ORIGINAL ARTICLES

In-Hospital Resource Utilization DuringMultiple Casualty Incidents

Sharon Einav, MD,* Limor Aharonson-Daniel, PhD,† Charles Weissman, MD,‡Herbert R. Freund, MD,§ Kobi Peleg, PhD, MPH†, and the Israel Trauma Group�

Objective: To suggest guidelines for hospital organization duringterror-related multiple casualty incidents (MCIs) based on the ex-perience of 6 level I trauma centers.Summary Background Data: Most terror-related MCIs are bomb-ings. The sporadic nature of these events complicates in-hospitalpreparation.Methods: Data were collected at all level I Trauma centers during/after MCIs for the Israel National Trauma registry. Patients wereincluded if they were admitted or died in hospital following injury insuicide bombings (October 1, 2000 to June 30, 2003), which fulfilledMinistry of Health suggested criteria for MCIs (number of admis-sions, severity of injury).Results: Included were 325 casualties from 32 events, 34% of whichhad an Injury Severity Score �16. A third of the admissions arrivedwithin 10 minutes and 65% within 30 minutes. Forty percent of thepatients underwent CT scans directly from the ED. Operative pro-cedures were performed on 60% of patients and 36% were trans-ferred directly from the ED to the OR. Initiation of surgical proce-dures peaked at 1 to 1.5 hours, mainly multidisciplinary abdominal,thoracic, and vascular surgery. Orthopedic and plastic surgery pre-dominated later. A third of the patients were admitted to ICUs, often(31%) directly from the ED.Conclusions: High staffing demands for ED, OR, and ICU overlap.Anesthesiologists, general, thoracic, and vascular surgeons are inimmediate demand. ICU admissions occur simultaneously withongoing patient arrival to the ED. Most patients operated within thefirst 2 hours require multidisciplinary surgical teams. Demand for

orthopedic and plastic surgery and anesthesiology services continuesfor �24 hours.

(Ann Surg 2006;243: 533–540)

The need to cope with terror-induced multiple casualtyincidents (MCIs) is becoming a reality for hospitals in a

growing number of countries.1,2 Aside from the challenge oftreating a large volume of casualties in the already crowdeddepartment of emergency medicine (ED)3 within a short timeframe,4 the unforeseen and sporadic nature of these incidentscomplicates in-hospital preparation. Explosions are by far themost common type of terror-induced MCI.5,6 The currentstudy undertook to suggest guidelines for entire hospitalorganization during MCIs resulting from bombings in acivilian scenario. The unified experience of all level I traumacenters in Israel who participated in trauma care duringsimilar-type terror-related MCIs was studied. Hospital re-sponses to prior MCIs were characterized, and this informa-tion was used to map the initial 24-hour response, enablinganticipation of the response to subsequent MCIs of a similarnature and magnitude.

MATERIALS AND METHODS

DefinitionsMCI

Based on the criteria suggested by a committee ap-pointed by the Israeli Ministry of Health Division for Emer-gency services, this study defined an MCI for the purposes ofin-hospital mobilization as an event which causes �10 casu-alties or �4 severely injured casualties (Injury Severity Score�ISS� � 16) to arrive within a brief time frame at a level Itrauma center.

Time ZeroTime zero was defined as the time of arrival of the first

admitted patient. This time was selected to avoid bias causedby the use of times reported by more than one source.

Clinical DatabasesValidated data regarding the number of casualties ar-

riving at each level I trauma center were provided by the

From the *Shaare Zedek Medical Center, Jerusalem, Israel; the †IsraelNational Center for Trauma and Emergency Medicine Research, theGertner Institute for Epidemiology and Health Policy Research, Israel;‡Department of Anesthesiology and Critical Care, Ein Kerem Campus,Hadassah Hebrew University Medical Centers, Jerusalem, Israel; and§Mt. Scopus Campus, Hadassah Hebrew Medical Centers, Jerusalem,Israel.

Presented at the Surgical Forum of the 90th Annual Clinical Congress of theAmerican College of Surgeons, October 2004.

�The Israel Trauma Group includes the following: Ricardo Alficci, MD, IgorJeroukhimov, MD, Yoram Klein, MD, Yoram Kluger, MD, MosheMichaelson, MD, Avraham Rivkind, MD, Gad Shaked, MD, DanielSimon, MD, and Michael Stein, MD.

Reprints: Sharon Einav, MD, Intensive Care Unit, Shaare Zedek Medical Center,POB 3235, Jerusalem 91031, Israel. E-mail: bromi@md2.huji.ac.il.

Copyright © 2006 by Lippincott Williams & WilkinsISSN: 0003-4932/06/24304-0533DOI: 10.1097/01.sla.0000206417.58432.48

Annals of Surgery • Volume 243, Number 4, April 2006 533

Operations Division of the Israeli National Emergency Med-ical Services (MDA) (including arrivals who were trans-ported by the EMS and those who were not). All dataregarding admitted patients for this study (including patientdemographics, mechanisms and type of injury, ICD-9-CMdiagnosis categories, therapeutic time frames, proceduresperformed, admitting departments, length of hospital stay,and patient outcomes) were retrieved from the NationalTrauma registry. Data regarding lengths of stays of casualtiesinvolved in motor vehicle accidents during the study period(presented for comparison) were also derived from the na-tional trauma registry.

Data Collection on Location for the NationalTrauma Registry

The Israel National Center for Trauma and EmergencyMedicine collects data for admitted trauma patients from the6 level I trauma centers in the country as well as 4 regionaltrauma centers. A trauma registrar stationed at each level Itrauma center collects the data during the admission period ofthe hospitalized patients. The quality and completeness ofdata collection in each level I trauma center are supervised bymembers of the trauma team: the coordinator (a nurse) andthe director of the trauma unit (a trauma surgeon). All dataare later transferred by the trauma centers to the registry.Identical methods apply to all hospitals.

Inclusion/Exclusion CriteriaAll terror-related incidents occurring over a period of 33

months (October 1, 2000 to June 30, 2003) were screened.Incidents were included if they fulfilled criteria, as previouslydefined, for MCIs and if they were caused by a suicide bombing.Data relating to civilians who were admitted to level I traumacenters as a result of these incidents were analyzed.

Data Presentation and Statistical AnalysesData were collected using designated trauma registry

software. Statistical analyses were performed using both SAS8.0 (SAS, Cary, NC) and SPSS 10.0 (SPSS Inc., Chicago, IL).

Descriptive statistics used included frequency distribu-tions, ratios, means, medians, ranges, and standard deviations.Demographic data (eg, sex, age groups, ISS) are presented as theprevalence of each characteristic within the study population.The number of patients having need of each in-hospital resourceis presented using medians and 75th and 95th percentiles toreflect the required in-hospital surge capacity.

RESULTS

Study Population (n � 325)During the study period, there were 1639 terror-related

civilian casualties. Included were 325 casualties from 25MCIs, which met criteria for the in-hospital definition of anMCI. Several of these MCIs produced sufficiently largenumbers of injured civilians to necessitate mobilization ofmore than one hospital. Therefore, 32 in-hospital mobiliza-tions are described (hitherto “events”).

The male-to-female ratio was 1.35:1 (Table 1). Themean patients’ age was 32.3 � 16 years, and none of the

admitted patients was younger than 2 years. Further relevantdemographic data are presented in Table 1.

Resources in the Department of EmergencyMedicine (n � 32 Events)Arrivals and Admissions

The average number of patients arriving at the ED was20.9 � 11.6 (median, 18 patients; 75th percentile, 23 patients;95th percentile, 54 patients). The average number of admittedpatients per event was 10.2 � 8.4 (median, 7 patients; 75thpercentile, 13 patients; 95th percentile, 28 patients). Overall,approximately 45% of the arrivals were admitted (in 2 events,there were conflicting data regarding the exact number of

TABLE 1. Patient and Injury Characteristics of the StudyPopulation (n � 325)

Demographic Detail

Prevalence AmongAdmitted Patients*

n %

Sex

Male 187 57.5

Female 138 42.5

Age

0–14 14 4.3

15–29 161 49.7

30–44 81 25.0

45–59 42 13.0

60–74 22 6.8

75� 4 1.2

Religion

Jewish 275 84.6

ISS

1–8 151 47.5

9–14 56 17.6

16–24 44 13.8

25–75 67 21.1

Admission Characteristics n %

Sent to CT scan from the ED 129 39.7

Underwent operation 196 60.3

Direct transfer to the OR 116 35.7

1st admitting department†

ICU 109 33.5

Orthopedics 61 18.8

General surgery 59 18.1

Other surgical 36 11.1

Nonsurgical 32 9.8

Pediatric surgery 6 1.9

Neurosurgery 5 1.5

Chest surgery 3 0.9

Mortality

Died prior to transfer to department 14 4.3

Overall inpatient mortality 26 8

*Missing data: for age, 1 patient; for religion, 2 patients; for ISS, 7 patients. Allpercentages are corrected.

†Vascular and plastic surgery wards are reported as “other surgical” wards.

Einav et al Annals of Surgery • Volume 243, Number 4, April 2006

© 2006 Lippincott Williams & Wilkins534

arrivals). However, there were marked differences betweenevents in the proportion of admissions to arrivals, rangingfrom 10% to 90%.

Arrival TimesThirty-four percent of the patients arrived within 10

minutes and 65% of the patients arrived within 30 minutes oftime zero. The average arrival time of admitted patients was24 � 30 minutes (n � 267).

Severe Injuries (ISS > 16)There were altogether 111 admitted patients with an

ISS � 16 (Table 1), constituting 34% of the admissions. Theaverage number of admissions/event with an ISS � 16 was3.5 � 3.4 (median, 3 patients; 75th percentile, 5 patients; 95thpercentile, 10 patients). The first admitted patient with an

ISS � 16 arrived within 7.36 � 11.1 minutes of time zero.However, in most events, the first admitted patient with anISS � 16 arrived within 5 minutes of time zero (median, 4minutes; range, 0–43 minutes).

Operating Room ResourcesOperative procedures were performed on 60.3% of

admitted patients (n � 196), 116 of which (35.7%) weretransferred by the triage surgeon directly to the OR (Table 1).There was a bimodal distribution of initiation of surgicalprocedures over time (Fig. 1A): a first peak at 1 to 1.5 hoursfrom time zero comprised predominantly (70%) of patientswith ISS � 16 and a second, smaller peak at 2.5 to 4 hoursfrom time zero. The number of severely injured patientsentering surgery increased 2-fold after 30 minutes and 3-foldafter 60 minutes (Fig. 1A). This was not due to late arrival

FIGURE 1. Operating room utiliza-tion for event-related proceduresfollowing an MCI. The percent ofpatients entering surgery is shownrelative to time zero (A) and rela-tive to their own time of arrival (B).The contribution of each ISS cate-gory to the total number of surgi-cal procedures being initiatedwithin each time frame is alsoshown. For patients who under-went more than one type of sur-gery, only the starting time of thefirst procedure is given (n � 149).

Annals of Surgery • Volume 243, Number 4, April 2006 Hospital Organization in Multiple Casualty Events

© 2006 Lippincott Williams & Wilkins 535

times of the more severely injured patients, as shown inFigure 1B, which relates patients entering surgery to theirarrival time. The number of surgical procedures being initi-ated plateaus within 6 to 7 hours (Fig. 1A).

The majority of patients entering surgery within thefirst 2 hours were severely injured (Fig. 2). During the firsthours, initiation of abdominal, thoracic, and vascular surgical

procedures predominated, and most patients entering surgeryrequired 2 or more surgical teams working in parallel (Fig. 3).

After several hours, the proportion of procedures beingstarted for patients with lesser injuries gradually increased (Fig.2). Orthopedic and plastic surgery predominated at this time andincreased demand for these surgeons and anesthesiology ser-vices continued for even more than 24 hours (Fig. 3).

FIGURE 2. Percent of patients en-tering surgery during each timeinterval by ISS. Note the predomi-nance of high ISSs (�16) in thefirst hours. The arrow denotes thedecreasing portion of high ISSs.

FIGURE 3. The percent of surgicalprocedures commencing over time(n � 196). Subdivisions denote thesurgical specialties performing theprocedures. A large number of pa-tients underwent multidisciplinarysurgery, yielding a proportion ofprocedures per patients greaterthan 100% (eg, 2.75 procedures/patient in the initial 0- to 30-minute time frame). The durationof surgery is not shown.

Einav et al Annals of Surgery • Volume 243, Number 4, April 2006

© 2006 Lippincott Williams & Wilkins536

ICU and Patient Ward ResourcesICUs were the first admitting department for a third

(33.5%, n � 109) of the admitted patients (Table 1). Theaverage number of patients admitted to the ICU was 3.4 � 3per event (median, 3.5 patients; 75th percentile, 5.7 patients;95th percentile, 8.5 patients). Almost a third of ICU admis-sions (31%, n � 34) were transferred there directly fromthe ED.

The average number of patients admitted to the ICUfollowing a surgical procedure in the OR was 2.3 � 2.5 per

event (median, 2 patients; 75th percentile, 4 patients; 95thpercentile, 6.0 patients). Patients who were admitted to theICU following a surgical procedure had higher ISSs thanthose who were transferred to the ICU directly from the ED(Table 2).

General and orthopedic surgery wards each admittedapproximately one fifth of the patients. Nonsurgical wardswere used as first admission wards for 10% of the overallnumber of patients since surgical ward capacities were easilyexceeded (Table 1). These nonsurgical wards included thedepartments of internal medicine (69% of the overall over-flow in 22 events) and various recovery rooms (22% of theoverall overflow in 7 events).

RadiologySlightly more than one third of the admitted patients

were sent to CT scan directly from the ED (Table 1), yieldingan average of 4 � 3.2 ED CTs per event.

Length of Stay (LOS) and MortalityLOS

The median length of hospital stay (LOS) for the entirestudy population was 7 days (75th and 95th percentiles, 15and 44 days, respectively) (Table 3) and for survivors it was7 days. When compared with victims of motor vehicle acci-

TABLE 2. Injury Severity Scores of Victims of MCIsAdmitted to Intensive Care Units Directly From theEmergency Department or Following a Surgical Procedure*

ISS

Transferred DirectlyFrom the ED

Admitted AfterSurgery

n % n %

1–8 4 12 5 7

9–14 8 24 11 15

16–24 10 29 17 23

25–75 12 35 41 55

Total 34 100 74 100

*Data missing for 1 patient regarding the mode of admission. All percentages arecorrected.

TABLE 3. Patient Length of Stay

Length of Stay

% of AdmittedPatients*(n � 325)

% of Survivors(n � 299)

A. Length of hospital stay following injury in anMCI (days)

1 23.2 19.5

2–7 31.6 32.3

8–14 20.1 21.2

15–28 15.5 16.5

29� 9.6 10.5

nMedian(days)

IQR 25–75(days)

Range 5–95(days)

B. Length of hospital stay of various civiliantrauma patient categories

In hospital, all patients

MCIs 323 7 2–15 1–44

MVAs 13945 3 1–7 1–23

Other trauma 40832 3 1–7 1–19

In ICU

MCIs 109 4 2–9 1–38

MVAs 1878 3 1–9 1–31

Other trauma 2053 3 1–8 1–28

In hospital, patients who were in ICU

MCIs 109 17 10–27 4–66

MVAs 1864 11 6–22 2–47

Other trauma 2039 10 5–22 2–56

*All percentages corrected for missing data (2 patients) and mortality where relevant.MCIs indicates multiple casualty incidents; MVAs, motor vehicle accidents.

Annals of Surgery • Volume 243, Number 4, April 2006 Hospital Organization in Multiple Casualty Events

© 2006 Lippincott Williams & Wilkins 537

dents and other civilian trauma, the average overall hospitalLOS of victims of MCI was 5 days longer, and the ICU stayof these patients was 1 to 2 days longer (Table 3). However,there was considerable variability in the LOS.

MortalityOverall inpatient mortality rate was 8% (n � 26)

(Table 1). Females had a higher mortality rate than males:11.6% for females (16 of 138) versus 5.3% for males (10 of187). The expired female population had higher ISSs than theexpired male population, but statistical significance was notsought because of the small sample size.

The majority of deaths (17 of 26) occurred within thefirst day of admission (Fig. 4). Deaths occurred in the fol-lowing locations: 5 patients died in EDs, 7 in ORs, 12 inICUs, and 2 patients in recovery rooms acting as provisoryICU locations.

DISCUSSIONIn-hospital management of MCIs is largely unstudied.

The current study is the first to draw upon the pooledexperience of several level I trauma centers who have par-ticipated in trauma care during terror-related MCIs resultingfrom suicide bombing to propose guidelines for initial 24-hour in-hospital mobilization, organization, and resource uti-lization.

In Israel, hospitals near an MCI are immediately noti-fied by the EMS of the occurrence, location, and mechanismof an event. EMS evacuation times from most MCI locationsare generally known and usually brief.7 These 2 referencepoints provide the initial time frame essential for in-hospitalmobilization and deployment. Each hospital has a contin-gency deployment plan that has been practiced and perfectedby the staff and may be implemented upon notification of anMCI within the vicinity. All medical, nursing, and ancillary

personnel have assigned functions in specific hospital loca-tions. Providing an adequate number of hospital beds at amoments notice to allow unidirectional flow through the EDand reduce crowding is a major challenge. It demands exten-sive in-hospital mobilization, dependent on the balance be-tween existing bed occupancy and the surge capacity requiredfor the event at hand.

Routine patients already in the ED are either all trans-ferred or triaged for transfer, based on the estimated numberof required beds as gleaned from the reports coming in fromthe field. Decisions regarding transfer are usually made by asenior ED physician and a senior surgeon. Depending on themagnitude of the MCI, a contingency ED, located at apredesignated alternative location in the hospital may beopened to absorb already existing or newly arriving routineED patients. This ED may also function as a temporaryadmission location until ward beds are available and all MCIcasualties have been provided for. Routine patients who areeligible and prepared for admission are transferred immedi-ately to any available hospital bed. Nonsurgical staff mem-bers (eg, physicians from the departments of internal medi-cine, neurology, pediatrics) participate in the transfer and careof these patients, providing both continuity of care from theED to the ward and improved transfer of information. Simul-taneously, ward beds (preferably surgical) are made availablefor MCI admissions.

ICU bed availability is one of the foremost concerns inthe hospital deployment plan. ICU beds are always in highdemand8 and during an MCI relatively more ICU beds areneeded than any other department beds. High individualpatient nursing and monitoring requirements9 and demandingpatient mobilization tax an already strained system. Patientsin general/surgical ICUs are transferred to alternative ICUs(cardiac surgery, cardiac care, internal medicine, and pediat-ric) or even to wards to vacate ICU beds. The issue of ICUtriage is both emotionally loaded and controversial. Thepostanesthesia care unit, where beds, monitoring equipment,and trained nursing staff are readily available, should be usedas additional ICUs. As noted elsewhere, there is an overlap inthe staffing demands of the ED, OR, and ICU during theinitial phase of an MCI.10 ICU admissions occur in parallel toboth emergency surgery and ongoing patient arrival at theED, making the situation even more challenging. ICU bedsmust be vacated and new patients admitted at the same timethat ICU physicians are also required in the ED and OR. Inthe current study, even though the ISSs of the patients whowere admitted to ICUs directly from the ED were lower thanthose of patients admitted after surgery, apparently they werestill critically injured due to multisystem blast and/or shrapnelinjuries.

Hirshberg et al11 suggested that the increasing need fordiagnostic computed tomography (CT) will make imaging thebottleneck of patient management during MCIs. In this study,40% of the admitted patients underwent a CT in the ED.Suicide bombs contain a considerable amount of shrapnel inthe form of nails, screws and bolts; therefore, multiple andoften insidious penetrating injuries and unpredictable trajec-tories are fairly common.12,13 Tomographic screening is man-

FIGURE 4. Time to death (admission days). The overall mor-tality was 8% (n � 26). Deaths during the first day of admis-sion occurred in EDs (n � 5), ORs (n � 7) recovery rooms(n � 2), and ICUs (n � 3). All of the later deaths occurred inICUs (n � 9).

Einav et al Annals of Surgery • Volume 243, Number 4, April 2006

© 2006 Lippincott Williams & Wilkins538

datory to locate the shrapnel and uncover unsuspected inju-ries. Even though the time needed to perform a CT scan hasshortened significantly due to the use of rapid, multislicespiral CT, performance of multiple CT scans remains time-consuming and staff intensive.

Accumulated experience has led to use of a simplifiedtriage algorithm in the ED14 for arriving MCI casualties. Thisalgorithm is based on the degree of urgency of necessaryinterventions rather than on the detailed severity of injury ofthe patient. Severely injured patients are not necessarily thefirst to arrive15 and may continue arriving for an unknownperiod of time. It is therefore prudent to reserve operativecapacity for as long as possible. Immediate surgery is pro-vided for patients requiring life-saving and/or limb-savingintervention. Only when the entire scope of the event be-comes clear are the surgical teams still in reserve used forcases in need of less urgent surgical intervention. With thisapproach, in an overwhelming event, life-saving surgicalcapacity will ultimately be limited only by the number ofexpert surgical teams on location.

During an MCI, the limit of hospital capacity to per-form optimal triage for all arrivals and provide quality care toselect critically injured patients is easily exceeded. The EDmay be inundated with minor/moderate injuries16 impairingthe ability for a sufficient and timely response to the moreurgent casualties.3 This issue was examined by Hirshberget al.17 who found that saturation may occur with a criticalcasualty load of 4.6 patients/hour using immediately avail-able resources and of 7.1 patients/hour during full deploy-ment of all hospital trauma assets. The delicate balance betweenover-triage and under-triage is easily disrupted, initially leadingto either missed diagnoses with an increased critical mortal-ity18 or a level of care that becomes dependent on availabilityof resources outside the ED.17 This paper outlines the need toconsider simultaneous appropriate resource allocation to theED and other hospital locations during and immediatelyfollowing MCIs.

“Surgical conservativism” (deliberate suspension ofsurgery for patients who were severely injured but not yet inimmediate danger of loosing life or limb) is being practicedduring MCIs in Israeli level I trauma centers. In our data,initiation of surgical procedures peaked at 30 to 90 minuteseven for more severely injured patients. This phenomenondid not derive from prolonged staff arrival times or latearrival times of these patients to the ED. High ISSs predom-inated in the patients who entered surgery within the first 2hours, indicating that for these patients surgery was oftenlifesaving and could not have been postponed despite ongo-ing patient arrival to the ED.

Our data indicate that the majority of critically injuredpatients have abdominal, thoracic, or vascular injuries, cre-ating an immediate demand for the relevant surgical servicesand for anesthesiologists. Upon notification of an event,elective surgery should immediately be suspended until thescale of the event is clarified. Elective patients who arealready in the OR holding area either remain there until moreinformation becomes available regarding the scale of the MCIor are sent back to the ward. The elective patient who has

already undergone anesthesia but not surgery should beconsidered individually. Other urgent surgery cases should beincluded in the triage list with the MCI casualties.

Severely injured MCI patients are often demandingeven when their injuries are not immediately life-threatening.In the OR, a significant part of the experienced staff may beengaged in treating one individual patient, particularly sincemultiple surgical procedures are often performed in parallelduring the first several hours. From the viewpoint of the ED,once thus occupied in the OR this staff will be unavailable totreat still arriving patients. Withholding less urgent surgeryalso allows these trained teams to be put to good use in theED, improving the efficacy of patient triage, management,and admission.

This study validates previous observations and findingsregarding the potentially prolonged demands imposed on theresources of several departments.10,19 General, thoracic, andvascular surgery will be occupied during the initial hourspost-MCI. Orthopedic surgery, plastic surgery, and anesthe-siology services should be prepared to handle an excessiveand ongoing OR load for more than 24 hours. Cessation ofelective surgery imposes a significant loss of hospital reve-nues from routine patient care. This is a major incentive torestore routine OR scheduling as soon as possible, to mini-mize further financial losses. High utilization of radiologyservices are also resource consuming and high ICU admissionrates and prolonged patient LOS impose further burdens.

We acknowledge a few limitations of this study. Togain a more focused view, only MCIs following suicidebombings in level I trauma centers were studied. Cautionshould therefore be exercised in extrapolating the findings ofthis study to different scenarios. The time between the oc-currence of the MCIs and arrival of the first patient is also notgiven. Each hospital must take into consideration its individ-ual geography and approachability at the different times ofday and night when planning for force-mustering. Finally, thetrauma registry includes only casualties who were eventuallyadmitted or died in the ED.

Several conclusions stem from this study: 1) In urbanbomb explosions occurring among a civilian population mostadmitted patients arrive within a brief time interval (�30minutes). However, the number of admissions cannot bepredicted solely upon the basis of the number of immediatearrivals. 2) To prevent ED backlogging, patient mobilization,whether by hospital discharge or interdepartmental transfershould occur timely and across the hospital. All departments,including those not directly involved in trauma care, shouldbe prepared to provide beds and staff. 3) Severely injuredpatients constitute approximately a third of the admissionsand tend to arrive early on in the event, mandating assign-ment of a large number of appropriately trained staff to theED right at the outset. Such patients may, however, continueto arrive �2 hours after arrival of the first patient. 4) Anes-thesiologists, general, thoracic, and vascular surgeons will bein immediate demand after an event in both the ED and theOR for life-saving surgery. Reserve surgical teams comprisedof this staff should function in the ED until either a clearerpicture of the entire event has emerged or it is no longer

Annals of Surgery • Volume 243, Number 4, April 2006 Hospital Organization in Multiple Casualty Events

© 2006 Lippincott Williams & Wilkins 539

medically feasible to withhold necessary surgery. 5) Hospitalcontingency plans must allow for the overlap in the staffingdemands for the ED, OR, and ICU. 6) Orthopedic surgery,plastic surgery, and anesthesiology services should be pre-pared to handle an excessive and continuous OR load formore than 24 hours. Although further studies are required tosubstantiate and broaden our findings, this study offers somebasics toward institution of routine in-hospital preparednessfor MCIs.

ACKNOWLEDGMENTSThe authors thank the Ops Division of the Israeli

National Emergency Medical Services (MDA) who providedus with information regarding the number of casualties ar-riving at each hospital.

REFERENCES1. World Health Organization, regional office for Europe: Coopera-

tion in the face of terrorism. http://www.euro.who.int/futuresfora/topics/20031114_1.

2. Terrorism Research Center; Terrorist attack archives. http://www.terrorism.com/modules.php?op�modload&name�Attacks&file�index.

3. Schneider SM, Gallery ME, Schafermeyer R, et al. Emergency departmentcrowding: a point in time. Ann Emerg Med. 2003;42:167–172.

4. Centers for Disease Control and Prevention. Rapid assessment of inju-ries among survivors of the terrorist attack on the World Trade Center–New York City, September 2001. MMWR Morb Mortal Wkly Rep.2002;51:1–5.

5. Peleg K, Aharonson-Daniel L, Stein M, et al. Gunshot and explosioninjuries: characteristics, outcomes, and implications for care of terror-related injuries in Israel. Ann Surg. 2004;239:311–318.

6. Kluger Y. Bomb explosions in acts of terrorism: detonation, wound

ballistics, triage and medical concerns. Isr Med Assoc J. 2003;5:235–240.

7. Einav S, Feigenberg Z, Weissman C, et al. Evacuation priorities in masscasualty terror-related events: implications for contingency planning.Ann Surg. 2004;239:304–310.

8. Vincent JL. Forgoing life support in western European intensive careunits: the results of an ethical questionnaire. Crit Care Med. 1999;27:1626–1633.

9. Henning RJ, McClish D, Daly B, et al. Clinical characteristics andresource utilization of ICU patients: implication for the organization ofintensive care. Crit Care Med. 1987;15:264–269.

10. Shamir MY, Weiss YG, Willner D, et al. Multiple casualty terror events:the anesthesiologist’s perspective. Anesth Analg. 2004;98:1746–1752.

11. Hirshberg A, Stein M, Walden R. Surgical resource utilization in urbanterrorist bombing: a computer simulation. J Trauma. 1999;47:545–550.

12. Kluger Y, Peleg K, Daniel-Aharonson L, et al. The special injury patternin terrorist bombings. J Am Coll Surg. 2004;199:875–879.

13. Almogy G, Belzberg H, Mintz Y, et al. Suicide bombing attacks: updateand modifications to the protocol. Ann Surg. 2004;239:295–303.

14. Hirshberg A, Holcomb JB, Mattox KL. Hospital trauma care in multiple-casualty incidents: a critical view. Ann Emerg Med. 2001;37:647–652.

15. Hogan DE, Waeckerle JF, Dire DJ, et al. Emergency department impactof the Oklahoma City terrorist bombing. Ann Emerg Med. 1999;34:160–167.

16. Biancolini CA, Del Bosco CG, Jorge MA. Argentine Jewish communityinstitution bomb explosion. J Trauma. 1999;47:728–732.

17. Hirshberg A, Scott BG, Granchi T, et al. How Does Casualty Load AffectTrauma Care in Urban Bombing Incidents? A Quantitative Analysis.Abstract presented at the 2004 joint annual meeting of AAST andJAAM, Maui, Hawaii, September 2004.

18. Frykberg ER, Tepas JJ 3rd. Terrorist bombings: lessons learned fromBelfast to Beirut. Ann Surg. 1988 Nov;208:569–576.

19. Cohen M, Kluger Y, Klausner J, et al. Recommended guidelines foroptimal design of a plastic surgery service during mass casualty events.J Trauma. 1998;45:960–968.

Einav et al Annals of Surgery • Volume 243, Number 4, April 2006

© 2006 Lippincott Williams & Wilkins540