Multipatient Disaster Scenario Design Using Mixed Modality Medical Simulation for the Evaluation of Civilian Prehospital Medical Response: A “Dirty Bomb” Case Study

Disasters are defined as events that overwhelm local response capability and present myriad challenges to responding medical personnel.1,2 Prehospital medical providers and hospital personnel can be confronted with a multitude of patients, unusual illnesses or injuries, failure to communicate, poor logistic support, the need for decontamination, and difficulties in perimeter control.3–8 Concurrent demands often lead to task overload and performance failures.9 After-action reports (Rhode Island Disaster Initiative Annotated Bibliography 2002, available at: http://www.ridiproject.org/downloads/annotated_bibliography.pdf) and disaster literature highlight recurring problem areas.10–16 The potential for a variety of natural and manmade catastrophic events makes health care provider preparation for mass casualty incidents (MCIs) both imperative and challenging. Analysis of disaster response often attributes failure to inadequate preparedness and the inability to follow designated plans.17 Disaster training needs of medical responders may not be adequately represented in current instructional paradigms,18 especially with respect to weapons of mass destruction (WMD) situations (Rhode Island Disaster Initiative Vulnerability Assessment 2002, available at http://www.ridiproject.org/downloads/vulnerability_assessment.pdf) In the absence of a published standardized methodology,19–22 investigators applied mixed modality medical simulation (SIM) techniques to generate a multipatient WMD disaster scenario and accompanying evaluation tool for education and research purposes. Study Objectives Study investigators aimed to create a repeatable and immersive medical simulation disaster scenario capable of eliciting and recording a representative selection of critical and objectively measurable aspects of on-scene civilian disaster medical response. Combining full-body manikin simulators and professional actors within a high-fidelity environment, the scenario was scheduled to be repeated with trained observers and comprehensive audiovisual systems for systematic collection of prehospital provider participants' clinical performance data and perceptions of simulation utility. MATERIALS AND METHODS Design of Simulated Disaster Scenario and Evaluation Tools Investigators planned a precisely controlled disaster simulation scenario to elicit and study fundamental interventions in the initial phases of scene medical response. A nine-victim incident set in a family medicine clinic was scripted. Simulated terrorist detonation of a low-yield conventional explosive device dispersing a chemical payload of Lewisite (dichloro-2-chlorovinyl arsine) was selected (see Web Enhancement 1). The scenario was designed to have sufficient realism and plausible complexity to immerse participants in autonomous engagement of disaster response practices based on their judgment and training. The capacity for monitoring of distinct and quantifiable events and activities in disaster response was also essential. A seven-member regional panel of disaster experts and an international advisory board provided confirmation of content validity. Application of High-fidelity Medical Simulation to Disaster Scenario High-fidelity medical simulation has previously been used in military23 and hospital staff24–26 disaster training. Investigators staged the study sessions in a regional simulation center with digital video-recording and wireless communication capabilities. High-fidelity techniques were used for consistent and realistic rendering of disaster conditions and challenges. Contribution of fully-interactive manikin features to the disaster simulation was maximized. Physical Layout The scenario's simulated medical clinic was precisely mapped for reproducible environmental conditions (Figs. 1 and 2). Assorted debris, construction parts, and office supplies were used to construct the disaster setting (Fig. 3). Minimal ambient lighting, Fog F/X generator (MME, Farmingdale, NY) for simulated smoke and dust, and Firefly lights (ACR, Fort Lauderdale, FL) as fire-alarm strobes set up the incident scene.FIGURE 1.: Multipatient Mixed Modality Medical Simulation “Dirty Bomb” Scenario layout.FIGURE 2.: Multipatient Mixed Modality Medical Simulation “Dirty Bomb” Scenario patient descriptions.FIGURE 3.: Multipatient Mixed Modality Medical Simulation Disaster Scenario in progress. Adult manikin on right, actor victim on center left and four prehospital study subjects in protective gear.Victims Patients varied in severity of injury to reflect estimates based on the Centers for Disease Control Mass Trauma Casualty Predictor.27 Underlying illnesses and blast injury patterns combining blunt and projectile trauma were detailed. Airway compromise, thoracoabdominal hemorrhage, life-threatening fractures, and severe neurologic trauma were distributed across victims (Fig. 2). Progressive deterioration from injuries was coordinated with a scenario clock and exactly described by vital sign cards, actor instructions or manikin programming. Lewisite exposure was made apparent through prominent cutaneous blistering moulage, respiratory distress and vascular third-spacing pathophysiology.28–30 Combined traumatic and toxicologic insults were modeled when appropriate, along with psychologic trauma. Entrapment and immobilizing injuries kept patients situated within the clinic environment at time of study responder arrival. These arrangements permitted use of nonportable manikins and the study of initial onsite disaster medical responder actions (as opposed to hospital-based disaster response expected of later disaster phases). Experienced simulation staff members followed detailed standardized scripts to control two SimMan advanced medical patient simulators (Laerdal Medical, Wappingers Falls, NY). Allowing for invasive procedures and tight control of physiologic parameters, the SimMan manikins were programmed as entrapped clinic workers with major trauma and Lewisite exposure (see Web Enhancement 2 for sample programming). Changes in respiration, lung sounds, pulse, blood pressure, oxygen saturation and electrocardiographic signals were measurable by study subjects. SimMan victims were monitored for endotracheal intubation, IV access, medication administration, cardiopulmonary resuscitation (CPR), and defibrillation. CodeBlue III (Gaumard Scientific, Miami, FL) pediatric manikins represented a 1-year-old exhibiting cardiopulmonary deterioration and a 5-year-old with nonsurvivable injuries. These manikins were compatible with intubation, IV cannulation, and CPR. Five professional actors intensified the dramatic conditions necessary for the simulation. Actors were employed in the same role for sessions and were given instructions online to reduce intersession variability. They participated as immobilized clinic patients, parents of the pediatric victims, and a clinic nurse acutely suffering from critical incident stress. Each simulated or live patient was clothed and professionally moulaged with Casualty Simulation Moulage Kit (Simulaids, Woodstock, NY). Medical Response Actions Patients' vital signs were able to be assessed either through palpation of manikin chest rise and pulses or by review of laminated cards distributed to actors. IVs could be started in manikins or verbalized for patient actors. Stated recognition of the need for decontamination and antidote administration was considered adequate for the purposes of this simulation. Offsite communication was facilitated with a simulated 911 emergency dispatcher who arranged contact with local and regional facilities. Reasonable resources were made accessible on a predetermined timescale. Victim extrication involved spinal immobilization and activation of an offsite structural collapse crew. Transportation from the scene required proper communication with dispatchers regarding hospital destination, request for supplemental emergency medical services (EMS) units and transfer of care to receiving personnel. Medical Kits All study participants were issued standard kits based on state Advanced Life Support ambulance requirements. Basic airway devices and intubation kits were made available with oxygen sources. Model M CCT defibrillator/cardiac monitor (Zoll Medical, Chelmsford, MA), intravenous (IV) supplies, and medications were provided. Wound dressings, spinal immobilization equipment, and splints were integrated into the kits. All subjects had access to personal protective equipment (PPE) in the form of PA3NBC-filtered PA30 Powered Air-Purifying Respirators (PAPRs; Bullard, Cynthiana, KY) and Tychem SL impervious total-body gowns (Lakeland Industries, Ronkonkoma, NY). Team members communicated through portable Radius GP350 wireless radios (Motorola, Schaumburg, IL) with modifications for PPE. Selection and Development of Objective Measures An evaluation tool was designed concurrently with the SIM scenario for objective measurement of critical on-scene medical response and timeliness (see Web Enhancements 1 and 3). The tool was constructed with expert panel input for onsite observers to directly record clinical performance and timeline data distributed across nine concurrent movable patients without disrupting the scenario. Simple and readily recognizable actions were chosen as binary data elements. The first data element was the “entry–no entry” strategic response decision. At scenario initiation, responders were faced with limited dispatch information to a potential bombing incident. Encountering a nonsecured and disarrayed clinic environment with smoke and flashing fire alarm strobes, responders could verbally communicate across a closed entrance door with victims, who were specifically instructed to draw rescuer attention to their Lewisite symptoms (blistering and respiratory distress). After this opportunity for rapid formulation of scene assessment (ie, explosive device detonation accompanied by atypical cutaneous and respiratory symptoms in victims both within and outside the immediate blast radius), teams' decisions to enter into the simulated hazardous environment and to use PPE were recorded. Those deciding to enter the disaster using inadequate PPE or without requesting PPE would be halted and only allowed to re-enter after donning proper PPE or activating hazardous materials (HazMat) teams. (As the environmental hazards would “incapacitate” unprotected responders, this provision was necessary for study scenario continuation.) Appropriate treatment action (ATA) checklists, personalized for each of the nine victims, were the next evaluation tool data elements. Each ATA was one of nine clearly identifiable and observable actions (eg, administration of oxygen) and was chosen as a marker of responder activity. Every victim was a priori assigned ATAs with input from the expert panel. Initial location and contact by responders was observed for all victims. Recognition of fatality constituted ATAs for two victims. The seven remaining patients required administration of oxygen, removal from scene and arrangement of hospital transport. Specific victims were observed for additional ATAs such as extremity splinting, IV cannulation and advanced airway management. In all, 39 ATAs were distributed over nine victims in the scenario. Optimal response for study purposes required completion of all 39 ATAs by each study responder team. The number of ATAs performed per victim was recorded with timeline information. If a particular victim was assigned five distinct ATAs, a study team could be credited with a maximum of 5 unique ATAs for that patient. A separate data element, total ATAs, accounted for repetitions of ATAs performed on the same patient by a particular team (eg, starting two IV catheters). Repeated ATAs were not interpreted in either a positive or negative manner relative to team performance. Separate from ATA nonperformance, responder commission of actions harmful to victims was recorded by observers in a descriptive manner. Distinct from treatment actions expected of all study subjects, WMD-specific supplemental treatment actions (STAs) were also recorded. Teams' recognition of need for decontamination and British anti-Lewisite (BAL) administration was tracked for specific victims. These STA data points were not classified as ATAs, due to the need to differentiate WMD-related actions from “routine” disaster response activities. An additional data element, the victim transport order strategic response decision, was nested into the scenario and noted for each study team. This data was obtained to determine concordance with the optimal transport arrangement as designated by investigators. Parent-child victim pairs and two seriously injured victims entrapped by minor structural elements posed challenges. Participant Orientation Participants were recruited from statewide prehospital systems and block-randomized into 12 teams. Consent and signed waivers of liability were obtained. A comprehensive orientation with video briefing familiarized subjects to the high-fidelity medical simulation environment and manikins without revealing details of the study scenario. The interactive nature of simulated encounters and facilitator roles were presented in detail. A methodical safety review for all staff and participants was conducted prior to starting study sessions. The hospital Institutional Review Board approved the study. Scenario Implementation Twelve study sessions were initiated and conducted with strict adherence to scenario script and parameters by a Simulation Session Director. Synchronized time-dependent progression of actors and manikin systems' physiologic states maximized scenario automation. Unless subjects completed specific defined actions to alter scenario progression, each actor and manikin deteriorated across all sessions in an identical sequence and manner, as reflected in mental status, verbalizations, and vital signs. Unexpected events and deviation from the script were minimized by debugging through a prestudy pilot session with trial subjects. During sessions, investigator intervention was instituted infrequently and solely to provide requested information and equipment or to guarantee the safety of study subjects. Sessions were terminated upon scene evacuation/hospital transport arrangement for all movable victims and determination of fatality for nonresuscitable victims. Two clocks were used to time the occurrences of disaster response activities. A scenario clock reflected time within the simulated environment, whereas a study clock tracked the passage of real time. Both clocks were started upon scenario initiation. Victim morbidity at the time of initial responder contact was based on elapsed scenario clock time. Times to ATAs and STAs by both groups were measured from scenario entry by the unadjusted study clock with 1-minute resolution. This arrangement permitted adjustment of the scenario's temporal parameters while continuously recording real-time study participant action. Four video-cameras and five observers with study evaluation tools were positioned for data gathering (Fig. 1). Observers had been instructed in standardized guidelines on scoring with the tool and had participated in a preparatory scenario “dress rehearsal” and one pilot session with trial subjects. Research subjects' perceptions of SIM aspects of the disaster scenario were anonymously surveyed upon session completion. Data Analysis Quantitative descriptive analysis of study participant demographics, recorded data elements, and survey responses was performed. Spearman rank-order correlation analysis was completed for victim transport order by teams. Software package SAS version 9.1 (SAS Institute, Cary, NC) was used for analyses. RESULTS Characteristics of Study Subjects Study sessions were repeated twelve times over 3 months to enroll 48 participants. Most were fire department-affiliated EMTs at the Intermediate (EMT-I) certification level, with 10 paramedics (Table 1). Clinical experience averaged 8 years and 8% were female.TABLE 1: Demographic Characteristics of Study SubjectsSession Conduct and Clinical Performance Data Acquisition All sessions were completed as scripted and lasted 36.3±7.4 minutes. Scene entry and victim transport order data were collected for 100% of teams. Out of 468 potential ATAs and 168 potential STAs, 450 and 51 data points were recorded as being completed by study teams (Table 2). In all, 93.8% of ATAs were attributable to individual responders. Times were available for all ATA data (see Web Enhancement 1).TABLE 2: Clinical Performance Data Measures Recorded During Multipatient Mixed Modality Medical Simulation Disaster Scenario (n = 12 teams)Clinical Performance Data Nine of 12 teams entered the hazardous scenario environment without PPE at 46±26 seconds into the drill; three teams donned PPE and entered at 798±150 seconds. In all, 348 unique ATAs representing 74.4% of optimal response were recorded across 12 sessions, with a mean ± of per team. ATAs were with each team an of ± and for unique ATAs by individual study subjects were and and and 1 1 to for all nine victims per session ATAs, 100% of optimal and of nonresuscitable patients across all sessions administration was for seven of the nine patients based on scenario whereas IV cannulation have been for three of nine In all, and of disaster victims in were with oxygen and IVs applied a and on three patients whereas of victims was performed for of over all study sessions. Advanced airway endotracheal intubation was by three teams victims had to a hospital setting Times to first ATA from scene entry were 2 for all teams. Times of transport ATAs for movable patients were and from entry ± of decontamination and antidote administration was by teams for and of victims, of patient transport arrangement by teams concordance with optimal patient transport order order correlation Participant Response In all, study subjects completed a The simulation was a mean of in from the was and for was was and for participants. on scenario clinical and participant were conducted after sessions, as was a and study Both specifically the in clinical parameters disaster and simulated victim respect to psychologic with the study was from perceptions of individual performance team of medical response with limited resources a hazardous This aimed for positive of responder and as the study of on-scene disaster medical response limited in real-time and limited to medical simulation may a for disaster on investigators a repeatable WMD scenario fully-interactive medical simulation full-body pediatric professional a controlled simulated environment, and an objective evaluation was on detailed environmental and victim to reproducible Objective of disaster medical response actions were with expert panel input for the disaster scenario and used to data simulation sessions. The on onsite performance of response made through use of simulation and the comprehensive data recording by trained observers with systems and audiovisual on prior challenges noted such as in responders when specific actions and monitoring for in patient care across were by the use of a defined SIM study the objective are to disaster medical response to be for the of multipatient mixed modality medical simulation disaster in emergency study sessions an regional need to aspects of current medical response. the limited environmental in the absence of live Lewisite study subjects were observed to enter a clearly scene as in in provider and recognition of need for decontamination procedures and antidote administration were subjects receiving in WMD and of scene safety prehospital interventions such as and extremity immobilization were performed on a of victims to SIM and with respect to prehospital provider preparation may SIM to be a for the assessment of on-scene disaster response. If disaster may in prehospital PPE and training to and to with scene Disaster SIM may in and of onsite medical response training of the for disaster response and health care professional preparation in the of resources have been for emergency Both of could be facilitated through of disaster into the of such as through the Disaster Medical and Disaster Life Support may In 12 sessions of a objectively mixed modality simulation disaster were and response data recorded with the accompanying evaluation tool the to information regarding clinical performance in an immersive and medical simulation disaster study information regarding and of disaster scenario and evaluation tool has been presented for simulation in disaster research and training. The evaluation tool used in this study was not to study and session investigators were to teams without PPE entered of the due to the simulated nature of the or not the Simulation of medical response actions and comprehensive disaster response of incident perimeter and was not due to logistic and study or responder actions were recorded by were not into objective of scene response completion due to regarding and of negative The scenario was not to determine and could not to disaster The of simulation was not in that study sessions using disaster without medical simulation were not scenario was used to only medical response actions by EMTs from a without provision to follow up on Objective assessment of intersession was not performed. An online only available for this