Disasters in Germany and France, 2006–2016

The distribution patterns for the types of most reported events (storms, extreme temperatures, floods, transport accidents) indicated a similar risk pattern for injuries. Casualties of 259 and 4973 were reported for disasters in Germany and France, respectively. Whereas in Germany most lives (96) were lost in transport accidents within the last decade, in France, most deaths were due to the heat waves of 2006 and 2015 (1388 and 3275 casualties, respectively), which indicate that heat waves struck France more significantly than Germany during that time.

Disasters in Germany and France Between 1900 and 2016 and Most Significant Events as Registered in the EM-DAT

We explored the overall pattern of events for Germany and France available in the database, which covered the years 1900 to 2016. The EM-DAT reports that 409 natural or technological disasters affected Germany and France, leading to the loss of 44 690 lives. Most single events in both countries were storms, transport accidents, and floods. Extreme temperatures, industrial, and transport accidents claimed most lives. The three single events with the highest death toll in Germany were the heat wave in 2003 (9355 casualties), the explosion of a chemical factory in 1921 (600 casualties), and a mining accident in 1946 (439 casualties), and in France the heat waves of 2003, 2015, and 2016 (19 490, 3275, and 1388 casualties, respectively). Otherwise, the distribution of events and casualties between Germany and France appeared to exhibit a similar pattern. A major limitation of this long time EM-DAT analysis is the finding that the historical data entries may not be complete. Specifically, the 1918–1919 influenza pandemic, which caused an excess mortality of 237 509 in France between August 1918 and April 1919, and 426 574 between March 1918 and January 1919 in Germany, is not captured in the EM-DAT.Reference Ansart, Pelat and Boelle¹⁹ Furthermore, there is no data entry for the other three major influenza pandemic waves in France and Germany of 1957–1958, 1968–1970, and 2009, which caused the loss of 29, 100, 46 900, and 350 lives alone in Germany.Reference Buchholz, Buda and Reuss²¹ In addition, as illustrated in Supplemental Figure 2 and Supplemental Figure 3, there are fewer events reported before 1988 compared with after 1988, the launch date of the database. Issues with historical data are common in registry databases, including registries for diseases; missing data, recall, and ascertainment bias therefore limit generalizability.Reference Mehta, Beck and Elliott^22, Reference Muenzer, Jones and Tylki-Szymanska²³ There are, however, situations where retrospectively collected data allow helpful insight into issues that cannot be quickly investigated otherwise, especially when prospective studies are not feasible or would require a long time to collect sufficient data.Reference Mechler, Mountford, Hoffmann and Ries^24, Reference Zielonka, Garbade and Kolker²⁵ Therefore, the supposedly incomplete historical data in this analysis may serve as a helpful orientation but should be interpreted with caution.

Pediatric Data – the Blind Spot

Information in the EM-DAT was not stratified for pediatric data. Therefore, the pediatric interpretation hereafter is deducted, rather than being based on age stratified subgroup analyses, which would have been the ideal situation and allow data-driven decision making in pediatric disaster epidemiology. The present data of disaster patterns in the EM-DAT suggest that heat and cold (extreme temperatures) and trauma (accidents, explosion, storms, moving elements) may represent an important risk to children in disasters in both countries. In addition, exposure to chemicals (explosions, industrial accidents), water (flood), and infectious agents (epidemics) are potential mechanisms of disaster-related injuries (Table 3). Although not specifically captured in the database, diseases due to heat waves would include heat stroke, heat exhaustion, and dehydration.Reference Basu^{26, 27} Substantial psychological burden is to be anticipated. Exposure to stressful life events may lead to PTSD following a complex etiology that is mediated by many factors, for example: (epi-) genetic, familiar, trauma characteristics, subjective, pre- and post-trauma variables.Reference Trickey, Siddaway and Meiser-Stedman^28, Reference Kilic, Ozguven and Sayil²⁹ Moreover, there is evidence that depressive disorders are more common resulting mental disorders after natural disasters.Reference Ying, Wu, Lin and Chen³⁰ Yet, scientific evaluation of various disasters in the last decades has shown that the majority of exposed children did not develop PTSD or disaster-related mental disorders.Reference Green, Grace and Vary^31–Reference Lengua, Long, Smith and Meltzoff³⁴ This can be attributed to the construct of resilience, which is widely researched but still poorly understood.Reference Hjemdal, Vogel and Solem^35–Reference Fayyad, Cordahi-Tabet and Yeretzian³⁹ We favor an age-appropriate, transparent communication with children about risks and measures of prevention and mitigation. This should occur in a confident setting (eg, school) and whenever possible with inclusion of trusted persons, for example, parents and teacher. Playful educational online tools also may be helpful.⁴⁰

TABLE 3

Selected Mechanisms of Injuries in Natural and Technological Disasters and Potential Preventive Measures

Rothstein reviewed strategies of pediatric care in disasters and proposed top 10 priorities in the emergency phase of a catastrophe (ie, initial assessment, measles immunization, water and sanitation, food and nutrition planning, shelter and site planning, health care in emergency phase, control of communicable diseases and epidemics, public health surveillance, human resources and training, and coordination).Reference Rothstein⁴¹ Likewise, one should keep in mind that the seven sins of humanitarian medicine (ie, leaving a mess behind, failing to match technology to local needs, failure of NGOs to cooperate with each other, failing to have a follow-up plan, allowing politics or training to trump service, going where we are not wanted or needed, and doing the right thing for the wrong reason) are to be avoided.Reference Rothstein^41, Reference Welling, Ryan, Burris and Rich⁴² This implies that the preventive and interventional approach toward pediatric care in disaster situations should be timely, transparent, coordinated and sustained, and ideally be data-driven. From an epidemiological and logistic perspective (eg, type of food needed, immunization, dehydration risk, drug supply including galenic formulation, surgery and anesthesia equipment, and need of infant transportation equipment), it would be important to know which age groups are specifically affected.

Challenges in taking care of children in disaster situations occur on four main levels. First, unaccompanied children are vulnerable. This is of relevance because geographical dynamics that change on a circadian pattern – children attending daycare or school and parents going to work – may lead to a separation of families in case of an unexpected event.⁴³ Lessons learned from the 2010 earthquake in Haiti were not to separate families in case of a disaster and to quickly establish communication channels between family members.Reference Auerbach, Norris and Menon⁴⁴ Second, the physiology of the pediatric organism changes with age from neonates at adolescents.Reference Kearns, Abdel-Rahman and Alander⁴⁵ This leads to particular vulnerabilities, such as the increased susceptibility to cold and the risk of rapid dehydration in neonates, infants, or toddlers. Children are more vulnerable to chemical contamination in the air, because their respiratory frequency is higher than in adults.Reference Markenson and Reynolds⁴⁶ Children may be less resilient to seasonal temperature changes and living conditions due to loss of shelter after evacuation, as experienced in Japan in 2011.Reference Ishii and Nagata⁴⁷ Furthermore, psychological or pharmacological interventions must be age appropriate; in addition, pediatric pharmacology deserves particular consideration for the availability of medical supplies and stockpiling.Reference Markenson^9, Reference Kozu and Homma^{48, 49} Third, disease-inherent dynamics can be different in children compared with adults, and chronic medical problems may exacerbate in a crisis situation.Reference Mechler, Mountford, Hoffmann and Ries²⁴ Therefore, the availability of personnel with training in pediatrics is of advantage.Reference Kozu and Homma^48, Reference Gnauck, Nufer and LaValley^50, Reference Weiner, Manzi and Waltzman⁵¹ Fourth, pediatric surgical expertise is essential for children in conflict and disaster situations.Reference Trudeau and Rothstein⁵² Orthopedic injuries like fractures or amputations, neurosurgical lesions and burns, merit special pediatric knowledge. Besides, pediatric perioperative care and safe anesthesia are required for survival and successful outcome.

Census data may be of value in identifying populations-at-risk, as recently demonstrated in Japan after the Great Eastern Japan earthquake on March 11, 2011; although, as in many disaster analyses, pediatric aspects were not specifically considered.Reference Ishiguro, Togita and Inoue⁵³ A Web-based centralized tracking system with the goal to reunite families as soon as possible was suggested as a lesson from the events in Japan.Reference Yonekura, Ueno and Iwanaka⁵⁴ An analysis of impact and long-term health effects of the disasters at Seveso, Three Mile Island, Bhopal, Chernobyl, the World Trade Center, and Fukushima concluded that health surveillance and treatment programs for affected responders and residents, including children, are important. Furthermore, detailed emergency preparedness plans are critical to mitigate future threats.Reference Lucchini, Hashim and Acquilla⁵⁵ In 2010, Starmer et al. emphasized the need to mitigate possible disasters by community and societal preparation by inclusion of families and children in the preparation to respond and through access to support services.Reference Starmer, Duby and Slaw^56, Reference Khorram-Manesh⁵⁷ Ideally, pediatricians and families should actively be involved in governmental advice for disaster planning.Reference Starmer, Duby and Slaw⁵⁶

Last, but not least, appropriate education of pediatric health care professionals is crucial for disaster resilience. Although education in pediatric disaster medicine is usually not a core component of medical school curricula or specialty training in pediatric residency programs, closing knowledge and competency gaps are important.Reference Khorram-Manesh, Lupesco and Friedl^58–61 In the United States, the National Center for Disaster Medicine and Public Health put forward an initiative to develop such a training program.Reference Siegel, Strauss-Riggs and Costello⁶² An update on the current state of this work in progress was published by Siegel et al. in 2014.Reference Siegel, Strauss-Riggs and Needle⁶³

Limitations and Directions for Future Research

This analysis is based on events registered in the EM-DAT because (1) this database is a major source of epidemiological information on disasters and (2) the database is easily accessible. The present analysis relies on the accuracy of the reported data. Detailed original source data verification was logistically not feasible and was therefore not undertaken. There are other databases available, such as NatCatSERVICE (Natural catastrophe know-how for risk management and research, owned by the insurance company Munich Re, Germany), Sigma (catastrophe database, owned by the insurance company Swiss Re, Switzerland), GLIDE (Global Identifier number, Asian Disaster Reduction Center, Japan), DesInventar (Sistema de Inventario de Desastres, Corporación OSSO, Columbia, and La Red de Estudios Sociales en Prevención de Desastres en América Latina, Panama), and SHELDUS (Spatial Hazard Events and Losses Databases for the United States, Hazards and Vulnerability Research Institute, University of South Carolina, USA), but these are either less comprehensive or do not cover both countries of interest in this report.¹⁶ As outlined previously, historical data in the database may be less accurate than the prospectively collected information due to ascertainment bias, illustrated by the finding that the 1918–1919, 1957–1958, 1968–1970, and 2009 influenza pandemic wavers with their substantial mortality were not covered at the time of data query. The impact of armed conflicts is not taken into account in the EM-DAT. We speculate that this may introduce a bias because the outcome of disasters in wartime may be worse. It is possible that resource constraints or decreased societal resilience in wartime may lead to higher infant mortality, hunger, and malnutrition in children. The major issue, from a pediatric perspective, is the fact that mortality data reported in the EM-DAT were not separately available for adults and children, which might introduce an age bias into this analysis if a certain age segment of the overall population has either a particular vulnerability or a specific resilience to certain injuries. Moreover, capturing pediatric information in the EM-DAT would allow better data-driven decision making for children in catastrophic situations in the future.