacks on hospital networks, there is previous experience with ransomware affecting hospital networks in the United Kingdom resulting in multiple, prolonged ED closures [26]. The increasing trend in ICT disturbances make cybersecurity and digital continuity issues that should be prioritized in hospital disaster preparedness[27]. Furthermore, our data corroborates a recent Dutch governmental report detailing an increasing trend of hospital ICT failures and their potential impacts on patient safety[28].
Cascading Events
Hospitals are complex organisations and structures with convoluted interrelated systems [29]. Hospital resilience in mitigating internal crises appears to be habitually challenged by multiple consecutive failures in the form of cascading events. Nearly a third of IHCDs express this trend (Table 2). Hospitals customarily seem to be able to absorb one “hit” to their internal infrastructure, such as a simple external power failure where the auxiliary power seamlessly starts and the hospital continues to function. Conversely, if somewhere in that succession of events an additional failure, such as a technological failure (i.e. short circuit) occurs hindering the emergency power sources, a hospital fails to be able to provide lifesaving services. Power outages often led to failures involving the computer networks, and technological failures also tend to lead to equipment malfunctions that in turn lead to fires and hazardous material releases. Knowledge of these predominantly occurring sequential events are beneficial for attenuation of safety checks in disaster preparation plans. Examples would include frequent auxiliary power tests, training of personnel to bypass problems with power and the creation of emergency procedure guides in the face of ICT failure.
Hazardous Materials
Hospitals are fertile grounds for disaster as they are packed with flammable gasses, toxic substances, biological sources and radioactive materials. Patients injured by these same types of substances ideally present to the ED in a decontaminated state. However, this is not always the case, endangering hospital employees or prompting evacuation. Hospital evacuations occurring from both threatened and actual hazardous material exposure, or release typically come with little warning or time to prepare, often occurring in EDs [30]. Between 1971 and 1999, 18% of U.S. hospital evacuations were due to hazardous materials[31]. Comparatively, we found 19 incidents occurring in the Netherlands in the past 20 years, each necessitating evacuation (33.9% of total evacuations). High consequence infectious diseases and emerging infectious diseases such as the 2019 Novel Coronavirus (COVID-19), also pose significant risks to hospitals and EDs, where hospitals need be able to safely treat infected individuals while maintaining surge capacity[32]. Unprepared hospitals receiving these types of patients lead to dangerous situations, panic and eventual evacuations. Our dataset entails 2 ED closures lasting several hours due to possible Ebola viral disease infection. Unpreparedness and no prior notice of patient arrival caused considerable delay while the emergency services scrambled to collect appropriate protective equipment and protocols.
Hospital Characteristics
Our data has shown that hospitals in the Netherlands are susceptible to sudden closure due to IHCDs irrespective of trauma level or academic specialties. The acute closure of academic hospitals and level 1 trauma centres places a large burden on emergency medical services by increasing the travel distance to an appropriate medical centre, or the need to transport severely injured patients to underprepared centres. Critical access hospital closures would exacerbate these problems in more rural settings. Closer hospital proximities and multiple trauma centres in larger urban areas make this problem less evident.
Advice to ED and Hospital
Hospitals shelter a high number of vulnerable and dependent individuals, sustained by high densities of medical and support staff. Risks and hazards to potential disasters must be made apparent and planned for. As part of a hospital’s accreditation in the Netherlands, a hospital disaster plan must be available, and the staff must be trained to act in extenuating circumstances to both external and internal sources of disruption [33, 34]. The most common types of internal disaster plans only encompass computer system failures, power failures, and fires [35]. We have enumerated the possible types of IHCDs having transpired in the Netherlands, which can be extrapolated and applied as potential hazards to the majority of modern healthcare facilities in developed countries. If these internal disaster plans are revised in line with our findings, this will have the potential to improve hospital recovery in times and possibly prevent unnecessary closure. Hospital management officials and emergency physicians should attenuate their hospital disaster plans based on the incidents we have enumerated to prevent future closure. Creation of generic crisis response templates based on the most common disaster and cascading events will allow hospital coordinators to create trainings to show employees how the various hospital systems are organised and how to respond to specific failures [29]. Training of hospital staff, notably ED staff, should occur frequently to ensure seamless activation of disaster plans. Furthermore, a unified body that registers IHCDs nationally and that acts as a hospital disaster preparedness expertise centre would be a valuable medium where learned lessons can be shared with other medical centres.
Limitations:
There are several limitations in our search method that may have led to under ascertainment of total incidents. Due to the subjective nature of press and news releases, bias towards sensationalistic and newsworthy events may have arisen, leaving out smaller and less impactful incidents. However, we deem it likely that events in which a department had to be closed urgently would have made the news. Underreporting of incidents may have occurred in earlier years due to inaccessibility of articles and decreased number of reporting news outlets (i.e. now more online news articles than print versions). Furthermore, we have only included acute care hospitals and our results cannot be applied to specialised clinics and hospitals. Potentially valuable information on the exact duration of department closure was not available in all incidents. Nonetheless, the sources and databases used are the best available. This is to our knowledge the first study utilizing this search methodology to characterize IHCDs, but can be applied to other regions in the world. Therefore, our classification system of IHCDs from news sources has not been validated. This study was geographically limited to the Netherlands, though we believe that the incidence and failure types to be applicable to all modern hospitals with the exception of natural disasters more common in other regions of the world.