Harnessing In Situ Simulation to Identify Human Errors and Latent Safety Threats in Adult Tracheostomy Care

Abstract

Background:

Tracheostomies are associated with high rates of complications and preventable harm. Safe tracheostomy management requires highly functioning teams and systems, but healthcare providers are poorly equipped with tracheostomy knowledge and resources. In situ simulation has been used as a quality improvement tool to audit multidisciplinary team emergency response in the actual clinical environment where care is delivered but has been underexplored for tracheostomy care.

Methods:

From July 2021 to May 2022, our team conducted in situ simulations of a tracheostomy emergency scenario at Montefiore Medical Center to identify human errors and latent safety threats (LSTs). Simulations included structured debriefs as well as audio-video recording that allowed for blind rating of these human errors and LSTs. Provider knowledge deficits were further characterized using pre-simulation quizzes.

Results:

Twelve human errors and fifteen LSTs were identified over 20 simulations with 88 participants overall. LSTs were divided into the following categories: communication, equipment, and infection control. Only 50% of teams successfully replaced the tracheostomy tube within the scenario’s five-minute time limit. In addition, knowledge gaps were highly prevalent, with a median pre-simulation quiz score of 46% (interquartile ratio 36 to 64) among participants.

Conclusions:

Our in situ simulation-based quality improvement approach shed light on human errors and LSTs associated with tracheostomy care across multiple settings in our health system. This method of engaging frontline healthcare provider key stakeholders will inform the development, adaptation, and implementation of interventions.

Tracheostomy, one of the most common procedures performed on critically ill patients, is associated with high rates of complications^1–4 and mortality.^5–7 Reliable teams and systems are required to effectively provide routine and emergency management for patients with tracheostomies to prevent catastrophic outcomes such as hypoxic brain injury. However, most providers who routinely manage these patients are poorly equipped to provide tracheostomy care, having little training and insufficient knowledge.^6–12 Although these tracheostomy complications are largely preventable, this lack of experience and knowledge regarding tracheostomies continues to contribute to patient harm. Thus, it is necessary to address these contributing factors before they lead to patient consequences. Previous studies have demonstrated improvement in providers’ knowledge and skills in managing tracheostomy-related complications with tracheostomy education over the course of several months, including non-otolaryngologist residents, fellows, nurse practitioners, nurses, and medical students.^13–16 However, published interventions that focused on individual provider knowledge and skills failed to test team behaviors in response to a tracheostomy emergency or to interrogate the complex person-centered work systems in which tracheostomy care is delivered.

Deficiencies in patient care include both human errors and latent safety threats (LSTs). LSTs are the underlying systems factors that can combine to contribute to patient harm. These deficiencies can be identified through a variety of methods. In situ simulation, wherein medical teams operate in their actual clinical environment using tools and resources typically available to them, is a powerful tool that can recreate the complex systems in which care is delivered. It allows for mimicking of real life scenarios to facilitate time-sensitive measurement of the quality of patient care.^17–20 Coupled with systems-based debriefing,^21,22 in situ simulation can thereby facilitate identification of errors in team performance and LSTs before they contribute to patient harm in a more comprehensive program to improve quality of care, evoking a Safety-II approach to patient safety and quality improvement.²³ Our simulation and quality improvement implementation team has previously used in situ simulation to identify LSTs associated with pediatric tracheostomy care^24–26 and has established infrastructure to scale this approach to other contexts.^27,28 The objective of this study was to use in situ simulation as a diagnostic tool to provide an initial examination of tracheostomy-related human errors and LSTs in adult tracheostomy management.

METHODS

Study Protocol, Setting, and Development

This simulation-based observational study was approved by the Albert Einstein College of Medicine Institutional Review Board (2018–9421) and conducted at Montefiore Medical Center from July 2021 to May 2022. The study protocol was designed to identify human errors, LSTs, and knowledge gaps in adult tracheostomy care in each department. The study was conducted in five hospital units at a tertiary care, urban medical center: the medical and surgical ICUs, patient floors, and the transplant unit. Physician and nurse champions were recruited from each unit through our institution’s multidisciplinary airway committee, which includes physician, nursing, and respiratory therapy leaders; clinical educators; patient safety and quality improvement experts; and representatives from critical care, anesthesiology, otolaryngology, pulmonology, and emergency medicine. Unit champions coordinated in situ simulation activities with the study team through recruitment of multidisciplinary teams representative of typical ad hoc emergency teams that would respond to tracheostomy emergencies in their clinical settings (one to two physicians, two nurses, one to two physician assistants or nurse practitioners, a medical student if present on the unit, and a respiratory therapist if present on the unit). These unit champions followed up with the study team regarding simulation findings and unit-based action plans.

Unit champions allowed for buy-in from the units through scheduling and recruitment, which was vital to the implementation of the program. Scheduling focused on day and night shifts to provide units with flexibility. Unoccupied rooms to host the simulation were identified by the charge nurse a few hours prior. The study team built in flexibility for a 20% to 30% cancellation rate and confirmed the simulation with the unit champions one day before. No-go criteria (for which a scheduled simulation would be postponed) included issues with staffing, clinical load, an active code, or admissions.²⁹ The team also kept strict time to allow for coverage of the participants’ clinical duties.

In Situ Simulation

In situ simulations were performed on the hospital unit during the clinical work hours of the team. Experts in nursing, pulmonology, otolaryngology, critical care, patient safety, and healthcare simulation developed the simulation scenarios to mimic common tracheostomy emergencies observed in our institution that could occur at any time in the patient’s hospital course after tracheostomy tube placement. Prior to the simulation, providers consented to participation and audio-visual recording. The simulations were facilitated by a member of the study team. The five-minute scenario involved an adult patient with a mature tracheostomy tube (placed one month earlier) that was either (a) partially dislodged or (b) obstructed due a mucous plug (see Appendix 1). These scenarios were chosen based on the large number of observed negative patient outcomes observed at our hospital system and were adapted from our group’s pediatric tracheostomy simulation/implementation science program. When feasible based on clinical load and staffing, two teams were recruited during the same shift and simulations conducted consecutively, with the more straightforward scenario of obstructed inner cannula performed first to reinforce basic skills. The simulation participants were not given information regarding the purpose of the simulation, and team leaders and participant roles were not assigned by the study team.

Debriefs and Latent Safety Threats

Immediately following the simulation, participants engaged in 10-minute system-focused debriefs led by the study team to identify LSTs and human errors. Facilitators, who included an otolaryngologist or critical care attending on the study team, identified errors in critical actions during the scenario. The attending on the study team demonstrated correct tracheostomy tube change technique and provided opportunity for practice, while the unit champions demonstrated the location of key airway equipment on the unit as well as any other unit specific information. LSTs were captured via the audio-visual recording of the simulation, each of which was reviewed by two blinded raters at a later time. A checklist of behaviors and actions that should be present in an optimally conducted scenario was used as the scoring guideline, represented in Tables 1 and 2. The two raters reviewed each recording separately, and discrepancies in scores were addressed by taking the mean.

Table 1.

Errors^* in Tracheostomy Emergency Team Response Observed Through In Situ Simulation

Category	Total Frequency (Human Errors / Total Simulations)
Continued bagging of obstructed tube	12/20
Type of tube not verbalized	12/20
Tracheostomy malposition not addressed through repositioning or replacing tube	9/20
Concern for tube obstruction not vocalized	7/20
Inner cannula not removed, inspected, and replaced	5/20
Ventilation not maintained throughout scenario	5/20
Failed to assess patency through auscultation	2/20
Patient intubated orally (without attempting to address tracheostomy tube issue)	1/20
Team not aware of timing of fresh versus mature tracheostomy	1/20
Tracheostomy ties not properly tightened	1/20
Uncuffed tracheostomies brought to bedside instead of cuffed	1/20
Wrong size tracheostomy brought to bedside	1/20

^*Human errors, in decreasing order of frequency.

Table 2.

Latent Safety Threats* in Tracheostomy Emergency Team Response Observed Through In Situ Simulation

Category	Total Frequency (Latent Safety Threats / Total Simulations)
Communication
Team leader not identified	11/20
Lack of designated runner	10/20
Did not call for help or activate rapid response/code team	6/20
Closed loop communication not used when appropriate	5/20
Call out strategy not used when appropriate	4/20
Briefs or huddles not implemented when appropriate	2/20
Equipment
Smaller tracheostomy not placed physically at bedside	20/20
Correct tracheostomy not available on unit	13/20
Down-size tracheostomy not available on unit	13/20
Same size tracheostomy not placed physically at bedside	11/20
Team not aware of spare tracheostomy tube location	3/20
Uncuffed tracheostomies brought to bedside instead of cuffed	1/20
Wrong size tracheostomy brought to bedside	1/20
Infection Control
Awareness of clean and sterile technique, PPE, not observed	2/20
Potential aerosol spread of contaminants observed	1/20

Latent safety threats (LSTs) by domain, in decreasing order of frequency.

PPE, personal protective equipment.

Knowledge Quiz

In addition to the simulation, baseline provider knowledge on tracheostomy care was further assessed through an online knowledge quiz administered through hospital intranet-based Talent Management educational software. This knowledge quiz was developed by otolaryngology-head and neck surgery tracheostomy experts based on best evidence,^30–32 guidelines and expert consensus in conjunction with an e-module and tracheostomy curriculum for healthcare providers. Providers were tested on best practices in managing tracheostomies (Appendix 2). The quiz questions were critically reviewed by airway committee leadership including nursing and respiratory therapy. The list of potential participants for the planned simulation were emailed to the study team by the unit champion in advance. The participants were then assigned the quiz via Talent Management. Participants were requested to take the quiz before the start of the simulation; however, participating in the quiz was voluntary and participants were able to participate in the simulation without quiz completion. Quiz responses were categorized by provider type (nurses, physician assistants, and critical care fellows) and a Kruskal-Wallis test was conducted to measure statistical significance.

RESULTS

Simulations

Out of 22 planned simulations, 20 were conducted. The transplant unit completed five simulations, the MICU completed four simulations, the patient floors completed six simulations, the SICU completed four simulations, and the NSICU completed one simulation. Eighty-eight unique clinicians (21 physicians [including attendings, fellows, and residents], 25 physician assistants and nurse practitioners, 37 registered nurses, 3 respiratory therapists, and 2 medical students), grouped into multidisciplinary teams typical of those that would typically respond to a tracheostomy emergency on the unit, participated in the 20 in situ simulations.

Human Errors and Latent Safety Threats

Twelve types of human errors and fifteen different LSTs were observed across the 20 conducted simulations (Tables 1 and 2). LSTs were categorized into domains, including communication, equipment, and infection control. Among the 20 scenarios run, 13 involved an occluded tracheostomy tube and seven involved a dislodged tube. The scenario was concluded when ventilation was restored through a correctly positioned, patent tracheostomy tube or at a maximum of five minutes. Eight of 13 (61.5%) teams successfully completed the occluded tracheostomy tube scenario in less than five minutes. Two of seven (28.6%) teams successfully completed the dislodged tracheostomy tube scenario within five minutes. Overall, 10 out of 20 (50%) teams completed their respective scenarios within five minutes.

Knowledge Gaps

Knowledge and skills gaps identified through simulation were corroborated in quiz findings. With 37 responses, the median score overall was 46%. Quiz data was further analyzed through breakdown by provider role (Table 3). Responses included 25 RNs, seven PAs, and five critical care fellows. The critical care fellows tested with the greatest quiz scores on average. Kruskal-Wallis analysis results indicated a statistically significant difference among provider roles, with a p value of 0.016, determined with an alpha score of 0.05. While other providers, including additional medical doctors and respiratory therapists, participated in the simulation and were assigned the quiz, no responses were recorded from these groups in the quiz.

Table 3.

Pre-Simulation Tracheostomy Quiz Scores by Provider Role (N = 37)

Provider Role	Quiz Score Median (Q1, Q3)	p value
Registered nurses (n = 25)	46 (36, 55)	0.016
Physician assistants (n = 7)	45 (36, 45)
Critical care fellows (n = 5)	64 (59.5, 73)
All participants (n = 37)	46 (36, 64)

Q1, first quartile (25th percentile); Q3, third quartile (75^th percentile).

DISCUSSION

This study is a novel application of in situ simulation to identify LSTs associated with inpatient adult tracheostomy emergency management across multiple clinical settings. Through implementing a proactive manner to recreate complex systems of care in hospital settings with in situ simulation, we have demonstrated the importance of not solely relying on root cause analysis of clinical adverse events, but also addressing the systemic causes. The prevalence of human errors and LSTs identified through simulation imply an emerging priority to promote systems-level interventions. Additionally, in situ simulation engages key stakeholders at the unit and health system levels in identifying strengths and gaps in tracheostomy care delivery, particularly through formal and informal debriefings.

Human errors leading to severe patient safety hazards and deficits in emergency readiness were exposed in the simulations, given half the teams participating were unable to successfully re-establish the airway within the allotted five minutes. The commonality of human errors suggests that yearly training and hands-on skills practice in conjunction with protocol safeguards are essential to improving tracheostomy care. Additionally, the prevalence of equipment LSTs indicate a necessity for unit equipment standardization, including adequate supply of multiple size tracheostomy tubes, to prepare for tracheostomy patients on unit. Low quiz scores among clinical providers expected to routinely care for patients with tracheostomies corroborated the presence of an eminent knowledge gap. While scores among provider groups varied, all three groups scored medians below 65%. Thus, it is advisable that future studies exploring interventions for quality improvement in tracheostomy care substantiate results with knowledge assessments as well, particularly for providers who may have received less tracheostomy-related education.

Certain deficiencies in tracheostomy care observed in the simulations were considered both human errors and LSTs in our scoring guideline. Whether a participant brought the correct type of tracheostomy tube to the bedside could imply a knowledge deficit or a lack of equipment available. Occasionally these deficiencies were explained in the debriefs; however, scoring of the simulation does not clarify the reasoning. Thus, human errors and LSTs are intertwined and must be addressed in conjunction to minimize gaps in tracheostomy care.

Limitations

This study has multiple limitations. Although we found in situ simulation to be practical and helpful in revealing system issues to frontline key stakeholders, unit leadership, and hospital administration independent of patient adverse events, the small sample size and low participation rates of this exploratory pilot program limit its generalizability. In addition, this study focused on several units within a single hospital system. To further explore the validity of these results, the study would need to be expanded to many different types of clinical settings in healthcare systems beyond the one studied here. Furthermore, potential selection bias may have been present during this study as the voluntary nature of participation might select for those more confident in tracheostomy skills or with more experience in simulation. We attempted to mitigate this limitation through the unit champion structure and scheduling of simulations during healthcare providers’ usual shifts. Lastly, as a baseline study, an educational intervention as a means of measuring improvement in LSTs or provider knowledge over time was not conducted, thus limiting our knowledge regarding the effectiveness of in situ simulation as an improvement tool rather than an observational one.

CONCLUSION

The study provided a baseline understanding of tracheostomy care-related errors and threats. Additional research is necessary to further explore these objectives. Next steps for implementation by our study team include the wide-scale rollout of a tracheostomy curriculum (involving both an online module and task- and scenario-based training) piloted among critical care fellows and other physician, nurse, and respiratory therapist groups. In addition, our team is currently addressing the standardization of beside and unit equipment stocking through unit checklists, as equipment issues were shown to be a consistent problem during these simulations. Our research is also focused on piloting a crisis-checklist cognitive aid for tracheostomy emergencies to improve provider care. Members of the study team and unit champions will continue to be involved in these further studies to address the gaps identified. Finally, our study team is developing a phone application to document LSTs and human errors that arise during simulations in real time to bypass video review and create ease of convenience in assessing simulations. Ideally, with the use of this application, the success of quality improvement interventions can be monitored more closely with the documentation of tracheostomy care discrepancies over time.