Prolonged, High-Fidelity Simulation for Study of Patient Care in Resource-Limited Medical Contexts and for Technology Comparative Effectiveness Testing

Supplemental Digital Content is available in the text.

Abstract

OBJECTIVES:

Most high-fidelity medical simulation is of limited duration, used for education and training, and rarely intended to study medical technology. U.S. caregivers working in prehospital, resource-limited settings may need to manage patients for extended periods (hours to days). This “prolonged casualty care” occurs during military, wilderness, humanitarian, disaster, and space medicine. We sought to develop a standardized simulation model that accurately reflects prolonged casualty care in order to study caregiver decision-making and performance, training requirements, and technology use in prolonged casualty care.

DESIGN:

Model development.

SETTING:

High-fidelity simulation laboratory.

SUBJECTS:

None.

INTERVENTIONS:

We interviewed subject matter experts to identify relevant prolonged casualty care medical challenges and selected two casualty types to further develop our model: a large thermal burn model and a severe hypoxia model. We met with a multidisciplinary group of experts in prolonged casualty care, nursing, and critical care to describe how these problems could evolve over time and how to contextualize the problems with a background story and clinical environment with expected resource availability. Following initial scenario drafting, we tested the models with expert clinicians. After multiple tests, we selected the hypoxia model for refinement and testing with inexperienced providers. We tested and refined this model until two research teams could proctor the scenario consistently despite subject performance variability.

MEASUREMENTS AND MAIN RESULTS:

We developed a 6–8-hour simulation model that represented a 14-hour scenario. This model of pneumonia evolved from presentation to severe hypoxia necessitating advanced interventions including airway, breathing, and shock management. The model included: context description, caregiver orientation scripts, hourly progressive physiology tracks corresponding to caregiver interventions, intervention/procedure-specific physiology tracks, intervention checklists, equipment lists, prestudy checklists, photographs of setups, procedure, telementor, and role player scripts, business rules, and data collection methods.

CONCLUSIONS:

This is the first standardized, high-fidelity simulation model of prolonged casualty care described in the literature. It may be used to assess caregiver performance and patient outcomes resulting from that performance during a complex, 14-hour prolonged casualty care scenario. Because it is standardized, the model may be used to compare differences in the impact of new technologies upon caregiver performance and simulated patient outcomes..

Medical care in extreme circumstances is challenging. Austere, resource-limited clinical contexts (1, 2) make casualty evacuation and resource availability/resupply less reliable and “golden hour” (3) evacuation to surgical resuscitation unfeasible. Consequently, caregivers with limited training and experience must manage complex casualties for extended periods of time (4, 5). The military has coined this contextual problem as “delayed evacuation,” “prolonged care,” or “prolonged casualty care” (PCC) depending on the reference and, in many ways, it is similar to wilderness, disaster, refugee, and space medicine (2, 6). The clinical context of PCC tests the limits of caregivers’ knowledge, skills, and abilities to make accurate, reliable, and efficient decisions and to perform necessary interventions and procedures until evacuation to a higher level of care.

Essential to optimal care in any context is an effective judgment: training and education do not always translate to desired results as they cannot perfectly replace real-life experiences. High-fidelity simulation can augment experience, improve clinical judgment, and hone procedural skills (7–9). Indeed, there are many currently available simulation models developed and used in training for specific procedures or tasks, but none match the potential complexity or duration of care anticipated in PCC.

Additionally, it is impossible to prepare PCC clinicians for all possible contingencies. In usual settings, this challenge is mitigated by specialist consultation and/or patient transfer to “higher” levels of care. When these are not possible in the context of PCC, technology solutions may enhance a caregiver’s capability to work beyond their normal scope of practice (10–12). Telemedicine is one such technology that has, on the whole, proven effective in the civilian literature to improve patient outcomes and reduce costs (13). The U.S. Military has been particularly interested in telemedicine capabilities for nearly 30 years (1) and more recently considered its potential to address the challenges faced in PCC (14, 15).

To effectively study caregiver effectiveness during PCC and test the impact of different technologies on caregiver performance, it is necessary to have standardized care models of fully contextualized PCC that allow for subject immersion in the scenario (16). Unfortunately, no such models of PCC exist. Task trainers and narrowly focused, objective-based, short duration training models do not account for resource limitations, fatigue, distraction, location changes, and environmental factors like heat, cold, rain, and darkness faced in PCC. Although technology is intended to improve work performance and reduce cognitive load, new technology may create unintended consequences and have variable effects on performance and cognitive load (17–19).

Similar to a well-characterized large animal model of hemorrhage, we sought to develop a high-fidelity, contextually relevant, standardized model of PCC including standardized outcome metrics that can be used to study caregiver performance and the impact that different technologies—like telemedicine—might have on caregiver effectiveness. The purpose of this paper is to describe how we developed the PCC model and our intent to study the impact of telemedicine on caregiver performance and simulated patient outcome during PCC.

METHODS

Model Development: Selecting the Clinical Scenario

We conducted informal interviews with military experts who identified the most difficult scenarios faced during PCC: large thermal injuries, crush with acute kidney injury, closed head injury, chemical/biological/radioactive/nuclear injuries; polytrauma with internal hemorrhage, extended tourniquet placement, sepsis, and hypoxia due to severe lung injury. Next, several “table-top” exercises (i.e., talking through a scenario from start to finish) identified the potential care tasks within a scenario and model characteristics necessary to test clinicians performing PCC of these casualty types. Models needed to be complex, rely on precise and accurate clinical judgment, evolve over many hours, and require knowledge or skills beyond those obtained during expected training for anticipated PCC caregivers like medics or general medical providers—including physician assistants—and require common (e.g., airway management) and uncommon (e.g., advanced ventilator management) prehospital procedural competencies. See Table 1 for specific objectives from two scenarios.

TABLE 1.

Comparison of Model Objectives Between Two Piloted Scenarios

Objectives	Burn Scenario	Hypoxia Scenario
Initial assessment and interventions	Accuracy of burn size	Identification of possible causes of hypoxiaAdministration of antibiotics for pneumonia/sepsis
	Initial fluid resuscitation rate
	Recognition of need to evacuate/attempts to do so
		Recognition of need to evacuate/attempts to do so
Resuscitation	Completion of Lund and Browder chart	20–30 mL/kg fluid bolus according to sepsis guidelines
	Calculation of formal fluid requirements	Urinary catheter placement
	Urinary catheter placement
Identification and management of complications	Lower extremity eschar syndrome	Worsening hypoxia with fluid resuscitation
	Lower extremity escharotomy
		Cessation of crystalloid fluids
Ongoing resuscitation	Adjustment of hourly fluid rate according to urine output	Tolerance of mild hypotension given adequate urine output
	Wound care	Initiation of oxygen therapy
		Patient positioning (upright)
		Possible noninvasive positive pressure ventilation
Identification and management of complications	Recognition of the need for intubation to protect airway	Recognition of need to intubate to protect airway and provide mechanical ventilation
	Intubate or perform cricothyrotomy
		Intubate or perform cricothyrotomy
Ongoing management	Analgesia and sedation management	Analgesia and sedation management
	Continued fluid adjustment according to urine output
Identification and management of complications		Tension pneumothorax from positive pressure ventilation
		Needle decompression followed by tube thoracostomy
Documentation	Care documentation accuracy and completeness throughout	Care documentation accuracy and completeness throughout
Duration	8 hr real time, 14 hr simulation time	8 hr real time, 14 hr simulation time

Both burn and hypoxia (acute respiratory distress syndrome [ARDS]) scenarios were initially developed and piloted. After pilot testing, however, the ARDS scenario was found to have more complexity, be less familiar to our target subject population (medics and general medical providers), and thus have more opportunities for a tool or technology solution (e.g., telemedicine) to demonstrate potential value. Therefore, we describe the development of the ARDS PCC high-fidelity human simulation model here.

Model Details

The simulated patient model required great thought, multiprofessional input, and iterative pilot testing. Aspects of the scenario were considered (Table 1), and then different tracks of patient disease and physiology progression were developed to account for anticipated caregiver (future subject) interventions (Supplementary Material, http://links.lww.com/CCX/A705). Anticipated subject actions like airway management could occur early or late in the scenario, must be accounted for by the model, and designed to be consistent between tests. For specific pharmacologic and procedural interventions, we designed simulated physiology to change in accordance with dose or successful procedure completion. After developing scenario progress tracks, we wrote background scripts about the patient and caregiver to promote an immersive, imaginative subject experience and thus rapid scenario engagement.

Following this initial development, the model was repeatedly tested and enhanced. First, “experienced” clinicians (i.e., senior medics, critical care nurses, and critical care trained physicians) participated in the simulation scenario in real time and provided feedback on the perceived believability. These pilot sessions identified necessary scenario modifications and “business rules” that enhanced the research team’s ability to “consistently” administer the scenario. Following tests with experts, we piloted the scenario with inexperienced clinicians (e.g., junior medics, interns) that represented the target subject population. Adjustments made following observations during these tests focused on balancing realism with scenario flow/completion. For example, the research team noticed that few medics were familiar with urinary catheter placement, thus creating a barrier to scenario progress. Consequently, we allowed a scenario “confederate” (i.e., a research team member embedded in the scenario that helps subjects engage with the scenario effectively—see below) to insert the urinary catheter. Physiologic response to medication doses were stratified to simulate a realistic dose response but to ensure scenario completion allowed life-threatening overdose (i.e., a subject could give super maximal doses of a medication but only experience a “too high” dose model response, instead of death, see below). This allowed scenario completion and correctness of medication dose to be a recorded quality metric. Consistency was established and subsequently maintained by two research teams conducting regular weekly meetings in which they discussed simulations, watched video recordings of each other’s performance, and, when necessary, agreed to the business rules noted above. Believability was established by asking pilot subjects if the scenarios, simulated physiology, equipment availability, and expected subject tasks made sense and could be expected during real patient care in a PCC setting.

Outcomes and Performances Metrics

The primary outcome metric was simulated patient survival. Secondary outcomes included the accuracy, efficiency, reliability, and quality of caregiver interventions (Table 2) (Supplementary Material, http://links.lww.com/CCX/A705). We chose the National Aeronautics and Space Administration Task Load Index (NASA-TLX) to measure cognitive load for individual tasks and scenario completion as a whole (20–22). An example of quality related to opioid therapy was the dose amount stratified as too high, too low, or appropriate.

TABLE 2.

Outcomes to Be Measured During the Simulation Model and Their Definitions

Outcomes	Definitions
Accuracy	The right decisions are made according to predetermined optimal patient care as judged by experts who create the testing scenario.
Reliability	Reproducibility of accurate decisions across subjects within group (intervention vs control) assignments.
Efficiency	The timing of decisions and the duration of procedures.

PCC requires care over an extended time, but the necessity of scenario completion, access to simulation labs, and subject availability required us to develop novel approaches to scenario time management including real time, episodes with skip periods, and compressed time. Similarly, it is impossible to identify all variations of subject action in a complicated, contextualized simulation model. To mitigate this uncertainty, we collected simulation team responses to a subject’s actions during pilot testing and built new business rules or physiology tracks to account for these variations (Supplementary Material, http://links.lww.com/CCX/A705).

We identified resources that PCC clinicians might expect to have available and standardized sets and setup of equipment, simulated medications, and other supplies (Supplementary Material, http://links.lww.com/CCX/A705). To balance model consistency (standardized equipment) with model realism (most providers pack their own equipment and train with it before use with real patients), we included a plausible explanation in the background story and made inventory location diagrams, color-coded simulated medications (vials of sterile water with simulated labels for IV medications and various hard candies for oral medications), mandated subjects orient to the care environment (e.g., notional clinic) (Fig. 1) and resources on the night prior to study, and allowed subjects to arrange them to their preference on the morning of study.

Figure 1.

The scenario setup is designed to feel like a “house” phase of prolonged flied care (6). The “Sim-Clock” can be seen in (A) top middle and (B) bottom right. Cameras are mounted high and at corners of the room to ensure optimal data capture. The room setup is photographed and can be reset to match these photographs after a subject completes a study.

Model Development: Technology Evaluation

We intend to study technology and how it impacts caregiver performance during PCC using this model. We plan to use telemedicine as the first technology for randomized testing, results of which will be reported in future publications. To minimize bias of the individual performing the telementoring, we also standardized telementor interactions by creating a telementor script (Supplementary Material, http://links.lww.com/CCX/A705).

RESULTS

We simultaneously developed burn and ARDS scenarios until we decided to finalize the ARDS scenario for initial technology studies. Many of the tasks necessary to complete the scenarios (i.e., airway management, ventilator management, etc.) were common to both scenarios. We pilot tested models 23 times (14 burn, 9 ARDS), in their entirety or in parts, with experienced (10 burn, 5 ARDS) and inexperienced (4 burn, 4 ARDS) clinicians. When queried, pilot subjects confirmed the realism and believability of the scenario’s context, resources, disease progression, and expected patient management by describing similar real-life experiences that were similar to their simulation experience. We were unable to collect survey data from our pilot subjects due to institutional review board (IRB) restrictions; however, we intend to report data from subjects who delivered care to the ARDS model according to IRB-approved protocols in the future. Here, we describe the final 6–8-hour PCC simulation model representing 14 hours of pneumonia care evolving to ARDS using the components which are summarized in Table 3 and available in the Supplementary Material (http://links.lww.com/CCX/A705). Figure 1 shows scenario setup.

TABLE 3.

Model Components, Their Description, and Their Purpose

Model Components	Description/Purpose
Scenario context description	The scenario context description provides a narrative the care context that subjects must imagine themselves to be a part of as they begin their care of the simulated patient. The description must be relevant to the subjects so they can imagine themselves in this context and how they would proceed with patient care were the scenario to be real.
Caregiver orientation scripts	The orientation scripts must: 1) instruct the subject about the simulation equipment and how it works, 2) orient the subject to the environment and the additional people they will interact with during the study, 3) data collection methods and tools they may interact with (e.g., cameras, physiology sensors), 4) allow them to explore the research environment prior to starting the study. We recommend a day-before and an immediately before (day-of) orientation. The day before orientation orients the subject to all but the scenario context; it allows the subject to explore the equipment and supplies available and to reflect on how he/she might conduct themselves during the scenario. The day-of orientation introduces the subject to the scenario context and allows him/her to organize the medical equipment and supplies to his/her liking prior to starting the scenario.
Hourly patient physiology tracks that correspond to caregiver interventions	These are the hardest scenario artifacts to produce because of the potential variability of subject performance. A subject who completes a procedure immediately in a scenario causes the simulation to proceed much differently than a subject who delays an intervention until much later in the scenario. Balancing complexity for the proctor/simulation technician and realism for medical care are constant challenges. Importantly, for unexpected care/decisions made by subjects, the simulation must respond consistently across subjects and across locations. For each “new” path a subject creates through variance in care, the research team must create a new physiology track or a business rule (see below) for how the research team and simulation will respond. This is the primary reason why a proctor should have medical training beyond the level of anticipated subjects during scenario development.
Physiology tracks specific to an intervention/procedure	Physiology tracks must be more detailed than hourly when subjects perform a procedure. For example, decompression of a tension pneumothorax must produce near immediate physiologic response in the simulated patient otherwise the simulation may cause cognitive dissonance for the subject who understands that physiologic improvement indicates a successful procedure. Similarly, the simulated patient’s physiologic response to an intervention, such as endotracheal intubation, must be well considered by the research team in the context of the research purpose. If an intubation takes too long, should the simulated physiology worsen? If so, how quickly? If the subject has difficulty with the procedure, should this cause the simulated patient to “die?” Again, it will be necessary to create physiology tracks or business rules for each of these alternative pathways to maintain consistency across subjects and study locations.
Anticipated caregiver interventions according to time/patient condition	The anticipated timeline is essentially the “correct” patient management and is used to measure caregiver performance metrics of accuracy, reliability, and efficiency. This timeline is both imagined during the scenario development and pilot tested with expert and novice subjects to ensure the baseline, or control group, is well described.
Intervention checklists, equipment lists, prestudy checklists	Checklists are essential for ensuring the simulation is reproducible. Scenario setup—the room appearance, the manikin setup, the control room monitors, the supplies—must be governed by checklists. Research personnel must practice using the checklists, preferably in a read-confirm team to ensure all necessary steps are identified and carried out. Similarly, within the scenario, caregiver procedures must be scripted by checklist so that a baseline of completeness is established. Because there is a high degree of variability in how caregivers perform procedures, procedural checklists must only account for the key steps of a procedure. Too much detail about steps that do not affect the success of the procedure are irrelevant. An exception to this rule, for research studies involving telemedicine, is the need to develop consistent telementor checklists/scripts such that telementors instruct subjects to complete procedures the same way each time.
Photographs/videos of setups or disease as necessary	Photographs that show how to consistently set up the scenario room, the manikin, the supplies, the control room, etc. are a best practice for consistency. Photographs particularly help to maintain consistency if new research team members perform setup. Photographs or videos of real patients (with consent) may be shown to subjects so they can better imagine the problem(s) they will manage during the study. For example, a video showing progressive respiratory distress helps subjects to understand the impact of hypoxia on the simulated patient. It can also add elements of sound to the scenario that might be difficult to consistently reproduce otherwise.
Role player scripts	Scripts are essential for anyone interacting with the subject during the scenario so that interactions are consistent across subjects and locations. The confederate must have a script defining his/her interactions with the subject. Similarly, the simulation technician who acts in the role of patients, must have a script defining what he/she can say on the patient’sbehalf. Other roles, like a concerned family member or a military commander, must be considered and scripted to ensure scenario realism. Telementors must have each interaction with the subject scripted. In situations when the subject varies from the anticipated flow of the scenario, the team must practice staying on script. If forced to leave script, the new narrative used by the role player must be captured and incorporated into the script or into a business rule such that all similar interactions in the future will remain consistent.
Business rules	Business rule help govern how the research team manages complexity during the scenario. These rules define when the scenario will be realistic and follow a physiology track or be unrealistic but allow the scenario to move forward. Some rules about sedative medication are described in the text. Additional rules may be found in the supplementary material.
Data collection methods	These are study specific but likely include audio/video recordings, surveys, and potentially subject physiology.

To address time constraints, we created “simulation time:” when few physical tasked occurred, we compressed 1 hour of simulation time into 15 minutes of real time. Because some physical tasks take time to complete and to not rush subjects, the scenario runs in “real-time” (1 hr = 1 hr) during the initial patient assessment, during the initial telementor session, and during all procedures. A custom software application (the “Sim-Clock”) tracks real time but displays simulation time to subjects to minimize cognitive dissonance as time advances. Overall, these time adjustments allow a 14-hour scenario to complete over a 6–8-hour study period. Completing the scenario during normal work hours is important for subject recruitment and resourcing research staff.

We use a SimMan 3G (Laerdal, Stavanger, Norway) manikin to represent realistic patient physiology and interactions. Hourly vital signs were programmed into the Laerdal Learning Application, LLEAP (Laerdal) according to the developed physiology tracks, including tracks for successful procedures that reverse programmed trends of worsening vital signs (e.g., the near instantaneous improvement of vital signs following decompression of a pneumothorax). The scenario is designed for gradual worsening of simulated physiology that can be temporarily improved by successful interventions or procedures (Supplementary Material, http://links.lww.com/CCX/A705).

Scenario Context and Physical Setup

Our scenarios take place in a fictional, austere location in Africa in a resource-limited clinic. This context is appropriate for subjects of a general practitioner background (e.g., medics, mid-level providers, and general practitioners) but not specialists, subspecialists, or surgeons. Study proctors control the mannequin’s physiology and physical examination from a separate room. Vital signs are displayed through a tablet monitor in the simulation room. The study room is monitored by the research team in the control room using audio transmitted by a baby monitoring system. A radio hidden in the confederate’s ear allows for direct communication between the proctor and the confederate. Streaming video from the study room to the control room facilitates recording of the scenario and continuous visual assessment by the proctors. The subject is able to communicate with their “higher headquarters”—a role played by the proctor—using a two-way radio. The subject is able to call the telementor using a phone or cellular phone with video-teleconferencing software like FaceTime (Apple, Cupertino, CA), Skype (Microsoft, Redmond, WA), or Google Meet (Alphabet, Cupertino, CA).

To ensure that all subjects are equally knowledgeable about the scenario, “day-before” and “day-of” scripts are read to subjects to orient them to the scenario background, manikin capabilities and limitations, available supplies and equipment, and subject monitoring equipment (see below and Supplementary Material, http://links.lww.com/CCX/A705).

Personnel

Multiple scripts were written and revised for the various roles in the scenarios (Supplementary Material, http://links.lww.com/CCX/A705). A minimum of three individuals are necessary to run the scenario: a proctor, a simulation technician, and a “confederate” (Supplementary Material, http://links.lww.com/CCX/A705). Additionally, a telementor is necessary for studies that include telemedicine.

The proctor is responsible for ensuring the overall scenario runs according to plan and should have medical training (e.g., nurse, physician assistant, or physician) to be successful. The simulation technician ensures the high-fidelity manikin runs properly and displays the appropriate physiology directed by the proctor. The confederate works directly with the subject to 1) increase engagement with the simulation by promoting scenario realism and 2) troubleshooting technical problems during the scenario that might affect good data collection (e.g., a video camera gets unplugged).

Role scripting and recorded practice followed by review maximized consistency and reproducibility. A script was written for the simulated patient, including relevant social, medical, and military history, with key written phrases that the simulation technician would use to “talk” to subjects through the manikin speaker. A telementor script was written to ensure the telemedicine “technology” was controlled across studies and did not cause undue bias in the research (Supplementary Material, http://links.lww.com/CCX/A705).

Because not all subject interactions with research staff can be scripted, “business rules” provided boundaries and expectations for nonscripted interactions. The confederate role was designed as a lay person who has “picked up” some medical skills volunteering in the clinic but has no medical knowledge. This allows them to facilitate the simulation (e.g., help the subject find equipment) and perform nonmeasured tasks (e.g., urinary catheter placement) necessary to help the scenario move forward. If a subject asks the patient for history that is not part of the script, the technician can repeat script information or deflect the question by saying “I’m not sure about that.” If asked for advice about an intervention or treatment, the confederate avoids encouraging or discouraging particular actions by saying “I trust your judgement.”

Data Collection Methodology

The model is designed to collect data that measures caregiver decision-making with respect to accuracy, reliability, and efficiency (Table 2) as well as subject (caregiver) cognitive work and stress. This is measured by the NASA-TLX with subcategories of effort, frustration, mental demand, physical demand, temporal demand, and performance (21–23) and using subject physiologic variables collected by wearable sensors and analyzed using heart rate and respiratory rate variability metrics (24, 25). Scenarios are recorded using at least two digital cameras: one camera with a wide-angle view from above and another positioned on the confederate’s chest so the confederate can obtain a close-up view of activities. Placement on the confederate’s chest allows the proctor to watch the video from the control room and, using the confederate’s earpiece, direct the confederate to a position that obtains ideal footage. Video footage is subsequently used to confirm that proctors consistently administer the scenario, complete case report forms, and adjudicate findings.

DISCUSSION

This is the first description of a standardized, prolonged simulation model for assessing medical care in a PCC context. The ability to study care in austere settings and resource-limited contexts like war, disaster, or pandemic response is extremely limited. Using a high-fidelity simulation model allows researchers to better understand how PCC caregivers use judgment to make value-based decisions regarding optimal care using available resources. The model was repeatedly tested and refined using feedback from PCC and critical care experts until it was consistently believable by experts asked to manage the simulated casualty.

To our knowledge, this is the longest and most contextually accurate model of PCC described in the literature. Our approach is similar to other high-fidelity simulation models evaluating medical care and technologies in realistic hospital settings (26, 27), but it is developed for study of care in a markedly different context. High-fidelity simulation has been used to study procedural telementoring (28), but not in a resourced-limited care context such as ours nor intended to assess decision-making “over time.” The successful simulation of the PCC scenario with multiple pilot subjects established a standardized, repeatable, and believable simulation scenario that may establish baseline performance metrics to compare with performance using new processes, tools, or technologies. Scripting and review of video recordings facilitate quality control and maintain consistency across the research team role players, study sites, and for individual study events.

Thus far, we have used pilot test videos to practice completing case report forms (Supplementary Material, http://links.lww.com/CCX/A705) intended for research studies. This work demonstrates the practicality of our model for identifying important data points and adjudicating observations. This process can effectively convert the complex data of clinical care into useful, simplified, definitive data points that can be analyzed and compared between groups.

The simulation format was intentionally developed to be overtly stressful to nonexpert medical professionals and to properly simulate the PCC environment. This scenario takes into account the complex nature of decision-making and the effect of stress upon nonexpert providers. This specific design will be useful for research; it may also be used to train clinicians before military or civilian deployment to resource-limited care environments like war zones and disaster response. The model provides an opportunity to experience the practical and psychologic challenges of PCC before experiencing them during real patient care.

The model has some limitations. As described, it requires tremendous human resources: three researchers, and one telementor with dedicated time away from clinical duties for the duration of the scenario (6–8 hr). Despite its immersive design, subjects may not fully engage in the scenario and may make different choices for the simulated patient than a real patient hindering generalizability (29). Noteworthy, when planning or creating long duration simulation studies, technology limitations in data collection tools (e.g., battery life, variability in network connectivity, overheating of devices, manikin automatic shutdown) must be taken into account. Our creative use of time, while necessary to increase subject recruitment, optimize research staff quality of life, and mitigate logistical challenges (e.g., after-hours building security), is also a potential limitation. “Sim-time” compression can rush a subject and reduce boredom during routine care, which was felt by some early test subjects during a long “real-time” simulation. Increased time pressure and decreased boredom affect decision-making, stress, and immersion (16), potentially introducing bias. We plan to test this hypothesis in future studies. Finally, complete removal of bias is not feasible with human actors. Although scripting mitigates bias, automation may maximally remove human bias. In future efforts, we intend to remove the simulation technician and proctor using automated physiology engines that respond to caregiver intervention.

CONCLUSIONS

We successfully developed a standardized, repeatable, and believable PCC simulation model. This model can be used to test the impact of telemedicine or other technologies like clinical decision support tools on caregiver performance during PCC. As additional casualty models are developed (e.g., trauma models) for studying PCC management, simulation could be added to create multicasualty scenarios and/or additional healthcare team members could be added for studying team dynamics or the impact of teams on clinical decision-making or both. Additionally, the model has the potential to test telementor efficacy via scripting of subjects or telementor/provider interactions through unscripted iterations. Ultimately, we intend to use high-fidelity, standardized simulation to study care in uncommon clinical contexts (military and civilian) and to use this understanding to better develop and test technology solutions in these challenging environments.

Supplementary Material

cc9-3-e0477-s001.pdf ^{(2.8MB, pdf)}