Abstract
BACKGROUND: Respiratory therapists (RTs) have historically performed safe and effective intubations, yet there are limited multi-center data assessing their intubation performance. Multi-center data can be used to compare RT intubation performance to that of other professions and identify quality improvement opportunities at hospitals where RTs perform intubation. We aimed to explore the feasibility of a multi-center collaborative to evaluate RT intubation outcomes.
METHODS: A data collection tool was developed by the authors and implemented at two institutions. Following institutional review board approval at each center and completion of data-use sharing agreements, data were collected between May 25, 2020–April 30, 2022, and combined for analysis. Descriptive statistics were used to compare overall success rate, first-attempt success rate, adverse events (AEs), and type of laryngoscopy.
RESULTS: There were a total of 689 intubation courses where RTs made an attempt, 363 from center A and 326 from center B. Center A captured 85% of all RT intubation courses, and center B captured 63%. Overall, RTs were successful in 98% of attempts. RTs made 86% of initial attempts. The most common indications for intubation were cardiac arrest (42%) and respiratory failure (31%). Videolaryngoscopy was used during 65% of initial attempts and was associated with higher first-attempt success rate, higher overall success rate, and fewer AEs. Airway-related adverse event rate was 8.7%; physiologic AE rate was 16%, and desaturation rate was 11%.
CONCLUSIONS: A collaborative examining RTs intubation performance was successfully initiated at 2 separate facilities. Intubations performed by RTs had a high success rate, with AE rates comparable to published results from other types of providers.
Introduction
Endotracheal intubation outside the operating room is a common procedure performed to facilitate invasive mechanical ventilation, to protect a patient’s airway, for stabilization prior to intra- or inter-hospital transport, or as a component of cardiopulmonary resuscitation. Respiratory therapists (RTs) frequently perform intubation, although practice varies widely throughout the United States.1,2 Early studies published in the 1980s and 1990s showed that RTs can perform safe and effective intubations.3,-,7 These early studies carry significant limitations as they were generally small single-center studies with differing outcome measures, subject populations, and training methods.
Contemporary clinical data on RT intubation outcomes are limited. A multi-center study in pediatric ICUs published in 2020 found that RT-performed laryngoscopies were associated with an increased risk of adverse events (AEs) although RTs had similar success rates to those of other providers in the center where RTs intubated most often.8 Importantly, that study showed that RTs rarely performed intubations in pediatric ICUs and that the use of videolaryngoscopy by RTs was protective against AEs.8
Prior work has indicated that RT training on endotracheal intubation varies considerably among centers.1,9 For instance, less than half of respiratory care departments provided education on videolaryngoscopy according to a survey published in 2017,9 and that figure was still less than two-thirds in a follow-up survey from 2020.1 Given the limitations of the current evidence and the lack of large studies describing intubations performed by RTs since the widespread availability of videolaryngoscopy, there is a need for multi-center studies characterizing RT intubation practices. In this study, we aimed to describe the development of a multi-center RT intubation collaborative and report early outcomes of RT-led intubations from 2 participating centers.
QUICK LOOK
Current knowledge
Endotracheal intubation is a low-frequency, high-risk procedure performed outside the operating room. Respiratory therapists (RTs) performing intubation is an established practice; however, the current data are outdated and consist of single-center case series of varying quality. These data also predate the widespread use of videolaryngoscopy.
What this paper contributes to our knowledge
This paper demonstrates the feasibility of a multi-center observational study of RT intubation practice. Early outcomes from the project indicate excellent success rates and acceptable adverse event (AE) rates. Important differences in outcomes and subject characteristics were noted between the 2 centers. Videolaryngoscopy was associated with higher first-attempt success rate, overall success rate, and fewer AEs.
Methods
We established a multi-center collaborative between two centers that included a series of virtual meetings and the development of a standardized data collection form adapted from NEAR4KIDS.10 Each center received approval from their respective institutional review board (IRB), and the study was considered exempt by both centers. A data user agreement was completed to enable data sharing between the two centers. We included intubations that occurred between May 25, 2020–April 30, 2022. Initially, intubation data were noted on paper data collection forms available at the bedside, completed as soon as feasible following the procedure; data were then subsequently entered into a secure REDCap database. Due to inconsistent completion of forms observed at an early evaluation cycle, concerns about lost forms, and increased resource utilization relative to redundant data entry, an all-electronic form was developed using Microsoft Forms (Microsoft, Redmond, Washington). The electronic form could be accesses via e-mail link or quick response (QR) code to enable its completion using a smartphone. Individual centers were allowed to customize data collection elements based on their IRB agreement; however, no protected health information was shared. Common data elements are summarized in Supplemental Table A (see related supplementary materials at http://www.rcjournal.com). For the purposes of this study, an intubation attempt was defined as placement of the laryngoscope within the oropharynx, whereas an intubation course was the complete attempt at placing an endotracheal tube.
RTs at center A perform intubation in all patient populations across the entire age spectrum and are the primary intubators during cardiac arrests outside the ICU. Center A’s training program includes in-person, 4-h classroom training that covered indications, anatomy, physiology, positioning, direct and videolaryngoscopy, and simulation training. Videolaryngoscopy included both traditional (curved or straight) blades and hyper-angulated blades, as indicated. After completing the training, each RT was required to perform 5 supervised intubations within one year to establish competency. Supervision was conducted by a competent RT or another credentialed intubation provider.
RTs at center B perform intubation in adult patients and are the primary intubators during cardiac arrests and emergent intubations in all adult patient care areas. Center B’s training program consists of a self-paced learning packet, classroom training, and an operating room rotation during which RTs are required to perform 10 endotracheal intubations under the supervision of anesthesia providers.
At center A, a flyer with the data capture QR code was conspicuously placed within each ICU to remind RTs to complete the data collection form. An electronic report was generated using electronic medical record data that identified all intubations done by RTs in the prior 2 weeks. If intubations in the report were not recorded within the study database, the RT who performed the intubation was contacted electronically with a copy of the study link with a prompt for the data to be entered. Reminders were sent weekly, as needed. The RT who performed the intubation was encouraged to fill out the database contemporaneously; however, any RT present at the time of intubation was also permitted to fill out the form. Preliminary data were analyzed in cycles of 50 intubations to assess progress, motivate participants, and report trends in outcomes.
At center B, RTs who attempted intubation were expected to complete the data collection form contemporaneously on a computer or work telephone. Verbal and electronic communications that included a link to the data collection form were used as reminders to the entire airway team and to specific members if identified as having performed endotracheal intubation without completion of the form.
Providers present for each intubation course were classified as anesthesia provider (attending physician or nurse anesthetist); emergency department or ICU attending physician; and resident, fellow, or advanced practice provider. More than one reason for intubation could be reported; however, intubations were later reclassified according to the primary indication. Indications were ranked using the following hierarchy, with higher order taking priority: cardiac arrest, respiratory failure, absent airway reflexes, unstable hemodynamics, airway clearance, procedure, and other. For the purposes of this study, ventilation and oxygenation failure were combined and classified as respiratory failure for analysis. The designation other was selected only if no other reason for intubation was recorded. The type of videolaryngoscope used was noted, but the specific blade type (traditional vs hyper-angulated) was not, and all videolaryngoscopy use was combined for analysis. Medications used to facilitate intubation were classified as none, analgosedation only, and analgosedation with neuromuscular blockade. Physical limitations recorded included limited mouth opening, presence of cervical collar, or limited neck mobility. The laryngeal view was classified using the Cormack-Lehane system.11 Oxygen desaturation was included only if initial SpO2 was > 90% and the subject had an SpO2 < 80% observed during the procedure. Cricoid manipulation before or during intubation attempt was recorded.
AEs were classified into airway and physiological. Airway-related AEs included esophageal intubation (immediate recognition), esophageal intubation (delayed recognition), main-stem intubation, direct airway injury, epistaxis, dental or lip trauma, vomit with aspiration, vomit without aspiration, and none. Physiologic AEs included cardiac arrest without return of spontaneous circulation, cardiac arrest with return of spontaneous circulation, hypotension requiring intervention, hypertension requiring intervention, medication error, dysrhythmia, pain or agitation, additional medications with delay in intubation, other, and none. Airway and physiologic AEs were combined for analysis.
We performed descriptive statistical analysis of the data set with SPSS v25 (IBM, Armonk, New York) using chi-square test for categorical variables and Mann-Whitney rank sum test for continuous variables. Primary analysis was performed for intubations courses where an RT made the initial attempt. Planned post hoc subgroup analyses were performed for cardiac arrest as primary indication for intubation and for intubations performed using videolaryngoscopy. Statistical significance was set at alpha < 0.05.
Results
There were 689 total intubation courses with at least one attempt by RTs, 363 at center A and 326 at center B. Center A captured 85% of courses with an RT intubation attempt, and center B captured 63%. The overall intubation success rate was 98% and did not differ between centers. RTs made the initial attempt in 86% of all intubation courses, and RTs at center B were more likely to make the initial attempts (91% vs 81%, P < .001). The most common indications for intubation were cardiac arrest (42%), respiratory failure (31%), and unstable hemodynamics (6.8%), with a distribution that differed between centers (P < .001). Videolaryngoscopy was used in 64% of initial attempts, and center B was more likely to utilize it (68% vs 60%, P = .03). Data for all intubation attempts are summarized in Table 1.
All Attempts
There were 589 intubation courses where RTs performed the initial laryngoscopy, with an overall success rate of 98% and first-attempt success rate of 90%, with no differences between centers (Table 2). Airway-related AEs occurred in 5.6% of courses, and center A had a higher AE rate (7.5% vs 3.7%, P = .048). Physiologic AEs were 14%, and center A had a higher AE rate (18.0% vs 9.5%, P = .002). Oxygen desaturation occurred in 7% of courses and was higher at center A (12.0% vs 3.6%, P = .008). The most common reasons for intubation when an RT made the initial attempt were cardiac arrest (40%), respiratory failure (32%), and unstable hemodynamics (6.3%). The indication for intubation differed between centers, with center A being more likely to intubate due to cardiac arrest (46% vs 34%), whereas intubation to facilitate airway clearance was more common at center B (0.7% vs 7.5%). Pre-oxygenation was provided using a bag-valve-mask (69%), non–rebreather or partial rebreather mask (12%), noninvasive ventilation (10%), high-flow nasal cannula (4.4%), and nasal cannula (4.1%). Videolaryngoscopy was used in 66% of initial attempts and was less common in center A (62% vs 71%, P = .003). Medications given to facilitate intubation included none (46%), analgosedation and neuromuscular blockade (37%), and analgosedation only (17%). Center A was more likely to intubate with no medications (54% vs 39%) or analgosedation only (24% vs 11%) and less likely to use analgosedation and a neuromuscular blocker (23% vs 50%, P < .001 for the distribution). Cricoid manipulation was used in 25% of all courses and was less common in center A (21% vs 30%, P = .01). Complete results for intubations where RTs performed the initial laryngoscopy are summarized in Table 2.
Respiratory Therapists Made Initial Attempt
There were 100 intubation courses where an RT was the second provider to attempt intubation, 69 (69.0%) at center A and 31 (31.0%) at center B. First-attempt success rate was 91%; overall success was 99% and did not differ between centers. Center B was more likely to have anesthesia providers present (23% vs 5.8%), whereas an attending physician was more likely to be present at center A (52% vs 39%, P = .042). The primary indication for intubation differed between centers, with cardiac arrest being more common at center B (71% vs 46%, P = .030). Medications used to facilitate intubation also differed, with center B being more likely to use no medications (74% vs 46%, P = .02). Center B reported more grade 3 and grade 4 laryngeal views (26% vs 11%, P = .039). There were no differences between centers for RT-attempted laryngoscopy, pre-oxygenation device, respiratory support prior to intubation, first-attempt laryngoscopy method, cricoid manipulation during or prior to attempts, physical limitations, airway-related AEs, and physiologic AEs. Results are summarized in Table 3.
Respiratory Therapists Did Not Make the Initial Attempt
Post Hoc Subgroup Analyses
Cardiac Arrest
Cardiac arrest was the primary indication in 234 intubation courses (Table 4). There were no differences between cardiac arrest and other indications for overall RT success rate, first-attempt success rate, physical limitations, airway AEs, and esophageal intubation. Providers present, pre-oxygenation device, respiratory prior to intubation, laryngoscopy method, cricoid manipulation during or prior to attempt, medications used, laryngeal view, and physiologic AEs were all different between groups. Cardiac arrest was associated with a lower likelihood of anesthesia or attending provider present (69% vs 80%, P < .001), more use of bag-valve-mask for pre-oxygenation (95% vs 52%, P < .001), less videolaryngoscopy use (54% vs 79%, P < .001), less medication administration (14% vs 86%, P < .001), and more physiologic AEs (26.0% vs 5.9%, P < .001). Results are summarized in Table 4 and Supplemental Table B (see related supplementary materials at http://www.rcjournal.com).
Cardiac Arrest and Videolaryngoscopy
Videolaryngoscopy
Due to the significantly lower use of videolaryngoscopy in cardiac arrest compared to intubations for other causes (54% vs 79%, P < .001), we chose to compare its use during intubation courses in non–cardiac arrest subjects. Videolaryngoscopy was associated with higher overall success rate (99% vs 95%, P = .040), first-attempt success rate (94% vs 82%, P = .001), improved laryngeal view (grade 1 view obtained in 75% vs 54% of attempts, P = .002), fewer airway-related AEs (2.5% vs 11.0%, P = .002), and fewer esophageal intubation (1.1% vs 11.0%, P < .001) compared to direct laryngoscopy. There were also significant differences in the pre-oxygenation device, medications used to facilitate intubation, and respiratory support prior to intubation. There were no differences for indication for intubation, cricoid manipulation, physical limitations, physiologic AEs, or desaturation. Results are summarized in Table 4 and Supplemental Table C (see related supplementary materials at http://www.rcjournal.com).
Discussion
We found a multi-center collaborative describing RT intubation practice was feasible, although there was variation in the percentage of intubation attempts captured by each center. The most common indication for intubation was cardiac arrest; videolaryngoscopy was used more frequently than direct; first-attempt success rate was excellent, and airway-related AEs were rare. RTs also demonstrated high success rates after a prior unsuccessful attempt by another provider. Despite different subject characteristics and indications for intubation, there was no difference in first-attempt success rate between centers. There were differences in AEs, with center B reporting lower rates of airway-related AEs, physiologic AEs, and rate of desaturation. These differences were likely related to the indication for intubation (nearly half of the intubation attempts at center A were for cardiac arrest, whereas a higher proportion of intubations at center B were for airway clearance), differences in training, or differences in type of laryngoscopy used between centers.
Intubation is a low-frequency, high-risk procedure with potential for significant harm; so it is imperative that programs routinely track and benchmark success rates, indications, and AEs. We were able to demonstrate the conduct of a multi-center collaborative was feasible and capable of characterizing intubation procedures across 2 institutions; such an effort could be easily scaled for the inclusion of additional centers. Current data indicate a limited number of RT departments with intubation programs have quality assurance programs in place.1 We believe an intubation registry to support intubation quality improvement should be a focus of all departments where RTs perform intubation. Importantly, our data collection tool can be used by other departments even if they are not part of our study collaboration. Intubation data can be used to assess a program’s internal performance and provide hospital leadership with valuable patient safety and outcomes data. The opportunity to perform intubation as part of a well-established RT intubation program may also be seen as a potential recruitment tool for departments aiming to attract staff wishing to practice at the top of their license.
Intubation performed by RTs is a long-standing practice; however, there are limited contemporary data, with most prior studies having been published prior to the year 2000.3,-,7 The most recent data from a large multi-center database of pediatric ICUs found that RT success rates were similar to those of other providers but were associated with a higher rate of AEs.8 That study also found videolaryngoscopy use by RTs was protective against AEs, but the findings were limited by the fact that RTs performed < 1% of total attempts.8
Our current study found RTs used videolaryngoscopy in 66% of initial attempts. Univariate analysis revealed higher first-attempt success rate and overall success rate with lower rates of AEs, better laryngeal view, and fewer esophageal intubations. Despite videolaryngoscopy being associated with a higher first-attempt success rate, there were significant differences in providers present, pre-oxygenation device, respiratory support prior to intubation, and medications used to facilitate intubation. Neuromuscular blockade has been noted to be associated with higher first-attempt success rate and may have affected the results.12 The use of videolaryngoscopy outside the operating room has been shown to increase first-attempt success rate, improve laryngeal view, and reduce the likelihood of difficult intubation,13 although not every meta-analysis has found a consistent benefit for videolaryngoscopy.14 This recent meta-analysis reported first-attempt success rate of 86% for all studies, whereas our results indicate RT success rate was higher at 94%. This could be related to the well-established RT practice at both centers, selection bias, or incomplete reporting of all intubation attempts. Importantly, our study found videolaryngoscopy was associated with a significantly better laryngeal view, consistent with prior meta-analyses on videolaryngoscopy.14,15
In our current study, cardiac arrest was the most common indication for intubation, and RTs were less likely to use videolaryngoscopy during cardiac arrest. At both centers, airway-trained RTs respond to all emergencies along with the charge RT. During cardiac arrests at both centers, a physician or another provider is present for all RT intubation attempts, and the decision to intubate is undertaken in collaboration with the RT and provider. The lack of portability for most videolaryngoscopy equipment likely limited its use in emergent situations, although there was no difference in overall and first-attempt success rate in cardiac arrests. RTs intubating during cardiac arrests may also be more experienced and preferred to use direct laryngoscopy, although we were unable to capture the experience level of the RT making each attempt or the rationale for why each method was chosen. Our results compare favorably to recent reports of physicians intubating during cardiac arrest.16,17
Our study has several limitations, primarily that all data were self-reported. Despite our best efforts, we were unable to capture every intubation attempt at each center. Because we were unable to capture all intubation attempts, our results may overestimate success rates and underestimate AE rates as it is possible RTs with lower skill levels or low success rates did not report their attempts; the opposite could also be true, as it is possible that documentation on straightforward/uncomplicated intubations could have been missed, whereas complex intubations or those with complications could have been overrepresented. Both centers in our study are large academic medical centers with well-established RT intubation programs; thus, the results may not be generalizable to other centers. Whereas not specifically tracked, we observed participants required regular reminders and prompts to fill in the database. Future studies should work to include community hospitals. Center B only intubated adults, whereas center A intubated all age groups. Due to the significant difference in videolaryngoscopy use in cardiac arrest and procedural peculiarities involved in this most emergent indication, we chose to compare videolaryngoscopy intubation attempts in non–cardiac arrest subjects only. We planned to perform logistic regression analysis to control for differences between centers; however, the small number of AEs and high first-attempt success rate precluded us from including all relevant variables without risk of model overfitting. We did not document RT experience level in our database, which is likely an important factor in intubation success. We do not have data on uncaptured intubation attempts; it is possible those could have affected our findings if RTs may have been reticent to report unsuccessful attempts. Lastly, we did not utilize medical records for data on AEs due to resource limitation and inconsistent documentation of transient events, such as desaturation or hypotension, that may not be captured in the electronic medical record.
Conclusions
A multi-center observational study of RT intubation practice was feasible. Preliminary outcomes indicate excellent safety outcomes and first-attempt success rates similar to those of physicians. Videolaryngoscopy was associated with higher first attempt and overall RT success rates, improved laryngeal view, and reduced AEs, including esophageal intubation.
Footnotes
- Correspondence: Andrew G Miller MSc RRT RRT-ACCS RRT-NPS FAARC. E-mail: Andrew.g.miller{at}duke.edu
See the Related Editorial on Page 1189
Supplementary material related to this paper is available at http://www.rcjournal.com.
Mr Miller discloses relationships with Saxe Communications and S2N Health. Mr Miller is a section editor for Respiratory Care. Mr Emberger discloses relationships with Dräger Medical and Avanos. Dr Rotta discloses relationships with Breas United States, Vapotherm, and Elsevier. Mr Gillin has disclosed no conflicts of interest.
- Copyright © 2023 by Daedalus Enterprises
