New York City (NYC) is the most densely populated city in the United States, with a population of over 8.5 million that swells to over 13 million when including commuters and tourists. Known as the world’s crossroads, it is filled with historical institutions, diverse architecture, and iconic landmarks. Given its population, infrastructure and fame, the potential for disasters, natural and manmade, is ever present. NYC has 62 acute care hospitals, including 16 trauma centers and 4 burn centers. While all hospitals can provide initial stabilization, a large-scale disaster could produce hundreds of injured patients, exceeding the capacity of even NYC’s trauma and burn centers. For example, between the 4 NYC burn centers, the total burn bed count is approximately 44 beds, with nearly half of those beds already occupied on any day. Surge capacity for burn treatment within those 4 centers is difficult to estimate, as it will vary based on existing census, acuity of patients and resources available at the time of the incident. However, based on planning assumptions and previous exercises, the maximal surge capacity for burn injuries is estimated to be approximately 1.5 to 2 times the normal total bed capacity. On 9/11, after the attack on the World Trade Center, over 80 burn patients were hospitalized.
Prehospital emergency medical service (EMS) mass casualty incident (MCI) protocols emphasize the need to triage trauma and burn patients to the nearest trauma or burn center whenever possible; in NYC the Fire Department, City of New York (FDNY) is responsible for coordination of prehospital resources. 1 However, in a large-scale disaster, EMS may transport patients to the nearest hospital regardless of burn/trauma designation due to the high volume of patients they receive and the need to provide immediate medical care. Furthermore, past MCI incidents have demonstrated that numerous patients self-transport themselves to the nearest hospital without knowledge of whether it is a burn or trauma center.
Due to these factors (limited number of burn/trauma beds, patient self-transport and the fact that the nearest hospital is often not a burn or trauma center) a secondary triage process for transporting patients for definitive care to a burn and/or trauma center is critically important. Once the primary EMS response is completed and patients stabilized at their initial hospital destinations, secondary transport decisions should follow a triage process attempting to match patient needs and survivability to available hospital resources within, and if necessary, outside of NYC. For this to occur across tens of hospitals involving a hundred or more patients would require a group of subject matter experts who are familiar with available resources and have access to essential patient information. The NYC Burn MCI Protocol was designed to meet this need. Reference Leahy, Yurt and Lazar2,Reference Yurt, Lazar and Leahy3 Completed in 2011 and developed jointly with NYC Emergency Management (NYC EM), the FDNY, NYC Hospitals, the NYC Department of Health and Mental Hygiene (NYCDOHMH), and the New York State Department of Health (NYSDOH), the protocol addresses coordination of secondary transport of patients with severe burn injuries. The NYC Burn MCI Protocol designs a centrally controlled response plan activated by the FDNY in the event of a large-scale MCI. It categorizes NYC hospitals according to Tiers: Tier 1 hospitals are NYS designated burn/trauma centers; Tier 2 are trauma centers; Tier 3 hospitals are large or specialty hospitals that are neither burn or trauma centers; and Tier 4 or “not tiered” hospitals are smaller community hospitals. All hospitals, regardless of their Tier designation, agreed to participate, thereby eliminating any EMTALA concerns.
Between December 2015 and March 2016, the NYCDOHMH and FDNY convened key clinical leadership from the 4 NYC burn centers to review the NYC Burn Mass Casualty Incident Protocol. As originally written, the NYC Burn MCI Protocol relied upon a Virtual Burn Consultation Center comprised of burn specialists outside NYC. The main goal was to provide 2 services: (1) prioritization of patients for secondary transfer to more definitive care and (2) technical support for nonburn centers caring for severely burned patients until secondary transport is available. However, NYC burn centers agreed to collaborate with FDNY to support the prioritization of patients for secondary transfer within NYC. This agreement was due to delays in establishing this external center and a growing realization that local centers are most knowledgeable of resources available within NYC.
The newly created Mass Casualty Consultation Center (MCCC) would only be activated for large-scale MCIs when available resources for managing acutely severely injured patients have been exceeded and the FDNY determined a need for centralized coordination of secondary transport decisions. The scope was expanded to include not only burn patients, but also trauma patients because (1) large-scale MCIs are likely to generate patients with combined burn and trauma injuries; (2) all burn centers are also trauma centers; and (3) while most trauma centers are not burn centers, all have the capability to treat burns for several days or longer. For this reason, the proposed MCCC team composition included a staff burn and a staff trauma surgeon assigned from each of the NYC burn centers and an FDNY EMS physician. According to the proposed plan, the MCCC is responsible for making inter-facility transfer decisions based on the expertise of the participating physicians. To test the NYC Burn MCI Plan’s MCCC, the FDNY, NYC DOHMH, and the Brooklyn Healthcare Coalition set up a multi-hospital functional exercise known as the “E-bombable Brooklyn”.
METHODS
Multi-hospital Large-Scale MCI Disaster Exercise to Test the NYC Burn Plan’s MCCC
The exercise “E-bombable Brooklyn” was conducted by the Brooklyn Healthcare Coalition for Emergency Preparedness on March 3, 2016, at FDNY Headquarters and Brooklyn hospitals. The exercise scenario was designed to assess the MCCC ability to prioritize secondary triage of patients roughly 24 h after several explosions detonated in Brooklyn, the most populous borough of NYC, potentially resulting in a multitude of injured patients in a short period of time. Eleven of 12 acute care hospitals in the Brooklyn Coalition participated, as well as the 3 largest burn centers in NYC, FDNY, NYC Emergency Management (NYCEM), NYC Regional EMS Council (REMSCO), Greater NY Hospital Association (GNYHA), NYCDOHMH, and NYS DOH. The exercise involved 69 simulated patients, many of them pediatric, who initially presented to local community hospitals, for whom the receiving hospital after evaluation and immediate stabilization then requested transfer to burn or trauma centers for definitive care. Patient profiles were developed by Board Certified emergency medicine physicians and designed to challenge hospital resources for managing medical surge, including patients requiring specialty care. Pediatric patients were of particular interest to the Brooklyn Disaster Coalition because there are no burn centers located in Brooklyn. Adult and pediatric case scenarios included trauma, burn, or mixed trauma and burn injuries, as well as critically ill patients presenting during the emergency.
A NYC-MCCC transfer request form was designed as a comprehensive worksheet containing detailed records of the patient’s medical history, hospital course and current medical state. These forms were seven pages in length and designed to give the triage physician team maximal information regarding the potential transfer of the patient’s medical status as well as information that might be needed to co-manage the patient if secondary transport was not possible for several days. During the exercise, each of these consultation worksheets were completed by clinicians in the local transferring hospitals and faxed to the centralized secondary triage/transfer center at the FDNY. Once all 69 patient transfer forms were received, they were to be distributed to the MCCC for triage prioritization. Due to concerns about the ability of hospitals to fill out these transfer forms within the artificial time constraint (ie, ~60 min) of the drill, a short summary for each of the 69 patients was provided to FDNY by the Brooklyn Coalition before the exercise. This summary included information needed to triage but not co-manage the patient. Once the decision was made to transfer the patient, the receiving burn/trauma center was notified to confirm an open bed existed; transport arranged; and the patient destination tracked.
Assessment of Effectiveness and Reliability of the MCCC During a Multi-hospital Large-Scale MCI Exercise
To test the effectiveness and reliability of the MCCC, 2 different physician MCCC teams (Teams A and B) were asked to prioritize identical patient profiles. Each MCCC team had 3 physicians with the same expertise, a trauma surgeon and a burn surgeon from 1 of the 3 largest NYC Burn/Trauma Centers and an FDNY EMS physician who served as the MCCC organizer/facilitator and provided expertise on transport capabilities. The 2 MCCC teams triaged and prioritized the transfer of the identical 69 simulated patients from local community hospitals to definitive care centers.
Each team was briefed at the start of the drill. The MCCC teams used simulated patient information, knowledge of available services at the patient’s current hospital location and transportation safety issues to prioritize patients for transfer to a burn center, trauma center, or to deny the transfer request and have the patient remain at the current hospital. Transfer denials were based on risk versus benefit to the patient and all transfer requests were made under the assumption that beds could be made available in a receiving Burn or Trauma center within NYC or the metropolitan area (upstate New York or New Jersey) without the need for air transport. The teams had no communication with each other; each team deliberated on its own private conference call line; and a transfer decision made by 1 team did not influence transfer options for the other team.
Postexercise Data Collection and Interview
A researcher was present during the MCCC team meeting, gathered paper documentation of transfer decisions and entered data into a spreadsheet for analysis. Patients were categorized according to characteristics such as age, gender, injury type, and mechanism. Burn patients were classified according to burn size and the presence of inhalation injury, while trauma patients were categorized according to affected organ system. In addition, the extent of injury for both burn and trauma patients was calculated using widely accepted outcomes predictors in their respective fields.
After the exercise, team decisions were compared and discordant results were analyzed further using a quantitative and qualitative approach. The quantitative approach used a burn or injury severity score to explain the discordance. These scores were not used for triage but only for posttriage analyses. Each simulated patient’s injury severity was quantified. For burn patients, the Revised Baux Score was used. This scoring system, which ranges from 0 to 220, is intended as a mortality predictor and is derived from the original Baux Score first introduced in 1961. The Revised Baux Score takes into account the amount of burn injury (percent total body surface area [TBSA] burnt), patient’s age, and the presence of inhalation injury. Reference Osler, Glance and Hosmer4 Recent studies suggest that a Revised Baux Score of 100 carries a mortality rate of 50% and a score of 165 predicts almost certain mortality. Reference Dokter, Meijs and Oen5 For trauma patients, the Injury Severity Score (ISS) was used. This score ranges from 1 to 75 and is calculated by taking the sum of the squares of worst injuries from the 3 most severely injured body areas according to the Abbreviated Injury Scale (AIS). The ISS correlates linearly with mortality, morbidity, hospital stay, and other severity measures. Reference Baker, O’Neill and Haddon6
For the qualitative approach, the researcher interviewed each team leader separately to determine how decisions were made and then reviewed each case where there was discordance in triage assignment. For each discordant triage assignment, the participants were asked to comment on why they believed their final transfer decision differed from the other team. The researcher reviewed the responses and classified them into four categories: (1) Insider knowledge of hospital capabilities; (2) Differing interpretation of trauma versus burn surgeon role in management; (3) Differing interpretation of burn severity; (4) Unexplained (there was not a clear reason for discordance even after reviewing the case with both team leaders).
Statistical Analyses
Demographic characteristics of patient profiles were treated as categorical variables and reported as frequencies and percentages. The McNemar’s test of groups was used to assess significant differences in secondary transport decisions made between the 2 triage teams. Results were considered significant at the P < 0.05 level. Overall agreement of secondary transport decisions by group was first evaluated through percent concordance, but an additional inter-rater agreement analysis using the Cohen’s kappa coefficient was conducted to account for any agreement that may have occurred by chance. Reference Viera and Garrett7 The strength of the inter-rater agreement was interpreted using the Landis and Koch scale. Reference Landis and Koch8 All data were analyzed using SAS version 9.4 (SAS Institute Inc., Cary, NC).
RESULTS
Table 1 summarizes the demographic and injury information for the 69 simulated patients developed by the Brooklyn Coalition for the triage scenario. Of the 69 patients, 52 (75.4%) were male and 17 (24.6%) were female. Patient ages ranged from 1 to 86 y; pediatric patients (age < 15 y) were almost 45% of the study population. Thirty patients (43.5%) presented with both burn and trauma mechanisms; 12 patients (17.4%) with only burn mechanisms; 24 patients (34.8%) with only trauma mechanisms; and 3 patients (4.4%) presented with neither burn nor trauma mechanisms. Of the 42 patients with burn injures, 16 patients had burns greater than 20% TBSA, and 8 had burns greater than 40% TBSA. Patients with inhalation injuries were categorized as burn patients, and if there were no surface burns present, they were assigned a TBSA of 0. Fifty-four patients suffered traumatic injuries with a median ISS of 9. Of the 34 patients with single organ traumatic injuries, 11 (20.4%) were neurologic, 15 (27.8%) were thoracoabdominal, and 8 (14.8%) were orthopedic injuries. Five patients (9.4%) had multisystem traumatic injuries, and 15 (27.8%) had “other” organ system injuries outside of the 3 predefined categories. The patients in this “other” category presented with traumatic mechanisms and concerning initial symptoms but were found to have minor traumatic injuries such as nonburn cutaneous wounds, eye injuries, or nontraumatic medical problems such as gastrointestinal complaints or respiratory difficulties without smoke inhalation. Thirteen of the 69 patients (18.8%) were mechanically ventilated.
TABLE 1 Demographic Information of Patient Profile
a Out of the 5 inhalation injuries, 2 patients are included in the count for the TBSA < 20 group, and the remaining 3 did not have any external burns and are listed under the TBSA = 0 group.
During the exercise, the MCCC teams were not able to receive the lengthy 7-page transfer request forms in the short time allotted (~60 min) due to 1) the extensive information requested, 2) lack of a common electronic communication system for patient transfer forms and 3) overloaded fax reception. To compensate, the exercise switched to using patient summaries organized in a table. Each MCCC team received all 69 patient summaries at the same time and was asked to make triage decision based on that information.
Secondary transport decisions made by the 2 teams are outlined in Table 2. After cases were transmitted to each MCCC, review and prioritization of the patient cohort for secondary triage took approximately 90 min. Of the 69 patients, Team A chose to transfer a total of 49 patients (71%); 26 (37.7%) patients were transferred to a burn center, and 23 (33.3%) to a level 1 trauma center. The remaining 20 patients (29%) were not transferred. Team B chose to transfer a total of 36 patients (52%); 16 patients (23.2%) to a burn center and 20 patients (29%) to a level 1 trauma center. Thirty-three patients (47.8%) were not transferred. There were a total of 21 discordant pairs, in which Team A and Team B recommended different dispositions for the same patient. Overall, there was a statistically significant difference in the total number of transfers between the 2 teams (49 vs 36; P = 0.003); team A transferred more patients overall and more patients with burn injuries to burn centers (26 vs 16; P = 0.016). There was no statistical difference between the 2 groups regarding the secondary transport to Level 1 trauma centers.
TABLE 2 Secondary Transport Decisions Made by Triage Teams (A vs. B) for the 69 Simulated Cases
a Decision groups are mutually exclusive.
b Significant at the 0.05 level using the McNemar test of groups.
Table 3 displays the percent agreement and inter-rater reliability between triage teams. Overall, the 2 teams agreed on the transfer status of 48 of 69 patients or 69.6%, with an inter-rater reliability of 0.55 (confidence interval [CI]: 0.40–0.71) and a moderate strength of agreement according to the Landis and Koch Guidelines.Reference Landis and Koch8 The 2 teams had substantial agreement when triaging to burn centers, with 85.5% overall agreement and an inter-rater reliability of 0.67 (CI: 0.49–0.85). Moderate agreement was seen when triaging to trauma centers (agreement 81%, inter-rater reliability 0.56 (CI: 0.35–0.77) or to not transfer (agreement 72%, inter-rater reliability 0.44 (CI: 0.24–0.64)).
TABLE 3 Percent Agreement and Inter-rater Reliability Between Triage Teams A and B for Secondary Transport
a Cohen’s kappa.
b Strength of agreement based on Landis and Koch guidelines: 0.00–0.20 slight, 0.21–0.40 fair, 0.41–0.60 moderate, 0.61–0.80 substantial, 0.81–1.00 almost perfect.
Table 4 summarizes the level of agreement by demographic information. The 2 teams tended to agree less for female versus male (64.7% vs 71.2%) and for pediatric versus adult patients (64.5% vs 73.7%). The teams also had less agreement for patients with burn mechanisms (58.3%) when compared with patients with trauma mechanisms (70.8%) and patients with mixed trauma and burn mechanisms (70.0%). The teams tended to have increased agreement in patients with a higher acuity of injuries, exhibiting 100% agreement in burn patient with TBSA > 20%, and in patients with Revised Baux Scores higher than the 45.5 median (85.7% vs 47.6%). Teams also tended to have higher agreement in patients with Injury Severity Score above the median (90.9% vs 38.1%). These results indicate that disagreements occurred in less severe cases.
TABLE 4 Level of Agreement According to Patient Demographics and Injury Factors
a Of the 5 inhalation injuries requiring mechanical ventilation, the 2 teams were in agreement (transfer to a burn unit) only for the 2 patients who also had surface burns.
The teams, however, had poorer agreement for mechanically ventilated patients than those breathing spontaneously (53.9% vs 73.2%). Qualitative analysis indicated that this was risk of transport versus the unclear benefit of transfer to a burn or trauma center when the sole injury was smoke inhalation or when there were both respiratory and central nervous system injuries.
Based on the qualitative analysis of the 21 discordant triage decisions, the most common reason for discordance was insider knowledge of local community hospital capabilities (n = 6). Other reasons for discordance included different interpretations for the roles of trauma versus burn surgeons (n = 2), different interpretations of what constitutes a priority burn (n = 3), combinations of the above (n = 4), and in 6 cases, the discordance could not be explained.
DISCUSSION
The information obtained from this large-scale disaster exercise gives important lessons regarding disaster response systems. First and foremost, this study demonstrates that local physician experts in burn, trauma and EMS were able to rapidly prioritize simulated patients in a mass casualty scenario for secondary transport from receiving hospitals to specialty hospitals within NYC. These physician teams performed the triage with relative consistency, with an overall concordance rate of 69.6% and a moderate strength inter-rater reliability using only their clinical training and experience. And, the 2 teams tended to agree less for pediatric as compared to adult patients (64.5% vs 73.7%). This raises the question of whether the variations seen in this study could be minimized with (1) protocols and training specific to secondary triage and transport decisions, and (2) the addition to the MCCC team of a pediatrician specializing in burn, trauma, intensive care, or emergency medicine.
We found that there was higher agreement among physicians regarding transfers to burn centers than transfer to trauma centers. This may be in part due to the direct correlation between mortality and burn size (TBSA) and the widely accepted burn center referral criteria from the American Burn Association. For burn and trauma patients, we found that there was increased concordance among higher acuity patients. For example, in burn patients there was 100% agreement among patients with higher than 20% TBSA burns and 90% agreement in trauma patients with higher than the median ISS for this exercise (Table 4). This suggests that physicians tend to recognize and agree on transfer for the sickest patients, but have differing opinions on lower acuity patients.
An exception was in mechanically ventilated patients, where the teams had less agreement in the higher acuity ventilated patients. The discordance was in 2 areas, in patients requiring mechanical ventilation due only to smoke inhalation (without surface burns or trauma) and in patients also suffering from head trauma. In cases with only smoke inhalation, this discordance is likely due to the lack of evidence-based guidelines for where smoke inhalation acute lung injury is best managed (burn center, medical intensive care unit [ICU] or surgical ICU).Reference Kadri, Miller and Hohmann9 In cases with head trauma the discordance may have been due to the differing perceptions of risk versus benefit.
Physicians may have varying opinions as to what hospitals not designated as burn or trauma centers can provide based on personal experience and practice pattern in this subset of patients (Table 5). We learned through our qualitative analysis that insider knowledge of the community hospital and its capabilities was the most common identifiable reason for discrepancies among physicians’ triage decisions, accounting for nearly 30% of the 21 discordant cases in this study. In 1 case, a head trauma patient was not transferred to a trauma center because of insider knowledge that the neurosurgical team at that local hospital was outstanding. There are also different perceptions as to the risk of transport. Although there is no evidence that inside knowledge of the community hospitals results in improved outcomes, we must acknowledge that these factors affect physician decision-making. Local physician team familiarity with the community infrastructure may result in triage discordance that is otherwise unexplained by the quantitative assessment of patient injury.
TABLE 5 Reasons of Discrepancies Between Teams A and B
The initial plan for this exercise was to have the referral hospitals fill out a detailed transfer request form for each patient. The reasons for including extensive details were to provide information not for secondary triage, but to co-manage the case if secondary transport was not feasible. The exercise proved that such detailed forms were too burdensome to complete and transmit in the short time allotted. This could be improved upon if more time was available or if data share/transfer through electronic medical records was an option. However, given the current resources available, it is our opinion that the best approach is to have 2 separate processes, a short form for secondary triage and an entirely different system (eg, long forms, full access to the electronic medical record and/or telemedicine) if distant prolonged co-management is required. Although the short summaries were far less detailed than what we had originally planned, the participating MCCC physicians unanimously agreed that the short summaries provided enough information to guide their secondary triage transfer decisions. Knowing this has informed ongoing development of a 1-page transfer request form.
A benefit of using a local MCCC for secondary triage is the knowledge of local conditions and expertise at each of the referring hospitals. A disadvantage is that the local MCCC members may not have time to devote to the MCCC if they are overwhelmed by patient care at their own center. We believe the latter would not be a concern for a major metropolitan area such as NYC because secondary triage would occur many hours and possibly even 1–2 d after the disaster; these centers are well staffed and as we have shown, the time needed to prioritize the cases took no more than 90 min. Depending on the size of the MCI and the types of injuries, the MCCC might convene several times daily, potentially reducing the number of patients and time required for each subsequent session. The presence of an EMS physician on the MCCC was useful in providing additional insider knowledge as to referral hospital capabilities and transport safety issues.
It is important to note that this exercise only tested the MCCC’s capabilities for secondary triage. We did not test mutual aid agreements between NYC and regional burn/trauma centers for patient evacuation. Nor did we test remote patient co-management. Therefore, we cannot comment on whether the physicians in a locally based MCCC would have the time to provide clinical support to local hospitals in addition to caring for their own patients and participating in secondary triage. It is our hope that secondary triage and transport would minimize or eliminate the need for distant co-management.
Furthermore, secondary triage decisions were based on data assessed in a short concise table format; it is possible that triage decisions would have been different if the more detailed seven-page MCCC form was used by the review teams. However, that was not the opinion of the MCCC physicians during our qualitative interviews. Finally, the discordance between female and male patients remains unexplained, but may be an anomaly due to limited numbers of female patients in this study.
Finally, this exercise was not designed to determine if specific protocols and training could reduce the discordance between MCCC teams, or if discordance in MCCC transfer decisions has an impact on patient outcome. This is partly because there is no current gold standard transfer protocol to which to compare our data. Rather, our study aims to gather useful quantitative and qualitative information regarding the MCCC secondary triage transfer process so that a future standard can be established.
CONCLUSIONS
Local experts in burn care, trauma care, and EMS were able to rapidly prioritize simulated patients in a large-scale MCI disaster exercise for secondary triage and transport from receiving hospitals to specialty hospitals within NYC using key information with relative consistency. Multi-disciplinary physician teams were able to complete the prioritization based on limited clinical data, suggesting that a brief clinical request form is sufficient for coordinated secondary transport of patients in a disaster. More extensive information sharing processes are likely to be needed to support prolonged co-management of disaster victims in hospitals that do not specialize in burn or trauma care. Variations in prioritization decisions might be addressed by developing additional training for MCCC teams. Training for MCCC teams should emphasize guidelines for transfer that are based on existing services at sending and receiving hospitals, as knowledge of local community hospital systems influence physician decision-making. And, due to lessons learned from this drill, a pediatrician (specializing in burn, trauma, intensive care, or emergency medicine) has been added to the MCCC and the MCCC has been incorporated into existing city protocols, both of which should strengthen NYC’s capacity to respond to large-scale disasters through improved coordination of patient movement.
Acknowledgments
We acknowledge the contributions of The Brooklyn Healthcare Coalition to the design and conduct of the exercise. We thank Dr. Robert Bristow, Dr. Bruce Greenstein, and Mordy Goldfeder for their contributions to the planning and conduct of the exercise.
Funding
This publication was supported by Cooperative Agreement Number 5U90TP000546-04 from the Centers for Disease Control and Prevention and/or Assistant Secretary for Preparedness and Response.
