Simulation to Implement a Novel System of Care for Pediatric Critical Airway Obstruction

Objective To implement a novel system of care for pediatric critical airway obstruction.

Design Retrospective, observational study of data gathered prospectively during high-fidelity simulations.

Setting Emergency department (ED) and operating rooms (ORs) of a pediatric referral center.

Subjects Health care provider simulation participants.

Main Outcome Measures Time from ED attending physician request to arrival of an otolaryngologist, participant survey responses, identified latent safety threats, and simulated patient outcomes.

Methods Twelve high-fidelity simulations were conducted: 6 to identify problems with an existing system of care, and 6 to implement a novel system. The simulation scenarios involved a 4-year-old patient with severe respiratory distress after foreign-body aspiration managed solely in the ED or in the ED and OR, depending on stability.

Results There were 196 participants in 12 simulations. The mean (SD) time from ED attending physician request to otolaryngologist arrival was 7.8 (1.6) minutes for the existing system simulations and 5.0 (1.1) minutes for the novel system (P = .001). Latent safety threats identified in the simulations included a lack of specialized airway equipment in the ED. Death of the simulated patient occurred in the ED in 2 of 6 existing system simulations; specialized airway equipment was available for neither. For the novel system simulations, specialized airway equipment was available for all 6, no simulated patient deaths occurred.

Conclusions High-fidelity simulation was an effective method to design and implement a novel system of care for pediatric critical airway obstruction. The novel system was associated with more rapid response times and elimination of simulated patient deaths.

Critical airway obstruction, which we define as an acute disruption of normal airflow in the airway sufficient to significantly compromise ventilation and/or oxygenation, can be rapidly progressive and life-threatening, particularly in children. Although certain conditions causing pediatric critical airway obstruction, such as epiglottitis, have become rare, others are still relatively common. Foreign-body aspiration has been reported to occur up to 2.5 million times annually in the United States and to result in as many as 2000 deaths.¹ In a study of a national database of pediatric hospital admissions, the mortality rate for foreign-body aspiration was approximately 3.4%.²

Management of critical airway obstruction must be delicate, thoughtful, and efficient. This is particularly true in children, who are recognized to present unique airway challenges and frequently require subspecialty management by an otolaryngologist and/or an anesthesiologist. Critical airway obstruction is a high-acuity but low-frequency event, and even the collective experience of the care team may be limited. Published reports suggest that acute epiglottitis is best managed through a predefined, coordinated system of care, including the requisite personnel and equipment and a standardized approach to common patient scenarios.^3-5 In the emergency department (ED) of our institution, the management of patients with critical airway obstruction has typically occurred ad hoc, with no coordination of a multidisciplinary response.

Medical simulation has been increasingly used for education and process improvement, particularly for high-acuity, low-frequency events.^6,7In situ simulation, or medical simulation that is physically integrated into the clinical environment,⁸ has been shown to be effective to evaluate new care environments and team composition.⁹ It has also been used in a variety of fields to improve hospital protocols or systems of care.^10,11 Within otolaryngology, medical simulation has been found to be useful in simulating pediatric airway endoscopy and foreign-body removal in the operating room (OR).^12-14

We reviewed the case of a child who presented to our institution's ED with critical airway obstruction due to foreign-body aspiration. Despite rapid diagnosis and subspecialty consultation, there was a delay in the transportation of this patient to the OR. The patient did well following endoscopic removal of the foreign body, although the delay in transport placed the patient at unnecessary risk of deterioration. This case prompted a detailed review of the existing system of care for all children with critical airway obstruction who present to our institution's ED, during which we designed a study using a series of in situ simulations. To our knowledge, no study has used simulation to implement a system of care for critical airway obstruction, and very few reports even describe such a system.^15,16 The objective of this study was to describe the implementation of a novel system of care for children with critical airway obstruction presenting to a pediatric ED, accomplished through (1) the use of in situ simulations to assess the existing system of care, (2) the design of a novel system of care based on findings from these simulations, and (3) the assessment of the novel system's impact on the timeliness and quality of care for simulated patients.

This was a retrospective study of data gathered prospectively from 12 in situ simulations of critical airway obstruction in a pediatric patient: 6 conducted to describe the existing system of care and 6 to assess the implementation of a novel system. Institutional review board approval was obtained prior to study commencement.

All simulations were performed in the ED resuscitation area of our institution, a tertiary-care children's hospital. The ED is the major provider of emergency health care to children in its region, with approximately 90 000 visits annually and 4 ED resuscitation bays. Airway equipment available in the ED includes standard and video laryngoscopes, laryngeal mask airways, and a tracheotomy tray. Prior to the development of the novel system of care, there was no bronchoscopic equipment readily available in the ED.

For this project, we developed a simulation scenario involving a 4-year-old boy who, after suffering a witnessed aspiration of a grape, is brought to the ED by ambulance. For each group of 6 simulations, 3 involved the patient remaining stable in the ED with a partial airway obstruction if managed appropriately and then being taken to the OR for foreign-body removal. In the other 3, the simulated patient developed complete airway obstruction shortly after arrival, forcing the care team to manage the patient in the ED resuscitation bay. The manikin used was the METI Pediasim ECS (METI Corporation) with a piece of green sponge in the trachea as a foreign body, and a 4.0 Storz bronchoscope (Karl Storz GmbH & Co KG) available for OR removal.

The simulations began with the designated ED care team being alerted by page that a child required evaluation in the resuscitation area. Health care providers in the ED were not aware that it was a simulation until they arrived in the resuscitation area, a standard practice for in situ simulations in our ED. Following arrival of the care team, a study investigator read the arrival information, including that the patient was in severe respiratory distress but maintaining oxygen saturation at 98% on a simple face mask. The team was then shown a video of a real child with severe stridor and retractions, which included audio. Per the existing system of care, ED airway emergencies were managed initially by the ED care team only, with the otolaryngology resident and anesthesiologist available for consultation. The otolaryngology resident, if consulted, would need to select indicated specialized airway equipment from the OR and proceed to the resuscitation bay. Separate calls were required to consult anesthesiology, OR staff, and/or pediatric surgery.

Our methods generally conformed to published guidelines for observational studies (STROBE recommendations¹⁷). Data were collected during the simulations by one of us (J.O.), using a standardized study form. Data collected included multiple time points: time of arrival of the ED team leader, otolaryngology resident, anesthesiologist, and airway equipment; time that pages were sent to otolaryngology and anesthesia; and time of scenario termination or transport of the simulated patient to the OR. Following the simulations, each participating health care provider was asked to complete an anonymous survey with questions regarding their previous simulation experience and their knowledge of and opinions regarding the effectiveness of the new system of care (eFigure 1 and eFigure 2). Following primary data collection, videos of the resuscitations were reviewed by 3 of us (K.J., J.O., and B.K.), both to confirm by consensus the data obtained through direct observation and to collect any missing data available on the videos.

The primary outcome was the time from ED team request to arrival of the otolaryngology resident. Additional outcomes were the time from ED request to the arrival of the anesthesiologist on call, the availability of specialized airway equipment in the resuscitation area, the survival of the simulated patient, and identification of any latent safety threat and/or team-level knowledge deficit. Latent safety threats were defined as system-based threats to patient safety, previously unrecognized by health care providers that could materialize at any time.¹⁸ Knowledge deficits were defined as departures from standard care directly related to deficiencies in the cognitive or technical proficiency of the health care providers.⁸

All data were tabulated and descriptive statistics generated for outcomes of interest and important data elements. Categorical variables are reported as absolute numbers and/or proportions. Continuous time data are reported as means and standard deviations. Time differences between the existing system and novel system simulations were analyzed via absolute differences along with the P value from a 2-sample independent t test for the difference in mean time. Statistical significance was set at P ≤ .05. All analyses were performed using SAS version 9.2 (SAS Institute Inc).

The novel system of care, developed following the first 6 simulations and tested in the second 6, consisted of (1) a critical airway team made up of health care providers from otolaryngology and anesthesiology, a respiratory therapist from our institution's complex airway unit, and OR front desk staff; (2) a written algorithm for the care of ED patients with critical airway obstruction (eFigure 3); (3) an airway cart with specialized equipment specific to the critical airway team (Figure); and (4) a critical airway team–specific paging system. The paging system simultaneously informs each critical airway team member of the event, and the airway cart is then transported to the resuscitation area by the respiratory therapist. The critical airway team is intended to be activated for any patient in the ED resuscitation bay with apparent critical airway obstruction for whom the ED care team anticipates the need for advanced airway intervention including bronchoscopy and/or a surgical airway.

The 6 simulations of the existing system were performed from November 2010 to January 2011. After a period of education, 6 simulations of the novel system were performed between June and July of 2011. Five of 17 scheduled simulations were cancelled (29%) owing to conflicts with real patient care. Surveys were collected from 160 of the 196 health care providers who participated in the 12 simulations (81%). Thirty-five percent of participants were physicians, 33% nurses, 14% respiratory therapists, and 17% categorized as “other.” Eighty-nine percent of participants had some prior simulation experience. Following education and then simulations of the novel system of care, 90% of health care providers responded that they knew how to activate the system, and 95% knew what the indications were for activation. When asked if our institution was better prepared for a real case of critical airway obstruction because of the critical airway team, the mean provider response was 6.4 on the 7-point Likert scale, with a median score of 7 (strongly agree).

The mean (SD) time from ED attending physician request to otolaryngologist arrival was 7.8 (1.6) minutes for the existing system simulations and 5.0 (1.1) minutes for the novel system simulations (absolute difference, 2.8 minutes) (P = .001) (Table). The times from simulation start to the ED team's request for otolaryngology consultation and the anesthesia response times were similar between the existing and novel system simulations. In the 2 simulations of the existing system where bronchoscopic equipment was brought to the ED by the otolaryngology resident, time to equipment arrival was 10 minutes or more. By design, specialized airway equipment was available for all novel system simulations and arrived an average of 3.9 minutes after being requested.

Death of the simulated patient occurred in the ED in 2 of 6 simulations of the existing system; specialized airway equipment was available for neither. For the novel system simulations, specialized airway equipment was available for all 6, and there were no simulated patient deaths.

There were 13 unique latent safety threats identified during the 12 simulations, 8 in the existing system simulations. Examples of latent safety threats included the lack of a tracheotomy tray in its designated location in the resuscitation bay, the lack of a needle cricothyroidotomy kit in the ED, and the lack of specialized airway equipment readily available to consultants. There were 11 unique knowledge deficits identified during the 12 simulations, 6 in the existing system simulations. An example of knowledge deficit was confusion surrounding the method for activating the appropriate paging systems. All latent safety threats and knowledge deficits identified during the existing system simulations were addressed either directly or by implementation of the novel system. Latent safety threats and knowledge deficits identified during novel system simulations were addressed either directly or by modifications of the novel system.

In a simulation-based study of pediatric critical airway obstruction in an ED, we found the use of in situ simulation to be highly effective to assess an existing system of care and to design and implement a novel system. Simulation participants believed that our institution was better prepared because of the novel system, and this was supported by the objective simulation data. Compared with simulations of the existing system, the novel system achieved faster response times for essential subspecialists and equipment and was associated with elimination of simulated patient deaths.

While avoiding issues of patient safety and confidentiality, and many of the costs and heterogeneity associated with clinical research, in situ simulation of critical airway obstruction allowed consistent reproduction of a high-acuity, low-frequency event and detailed, systematic analysis of the 2 systems of care. This use of medical simulation is analogous to animal models for analyses of complex physiologic processes, where interventions can be analyzed in a controlled environment and potential patient harm avoided.

In situ simulation has been used in a variety of settings to assess and improve systems of care. It appears to be particularly well suited for studies of high-acuity, low-frequency events, and has been used to assess and implement protocols for the management of postpartum hemorrhage,¹⁹ shoulder dystocia,¹⁰ ED procedural sedation,¹¹ and dental clinic procedural emergencies.²⁰ To our knowledge, this is the first study to use in situ simulation to assess an existing system of care for patients with critical airway obstruction and to design and implement a novel system. We also report a novel ED-based management algorithm for these patients (eFigure 3). The novel system of care was implemented for real patients in March 2012, with several successful activations at the time of this submission. We believe that the use of in situ simulation to assess and improve complex systems of care will be an integral part of medical process improvement work in the future.

Planned studies include (1) a more detailed description of the multidisciplinary development of the airway foreign-body simulation scenario (including survey data regarding the simulation scenario realism and educational benefits); (2) use of in situ simulation to implement the novel system of care in our institution's intensive care units and evaluation in this environment in a similar fashion; and (3) collection of prospective and historical control data for real patients who qualify for activation of the novel system. An area of future research should be further validation of this technique for process improvement in additional high-acuity, low-frequency events and in other clinical areas.

Our study has 2 obvious limitations: it involves simulated rather than real patients, and it includes a relatively small number of simulations. The use of simulations limits the validity of transferring these findings to real patients. However, studying the 2 systems of care for real patients would have been complicated by the heterogeneity of patient characteristics and presentations, and the time required to enroll a sufficient number of similar patients. In the absence of relevant prior studies with applicable critical airway process improvement data, we were unable to perform a meaningful power analysis a priori. This study was designed to provide data that will be useful for power calculations in subsequent studies of these processes. The logistics of planning 12 in situ simulations that monopolize the resources and personnel of the ED resuscitation bay and an OR was also a significant limiting factor in the number of simulations performed.

A third limitation is that members of each health care team could not be blinded to which system of care was being tested in a given simulation. It is possible that some of the improvement in response time was a version of the Hawthorne effect; ie, the health care team and otolaryngology resident performed better because they knew that the novel system was being studied and was supposed to be “better.” This effect was minimized by using (1) a primary outcome measure that was relatively independent of health care team attitudes toward the system of care (ie, participants were just responding to the resuscitation bay to generate that time point without any specific notification of which system of care they were testing or that a study was being conducted) and (2) a uniform simulation scenario and consistent manikin responses to health care team interventions.

A fourth limitation is the potential for health care team performance during each set of simulations to improve over time. The effect of this potential bias was minimized by the high variability of team members in each simulation, including having no repeat otolaryngology residents by design. Although repeated participation by anesthesiologists and ED health care providers was to some degree unavoidable, post hoc analysis of our data suggested no trends toward improved team performance over time within each set of simulations.

A fifth limitation is that data collection was performed by 3 of us (K.J., J.O., and B.K.), none of whom was blind to the study objectives. This bias was minimized by using a standard data collection form and consensus definitions for important data elements.

A final limitation is the generalizability of our findings. Given the likely strong effects of the personnel and health care environment specific to our large, tertiary care ED, as well as the use of simulated patients, the reported response times are likely to be different if a similar system were to be implemented in a different ED. The use of in situ simulation to assess an existing system and implement a new one is the most generalizable aspect of our study.

In conclusion, high-fidelity simulation was an effective method to implement a novel system of care for pediatric critical airway obstruction. The novel system was associated with more rapid response times and elimination of simulated patient deaths.

Correspondence: Kaalan Johnson, MD, Division of Pediatric Otolaryngology–Head and Neck Surgery, Cincinnati Children's Hospital Medical Center, MLC 2018, 3333 Burnet Ave, Cincinnati, OH 45229 (kaalan.johnson@cchmc.org).

Submitted for Publication: April 23, 2012; final revision received July 3, 2012; accepted August 12, 2012.

Author Contributions: Drs Johnson, Geis, and Kerrey and Ms Oehler had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Johnson, Geis, Meinzen-Derr, Myer, and Kerrey. Acquisition of data: Johnson, Oehler, Bauer, and Kerrey. Analysis and interpretation of data: Johnson, Geis, Meinzen-Derr, and Kerrey. Drafting of the manuscript: Johnson, Geis, Meinzen-Derr, Bauer, and Kerrey. Critical revision of the manuscript for important intellectual content: Johnson, Geis, Oehler, Meinzen-Derr, Myer, and Kerrey. Statistical analysis: Meinzen-Derr. Obtained funding: Geis. Administrative, technical, and material support: Johnson, Geis, Oehler, and Bauer. Study supervision: Geis and Kerrey.

Previous Presentation: This article was presented at the American Society of Pediatric Otolaryngology 2012 Annual Meeting; April 21, 2012; San Diego, California.