COVID-19 Transmission to Health Care Personnel During Tracheostomy Under a Multidisciplinary Safety Protocol

We conducted the present study at Michigan Medicine, which provided tertiary and referral services during the initial pandemic surge. According to the protocol at our institution, tracheostomies were performed either at the bedside in a negative-pressure room or in the operating room, depending on the proceduralist’s judgment. The institutional protocol for performing tracheostomy during the COVID-19 pandemic was developed through multidepartmental stakeholder engagement that specified standards and practices for performing tracheostomy (Supplement 1, available online only at ajcconline.org). This review of provider safety under the protocol was determined by our institutional review board to be a quality assurance/quality improvement initiative (HUM00187214) and was not regulated as human subjects research.

All health care personnel involved in tracheostomy procedures on COVID-19–positive patients from April 9, 2020, through July 11, 2020, were invited to participate in a postprocedure survey (Supplement 2, available online only). Proceduralists, anesthesiologists, and surgical technologists were identified by retrospective medical record review of procedure notes. To identify other members of the tracheostomy team, we also included any respiratory care practitioner or nurse who was charting vitals or respiratory parameters at the time of tracheostomy. Because many individuals performed more than one tracheostomy and each tracheostomy involved multiple providers in the room, adherence to the use of personal protective equipment (PPE) and use of properly filtered rooms are reported as percentages of the total number of provider–tracheostomy encounters. Patient demographics, procedure type, COVID-19 status, and duration from intubation to tracheostomy were also obtained via the electronic medical record and stored in a password-protected REDCap (Research Electronic Data Capture) database (version 10.0).²⁶ 

A multidisciplinary team comprising frontline health care personnel involved in COVID-19 tracheostomy care, individuals well-versed in COVID-19 tracheostomy literature and survey design, and representatives from occupational health services iteratively developed a 22-question survey using Qualtrics software, version 2020 (Qualtrics, LLC). The survey was assessed for clarity and vetted against a quality checklist to eliminate design flaws, such as leading questions, dual questions, or inadequate response options.

The survey was distributed via email, and follow-up reminders were sent to nonrespondents at 1, 2, and 3 weeks. The survey included questions regarding the health care personnel’s use of PPE, development of symptoms, or receipt of a positive result on a COVID-19 test within 21 days of the procedure. Individuals were not mandated to complete the survey and were given the option to participate or not as part of the community reviewing the institutional protocol. COVID-19 can present with a variety of symptoms; therefore, we used the symptoms employees are screened for every day at our institution (eg, fever, cough, shortness of breath). The survey also included questions regarding familiarity with the protocol, perceived provider safety, breaches of protocol, and suggested areas for improvement. Participants were given the option of choosing to answer or not to answer questions pertaining to COVID-19. Survey responses were maintained on a password-protected folder behind the institutional firewall that could be accessed by the study team only. The study team had no other access to individual records.

The occupational health services department reviewed the list of all health care personnel present when a tracheostomy was performed in a COVID-19–positive patient to identify instances in which individuals were tested for COVID-19 in the 3-week interval after participation in a tracheostomy. The results were reported to the study team in aggregate only to maintain individual privacy. We compared responses to the survey regarding COVID-19 status and symptoms with the data from occupational health services to assess survey reliability.

All survey data were collected from December 12, 2020, through January 15, 2021. Descriptive statistics of counts, proportions, means, and SDs were used to analyze closed-ended survey items using Excel, version 2020 (Microsoft Corporation). Times to tracheostomy and number of procedures were reported as medians and interquartile ranges. Given the brevity of responses, open-ended survey items were reviewed and analyzed using inductive content analysis. Inductive content analysis is recommended when there is limited literature on the phenomenon under study.²⁷  Author T.C.S. generated the initial codes and categories. Author P.K.P. reviewed the responses and categories and provided additional input. Open-ended questions left blank or that received an “N/A” response were excluded from analysis. Differences in the ratio of respondents to nonrespondents between health care provider groups were compared by using χ² testing.

A total of 64 tracheostomies were performed on COVID-19–positive patients from April 9, 2020, through July 11, 2020. Among the cohort of 64 patients who had tested positive for COVID-19 by quantitative polymerase chain reaction (qPCR) at their initial presentation, the mean time from initial diagnosis to tracheostomy was 24 days (SD, 10 days) and the mean time from last positive test to tracheostomy was 19 days (SD, 11 days). Thirteen patients (20%) had a negative result on a follow-up test from tracheal aspirate or nasopharyngeal swab before tracheostomy. Of 146 health care personnel initially identified as being eligible for survey distribution, 7 were travelers or were not available for further contact, and 6 responded that they did not participate in a tracheostomy procedure, for a total of 133 unique individuals. Of the 133 individuals, 50 were physicians, 49 were nurses, 31 were respiratory care practitioners, and 3 were surgical technologists. Proceduralists included physicians from multiple disciplines, including otolaryngology, thoracic surgery, acute care surgery, neurocritical care, and interventional pulmonology. Surveys were completed by 80.5% of participants (Table 1). Characteristics and outcomes of the health care personnel who responded to our survey are presented in Table 2. No significant difference in occupation was found between respondents and nonrespondents when occupation was self-reported (ie, proceduralist, anesthesiologist, fellow/resident, nursing, respiratory care professional, surgical technologist); however, we did note differences in disciplines when physicians were considered as a single group (χ² = 8.4, P = .04).

Most providers (91%) had previous experience with tracheostomies before the procedure on a COVID-19–positive patient. A total of 325 provider–tracheostomy encounters were reported. Respondents reported using standard PPE (an N95 mask with eye protection, a powered air-purifying respirator [PAPR], or a negative-pressure tent [AerosolVE]) in 100% of the tracheostomies performed. About 33% of cases were completed with an N95 mask and eye protection; 67% were performed using a PAPR. Four respondents reported having used a negative-pressure tent in addition to an N95 mask with eye protection. Three provider–tracheostomy encounters were reported in a neutral pressure room; one respondent reported being “unsure” of what type of room was used. A total of 99% of provider–tracheostomy encounters were reported as being performed in either a negative-pressure room or a room with a high-efficiency particulate air (HEPA) filter.

Five respondents (5%) reported new symptoms (n = 3) or a positive COVID-19 test result (n = 2) within 21 days of performing a tracheostomy. In contrast, no cases were identified by cross-matching with institutional occupational health services data. Three of the 5 respondents reported other high-risk exposures occurring during the same time frame.

Two respondents reported a positive COVID-19 test result within 21 days of performing a tracheostomy procedure (Table 2), one being the primary proceduralist and the other a resident physician. One respondent reported participation in 3 tracheostomies, and the other reported performing 4 tracheostomies during this period. During each tracheostomy, these providers used standard PPE and performed the procedure in a negative-pressure setting, either at the bedside or in the operating room. Neither provider reported having any other high-risk exposures outside the tracheostomy; however, one noted in the comments section that they most likely contracted the virus at another health care facility or via community spread. These 2 respondents never worked together in performing a tracheostomy during the study period.

Three other respondents reported new symptoms within 21 days of participating in tracheostomy procedures, including one nurse, one respiratory care practitioner, and one proceduralist. One respondent reported a new cough, one had a baseline increase in their inhaler usage, and one reported “other” symptoms without specifying the symptoms. One of these respondents had performed more than 4 tracheostomies and the other 2 respondents reported having performed one tracheostomy. During each tracheostomy, these providers used standard PPE and performed the procedure in either a negative-pressure room or a room with a HEPA filter. Two of these individuals did not receive a COVID-19 test and one respondent received a negative result on a COVID-19 test within 21 days of the tracheostomy. Two respondents reported having other high-risk exposures outside the tracheostomy and definite causality could not be determined for any of the respondents.

For the 5 health care personnel reporting positive test results or symptoms, patient data regarding ongoing infectivity were assessed by examining follow-up COVID-19 testing. Four of these 5 survey respondents performed at least one COVID-19 tracheostomy in a patient for whom a negative result on a COVID-19 test was not available at the time the tracheostomy was performed.

Overall, most respondents were aware (63%) or somewhat aware (28%) of the institutional tracheostomy protocol. The Figure outlines the respondents’ feelings of safety; 92% of respondents reported feeling extremely safe or moderately safe at the time of the procedure. Seven respondents reported perceived breaches of protocol. These breaches included having open or unsecure airways (n=2), staff who were unfamiliar with PPE donning and doffing or the tracheostomy procedure itself (n = 2), and ancillary staff entering and exiting an operating room, which was perceived as potentially contaminating the sterile field (n=3). One of these reports was by a health care worker who also reported new-onset symptoms within 21 days of the tracheostomy procedure.

Table 3 outlines the 4 key categories and representative excerpts abstracted from 2 of the open-ended questions on the survey. Some (40%) of the open-ended questions were left blank or received a response of “N/A [not applicable].” Respondents reported adequate PPE and deliberate, closed-loop communication as key elements that enhanced their feelings of safety. Two opportunities for improvement included increasing PAPR availability for all providers and forming dedicated tracheostomy teams to perform all COVID-19 tracheostomies throughout the hospital.

Some of the most provocative data from this study relate to the 5 survey respondents who reported either positive test results (n = 2) or symptoms (n = 3) within 21 days of the procedure, but for whom no causal link with tracheostomy could be established. Four of the 5 individuals with symptoms or infection reported that they had been involved in at least one tracheostomy in a patient with COVID-19 without documented conversion to COVID-19–negative status. Yet, this 80% rate of “no negative test” exactly matches the base rate in the entire sample, in which only 20% of patients undergoing tracheostomy had negative results on follow-up COVID-19 testing (our institutional protocol did not require a negative COVID-19 test result in patients undergoing tracheostomy). For reference, the overall rate of community spread for the county in which this study took place (Washtenaw County) was 4673.5 cases per million (0.47%) throughout the study period.²⁸ 

Achieving the buy-in of our entire multidisciplinary team was an important aspect of ensuring adherence to the protocol. Many have reported success engaging multidisciplinary teams when caring for tracheostomy patients both before^29,30  and throughout the COVID-19 pandemic.^31–33  Our collaborative approach emphasized diverse stakeholder engagement across the health care enterprise. This collaboration was particularly important in enabling providers to feel safe during the acute phase of the pandemic, during which time a complex relationship existed between ensuring the safety of health care personnel and ensuring patient-centered outcomes.³⁴  Just as interdisciplinary rounds have been shown to improve patient outcomes,³⁵  cultivating this interprofessional approach was essential to ensuring both staff safety and optimal patient care.

Our protocol had several important elements. Tracheostomies were to be performed preferentially in the intensive care unit to minimize transport of patients, and procedures took place in either negative-pressure rooms or, as a secondary choice, in externally vented rooms with a HEPA filter. In such cases, we favored an open versus a percutaneous procedure, given the lack of high-quality data on aerosolization and historical reports from the severe acute respiratory syndrome (SARS) pandemic.^36,37  Additionally, our protocol recommended the use of a PAPR or N95 mask with eye protection for all tracheostomy-related cases and procedures for COVID-19–positive patients and patients with an unknown COVID-19 status. Our survey results indicated that respondents were largely aware of and adherent to these protocol recommendations.

It is important to note that the survey response rate varied between provider groups, with responses from nursing, respiratory care practitioners, and surgical technology personnel being less likely than responses from physicians. It is not clear whether this difference reflects particular group engagement, responsiveness to email, or other factors, but it highlights the critical importance of tailoring assessments of safety interventions to reach all members of the health care team. Similarly, self-reporting identified a small number of cases and symptoms that were not detected by routine occupational health processes, highlighting the importance of communication and the challenges of process-building for surveillance.

Testing accuracy throughout a patient’s disease course is still not well understood,³⁸  and guidelines on pretracheostomy testing vary widely.²⁵  The results of follow-up COVID-19 testing before tracheostomy may be misleading because counts of dead or inert virus can be amplified for weeks or months after infectivity has subsided.^39,40  Given these constraints, as well as the limited availability of testing materials at the beginning of the pandemic,⁴¹  our providers did not rely solely on COVID-19 testing to influence their decision-making. Given that widescale testing has become more readily available, a negative COVID-19 test result before a tracheostomy procedure may provide additional assurance of waning of viral load with corresponding reduction in transmission risk. At the same time, the role of antibody testing in obviating the need for PCR testing in this setting is unclear, given that vestiges of former infection may persist, yielding positive test results in patients who are no longer infectious.

Knowing when to perform tracheostomy can be challenging.^42–44  Tracheostomy practice remains diverse in terms of timing, with some institutions still preferring longer periods of use of a closed ventilator circuit versus earlier transition to an open system. Some patients may benefit from early tracheostomy (<4-7 days) and others from delay (>14 days), and optimal approaches to ventilation, weaning, and decannulation continue to be refined.^45–49  In the landmark Trac-Man trial, early tracheostomy (within 4 days) was not found to improve 30-day mortality or secondary outcomes in intubated patients.⁵⁰  This finding has informed evidence-based recommendations to perform tracheostomy at or after 10 days. Viral load and infectivity are now believed to be markedly reduced after 10 days of ventilatory support, which typically corresponds to a significantly longer time since initial infection.⁴²  Although our protocol initially endorsed deferral of tracheostomy in patients with COVID-19 until after 21 days, practices at our institution evolved in response to emerging evidence, resulting in guidance that earlier timing of tracheostomy might improve patient outcomes without increased risk to health care personnel.²³ 

Our survey participants provided suggestions for improving the tracheostomy protocol. For example, although the extant literature⁵¹  and our study data provide no evidence of the superiority of PAPRs over N95 masks in reducing the risk for transmission, dedicated PAPR availability may improve perceived safety. Given the high rates of anxiety, moral distress, and health care provider burnout documented amid the pandemic,^52,53  such requests warrant due consideration.

Two other quality improvement opportunities involved communication and developing dedicated tracheostomy teams. Some respondents advocated for team huddles, especially with physician leadership, before entering the room. Other respondents proposed dedicated tracheostomy teams—which have been recommended³⁶  and implemented at other institutions^19,54 —to ensure each provider is comfortable with the protocol and to minimize the number of providers in each room. The value of dedicated equipment and teams must be weighed against workforce demands amid acute surges.

Last, the need for wide dissemination of protocols was evident in survey responses. Although most participants (92%) described being either familiar or somewhat familiar with our tracheostomy protocol, 9 participants were unfamiliar with it. Additionally, 7 breaches of protocol were reported in our study. These findings underscore the need to circulate and familiarize staff with key protocol measures, ensuring education for every individual who may be involved in a tracheostomy for a patient with COVID-19.

Our work has several implications for nurses involved in the care of patients with a tracheostomy tube, both during the initial procedure and during subsequent patient care. Nurses should follow the same protective measures as proceduralists and should recognize the potential for aerosol generation during routine postoperative procedures, such as suctioning of the tracheostomy tube, cleaning of the tracheostomy site, and any maneuvering of patients or circuits. Nurses also play a critical role in partnering with physicians and allied health professionals to ensure safe and coordinated care, including care provided by respiratory therapists, speech language pathologists, and other members of the multidisciplinary team.⁴⁵ 

This study is not without limitations. Our method of identifying health care personnel involved in tracheostomies was imperfect, as evidenced by the 6 individuals surveyed who reported not being involved in a tracheostomy during this period. Similarly, some individuals involved in tracheostomies were not surveyed, including those who were travelers or who may not have been documented in charting. The survey was also subject to response bias, though our populations of respondents and nonrespondents were similar, but not identical, with respect to health care role.

Our assessment of infection control practice and adherence to protocols (including proper donning and doffing of PPE) is limited by the self-reported nature of our survey; however, our findings do reflect true clinician practice. Among respondents who reported positive symptoms or testing, we were unable to confirm whether transmission occurred during tracheostomy procedures. It is possible that respondents with nonspecific symptoms did not contract COVID-19 altogether or that the 2 participants with confirmed infection contracted the virus in another health care setting or in the community. Community exposure was a significant predictor of infection of health care personnel during this period.⁵⁵  Three of these 5 individuals described other high-risk exposures within the same time frame.

We also did not study COVID-19 infection in personnel providing postprocedural care. Last, participants in this study were from a single institution, limiting our study’s generalizability. Considerations may differ in different geographies, particularly in relation to available supplies that may differ widely in resource-rich versus low- and middle-income countries.

With the shifting landscape of the pandemic and the evolution of COVID-19 treatment, this work continues to be relevant. Rapid viral mutations and pockets of persistent infection in undervaccinated regions keep institutional commitments to risk reduction for health care providers as a vital concern. Despite the inability to determine causality, we were able to demonstrate that, even before vaccination was available, participation in tracheostomy under an institutional protocol was safe for multidisciplinary personnel and was not associated with high rates of transmission. The findings of this study inform the tracheostomy safety literature and our ongoing responses to future emergent challenges.

Our survey results indicate that health care personnel were largely familiar with and adherent to institutional protocols and that most providers felt safe while performing tracheostomies on COVID-19 patients. However, a small portion of our survey respondents reported possible infection during the 21-day window following tracheostomy, introducing the possibility of rare instances of viral transmission among physicians, nurses, surgical technologists, or respiratory care professionals. Procedural risks and safety should therefore be considered for all health care personnel participating in tracheostomy. Potentially preventable breaches in protocol were also identified, as was inconsistent reporting to central health services. Interprofessional teams developing and adapting protocols should consider elements of equipment, perioperative testing, continued education, and preprocedure safety huddles to ensure closed-loop communication, with the goal of reducing transmission risk to zero.