METHODS

A comprehensive medical literature database search was performed using the MEDLINE/PubMed search engines. The investigators identified reports of patient care after disasters or major emergency incidents. Reports were queried for specific information on how the patients presented to or were referred to hospitals immediately after a disaster or significant event. Final articles were selected based on their reporting of how victims were transported to medical centers and hospitals after the incident. The time frame was limited to 25 years of the published literature, 1983-2008.

Articles were identified through a targeted key word search that identified articles that appeared to contain data relevant to our investigation. Key words included but were not limited to EMS, ambulance, paramedic, MICU, hospital, emergency department, triage, disaster, terrorism, MCI, self-referred, referral patterns, and transport. Special attention was paid to articles that described after-action reports, lessons learned, and emergency medical care in a postdisaster environment. Articles were specifically reviewed to determine whether the article reported the number of patients and the method of transportation of victims to hospital EDs. Victims who were explicitly not transported, in other words, patients who died at the scene, refused care, or were treated and released by on-the-scene personnel, were excluded from the study. The authors abstracted data from each manuscript and entered it into a Microsoft Excel spreadsheet (Microsoft, Redmond, WA). Basic descriptive statistics are reported in aggregate (Table 1).

TABLE 1 Categorization of Patients From Each Study ^a

Data were divided into 3 categories: transported by ambulance, self-transport, and other and were placed into each category in this analysis using the following parameters:

• • Ambulance: Patients entered into this category were specifically designated by the study authors as having been transported in an ambulance (air or ground) or specifically transported by emergency medical services.

• • Self-transport to the hospital: Any study that indicated that patients “presented themselves,” were considered “walking wounded,” or used “private vehicles” was placed into this category.

• • Other: This category was reserved for patients who did not ride in an ambulance (air or ground) or who explicitly did not meet the criteria for self-transport. Patients in this category arrived at hospitals mostly by bus, mass transit, or did not have the mode of transportation clearly delineated in the study but were certain to have been treated at the hospital.

RESULTS

A total of 166 reviews of disasters or major incidents were identified in the literature; 22 (13.3%) reported specific data on the transportation of casualties from the scene to hospitals (Table 1).Reference Carresi23Reference Hogan, Waeckerle, Dire and Lillibridge4Reference Teague5Reference Casey, Christy, Ellis and Wade11Reference Evans, Isgar, Bruins and Glover12Reference Greenwald, Rutherford, Green and Giglio13Reference Howells, Dunne and Reddy14Reference Karp, Sebbag and Peiser15Reference Kossmann, Wittling, Bühren, Sutter and Trentz16Reference Madzimbamuto17Reference Mohammed, Mann, Nawabi, Goodier and Ang18Reference Platz, Cooper, Silvestri and Siebert19Reference Richardson and Kumar20Reference van Harten, Welling, Perez, Patka and Kreis21Reference Wenck, Van Sickle and Drociuk22Reference Ağalar, Cakmakçi, Er, Akçakanat and Sayek26Reference Allister and Hamilton27Reference Beyersdorf, Nania and Luna28Reference Katz, Ofek, Adler, Abramowitz and Krausz29Reference Leiba, Halpern and Priel30Reference Raiter, Farfel and Lehavi31Reference Vilke, Smith, Stepanski, Ray, Murrin and Chan32 The studies included contained a heterogeneous cross-section of disasters from natural and human made to acts of terrorism (Figure 1).

FIGURE 1 Types of disasters by category.

The final analysis consisted of 8303 patients from 24 separate disasters or terrorist events. Nearly all of the case reports reviewed discussed the importance of the prehospital emergency care system. However, most suggest that only a fraction of the patients who are treated arrive at the ED via ambulance, particularly in the early postincident stages of a disaster.Reference Carresi23Reference Hogan, Waeckerle, Dire and Lillibridge4Reference Teague5Reference Casey, Christy, Ellis and Wade11Reference Evans, Isgar, Bruins and Glover12Reference Karp, Sebbag and Peiser15Reference Kossmann, Wittling, Bühren, Sutter and Trentz16Reference Madzimbamuto17Reference Mohammed, Mann, Nawabi, Goodier and Ang18Reference Platz, Cooper, Silvestri and Siebert19Reference Richardson and Kumar20Reference van Harten, Welling, Perez, Patka and Kreis21Reference Wenck, Van Sickle and Drociuk22 Overall, only 2980 (35.9%) of patients involved in these 24 disasters were transported to hospitals in ambulances; 5244 (63.2%) were transported to medical facilities via alternate means of transportation (Figure 2). Of the patients who were not transported in ambulances, 4813 (91.8%) transported themselves to hospitals for medical care and 431 (8.2%) sought other means of transportation to hospitals.

FIGURE 2 Patient data by transport category.

The categories most clearly defined across the studies were patients who were transported in ambulances and patients who transported themselves to hospitals. There were a variety of alternative means of transport to EDs in the 24 studies. These nonambulance methods were placed into the “other” category. Some studies described these methods, such as bus or coach transportationReference Teague5Reference Howells, Dunne and Reddy14Reference Karp, Sebbag and Peiser15Reference Mohammed, Mann, Nawabi, Goodier and Ang18Reference van Harten, Welling, Perez, Patka and Kreis21Reference Wenck, Van Sickle and Drociuk22; others did not describe these methods.Reference Hogan, Waeckerle, Dire and Lillibridge4Reference Casey, Christy, Ellis and Wade11Reference Evans, Isgar, Bruins and Glover12Reference Vilke, Smith, Stepanski, Ray, Murrin and Chan32

In general, patients who were not transported were placed into 1 of 3 categories: treated at the scene and released, died at the scene, or refused medical treatment or transportation. For the purposes of this analysis, these patients were excluded and only amounted to 79 (1%; Table 1).

COMMENT

Some assumptions are made by emergency planners about the referral patterns of patients to hospitals after a disaster, act of terrorism, or public health emergency. One assumption, as noted in previous investigations, is that if patients are acutely ill or injured, then they will arrive at hospitals, primarily by ambulance.Reference Auf der Heide6 Another common assumption is that when patients do arrive at hospital EDs, they will have been transported to facilities that can provide the appropriate specialty care that is needed, based on their injuries (eg, burn care, trauma care, pediatrics).Reference Auf der Heide6 The final assumption is that patients arriving by EMS-initiated transportation will have been appropriately triaged, decontaminated if necessary, and given medical treatment or stabilized before arrival.Reference Auf der Heide6 These assumptions are usually based on the perceived strength of the EMS system.Reference Bledsoe, Porter and Cherry1

Although during normal EMS system operations patients typically are triaged to receiving facilities based on their condition and any specialty care that they may need, during an MCI or disaster patients may need to be transported to facilities that lack certain specialty capabilities to deliver all of the victims to an acute care hospital. As a result, the hospital may need to alter its normal standard of care to meet the needs of patients with higher acuity or more complex clinical needs. In addition, patients who require decontamination may present to the ED. If a hospital is in close proximity to the incident location, then multiple patients may converge rapidly on the facility and special security and/or lockdown considerations may exist. Hospital emergency managers who are not planning for the likelihood of a mass convergence of self-referred patients from a disaster, act of terrorism, or public health emergency may be putting the safety and health of their staff, patients, and visitors in jeopardy.

In many locations, the EMS system is strong; however, our investigation has validated previous work showing that patients often disregard EMS and show up at the ED on their own.Reference Carresi23Reference Hogan, Waeckerle, Dire and Lillibridge4Reference Teague5Reference Auf der Heide6Reference Casey, Christy, Ellis and Wade11Reference Evans, Isgar, Bruins and Glover12Reference Karp, Sebbag and Peiser15Reference Kossmann, Wittling, Bühren, Sutter and Trentz16Reference Madzimbamuto17Reference Mohammed, Mann, Nawabi, Goodier and Ang18Reference Platz, Cooper, Silvestri and Siebert19Reference Richardson and Kumar20Reference van Harten, Welling, Perez, Patka and Kreis21Reference Wenck, Van Sickle and Drociuk22 Many victims who survive these events flee the scene well before the first arriving emergency vehicles and may well be on their way to the closest medical facilities. In fact, many will have minor injuries or illnesses that do not inhibit patients who are able to self-transport to hospitals after the event.Reference Carresi23Reference Hogan, Waeckerle, Dire and Lillibridge4Reference Teague5Reference Auf der Heide6Reference Casey, Christy, Ellis and Wade11Reference Evans, Isgar, Bruins and Glover12Reference Greenwald, Rutherford, Green and Giglio13Reference Howells, Dunne and Reddy14Reference Karp, Sebbag and Peiser15Reference Kossmann, Wittling, Bühren, Sutter and Trentz16Reference Madzimbamuto17Reference Mohammed, Mann, Nawabi, Goodier and Ang18Reference Platz, Cooper, Silvestri and Siebert19Reference Richardson and Kumar20Reference van Harten, Welling, Perez, Patka and Kreis21Reference Wenck, Van Sickle and Drociuk22 As a result, hospitals must be prepared to perform initial triage of “the walking wounded” (outside the hospital, in case of contaminated patients). In addition, facilities must have the ability to perform emergency decontamination of multiple victims, enforce tight perimeter security, including lockout or lockdown procedures when necessary, and create an area for “the worried well” and people with minor injuries to be cared for while awaiting the potential arrival of more critically injured or ill victims from the scene.Reference Auf der Heide6

In addition to the implications for hospital emergency planners, this study also has implications for incident command personnel at the scene of an event. Although scene controls, such as establishing a perimeter, delineating hot/warm/cold zones, estimating numbers of victims, and assigning kinds and types of resources, take time to be put in place, many ambulatory patients will leave the scene. The self-extrication of “walking wounded” victims may pose special challenges. It may be more difficult to first estimate the true number of victims from an event, when many may be actively leaving the scene. “Walking wounded” patients may be contaminated or infectious and may spread illness or cause secondary exposure to other people or locations. Mass dispersal of victims with minor illnesses or injuries may result in the loss of forensic evidence or make witness interviews after a criminal or terrorist event difficult. Epidemiologic investigations may also be affected by limiting the tracking of cases and ability to perform case and contact tracing. All of these reasons are important to emphasize scene security and perimeter controls as soon as possible when responding to a major incident.

There were several limitations to this study. First, our literature review initially returned 166 articles; however, only 22 reported complete data that were useable for analysis of this research question (Table 1). One reason for this low number of included studies is the underwhelming number of published case reports or after-action reports in the peer reviewed literature that reported, in sufficient detail, how victims arrived at hospitals after the event. A second reason was that many articles discussed transportation of casualties to hospitals, but they did not provide quantifiable data to report in this study. In addition, we purposely excluded articles published in non–English-language journals, which may have contributed to lower numbers for our analysis. Because these articles were not read for inclusion, it is impossible to estimate how many of them would have been appropriate for inclusion in this review. To remedy this in future analyses, we suggest that future investigators of emergency care and treatment of casualties postdisaster report data on how patients arrived at hospitals, trauma centers, and EDs. This may help provide a more accurate picture of the ways in which patients access emergency care after a disaster.

Second, the scope of our literature review was somewhat narrow—only 25 years retrospectively using only the PubMed/MEDLINE database. Because not all after-action reports appear in the medical peer-reviewed literature, the Lessons Learned Information System database operated by the Department of Homeland Security was also consulted, but it contained mainly executive summaries of incidents without actual numbers and dispositions of casualties. With such few investigators reporting useable data, this time frame may have been too short to yield enough data to generalize to all types of disasters. Future studies should look at longer time periods and multiple search engines to gather a more representative sample.

Third, because cases in which victims were transported by bus to hospitals for treatment were placed in the “other” category, we did not have enough information to determine whether this meant that patients chose to take buses to hospitals or whether EMS personnel arranged for ambulatory or “walking wounded” patients to ride in chartered buses for the purposes of mass transportation to medical centers. If this were part of the local EMS response, then these patients would have been triaged and possibly treated on the scene by prehospital personnel and hospitals notified of their impending arrival. Even if we assumed that all of the patients in the “other” category were EMS-initiated transports, then it would still result in only 40% of the disaster victims arriving via ambulance. This result would not change our interpretation of the results of our study; more victims of disasters who choose to seek immediate medical attention will arrive at hospitals through self-transport than via EMS or ambulance transportation.

Fourth, because it is, in general, easy to quantify the number of patients transported in ambulances to hospitals through EMS call logs and data available in patient care reports, there is no mechanism for reliable quantification of patients who self-refer to medical treatment after disasters. This is of particular concern when victims seek treatment at clinics, urgent care centers, or their primary care physicians and not hospitals. This may have resulted in an underreporting of patients in the self-transport category, biasing our results in favor of EMS-initiated transport mechanisms to hospitals and medical centers.

CONCLUSIONS

Our study suggests that after a disaster, act of terrorism, or public health emergency, victims most often present to hospitals and EDs through self-referral rather than via ambulance transportation, validating previous work in this area. The implications for hospitals should include the development of emergency procedures and preparedness plans that address the complications associated with a mass influx of walk-in or self-referred patients after a disaster. Emergency procedures and preparedness plans should include staff 's proficiency in triage, ability to perform emergency decontamination, and safety and security procedures should they encounter a patient surge at their facility. Future studies should attempt to report how victims reach emergency care in disasters so that we may better understand referral patterns after mass casualty events.