Risk for Acquiring COVID-19 Illness Among Emergency Medical Service Personnel Exposed to Aerosol-Generating Procedures

Aubrey Brown; Leilani Schwarcz; Catherine R. Counts; Leslie M. Barnard; Betty Y. Yang; Jamie M. Emert; Andrew Latimer; Christopher Drucker; John Lynch; Peter J. Kudenchuk; Michael R. Sayre; Thomas Rea


Emerging Infectious Diseases. 2021;27(9):2340-2348. 

In This Article


Study Design, Population, and Setting

We conducted a retrospective cohort study to evaluate the risk for COVID-19 infection among EMS providers caring for patients in King County, Washington, USA, during February 16–July 31, 2020. When determining risk for COVID-19, we considered all EMS provider-patient encounters and individual EMS providers involved in those encounters. The investigation was designed and reported with consideration of the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines[14] and approved by the University of Washington and Seattle and King County Public Health and University of Washington public health review boards.

King County is a large metropolitan region encompassing the city of Seattle and covering ≈2,300 square miles with ≈2.3 million residents living in urban, suburban, and rural areas. The EMS system is 2-tiered, the first tier comprising 27 firefighter and emergency medical technician departments and the second tier 5 paramedic departments serving multiple emergency medical technician departments for responding to more serious medical emergencies. EMS teams of 2–7 providers respond to calls based on dispatcher-determined acuity. In general, fire department or private basic life support ambulance units transport medically stable patients to hospitals and advanced life support paramedic units transport patients needing more acute care.

EMS COVID-19 Protocols

Seattle and King County EMS management developed protocols for screening and care of patients at risk for having COVID-19.[15] EMS PPE protocols include wearing a mask, eye protection, gloves, and a gown. Surgical masks were considered sufficient for treating patients not requiring AGP, but an N95 respirator was required when patients underwent AGPs. HEPA (high efficiency particulate air) filters were added to ventilation bags. Otherwise, clinical protocols did not change in response to the pandemic. For example, the EMS system continued to support the use of endotracheal intubation and manual CPR to treat out-of-hospital cardiac arrest.[13]

Data Sources, Linkages, and Abstraction

The Seattle and King County EMS Division of Public Health maintains an encounter-level electronic health record of each EMS response using software from ESO Solutions Inc. (https://www.eso.com). The EMS record for each incident contains information about patient and EMS provider identities, chief complaints, signs and symptoms, EMS care, and PPE use by providers. The state of Washington Disease Reporting System (WDRS) contains names, dates of birth, test dates, and results for all persons who have been tested for SARS-CoV-2 within the state. Seattle and King County Public Health administers the EMS system, enabling identification of EMS encounters with patients who have COVID-19.[15] To obtain patient COVID-19 status, we linked WDRS with EMS electronic health records using a multistep algorithm including the patient's first and last names and date of birth; identification through this linkage was followed by human confirmation of the potential link.

In addition to the linking process for COVID-19 status, we determined the health-related vital status of patients with COVID-19 by linking those patients with Washington State Department of Health vital records available through December 1, 2020, All study information for COVID-19 patient encounters was abstracted into a secure Research Electronic Data Capture (REDCap, https://www.project-redcap.org) platform by using a uniform data abstraction form supported by a data dictionary.[16] The abstract recorded a review of the narrative and discrete data fields from the dispatch and EMS records.

Exposure and Data Definitions

COVID-19 Patient Classification. A provider was considered to have encountered a patient with COVID-19 if the patient had a positive SARS-CoV-2 swab sample result determined by using real-time reverse transcription PCR (rRT-PCR) ≤10 d before or ≤3 d after an EMS encounter, on the basis of data from the linked EMS and WDRS records. We chose ≤10 d as a criterion on the basis of the 10-day infectious window after onset of symptoms. We used ≤3 d as a criterion after the EMS encounter recognizing that not all patients had been tested upon hospital arrival, especially in the first few months of the pandemic. In addition, a minority of patients were not transported by EMS and had subsequent follow-up for testing even though the EMS encounter appeared to be for illness consistent with COVID-19.[2]

AGP Definition and Classification. For this study, we classified endotracheal intubation, supraglottic airway insertion, bag-valve-mask (BVM) ventilation, continuous positive airway pressure nonrebreather mask oxygen, and nebulizer medication therapy as AGPs.[4] Although the standards for AGP are not fully defined, nonrebreather masks routinely involve using higher-flow oxygen (15 L/min) and require applying and manipulating face masks, which may increase transmission risk.[4,17,18] We did not classify use of low-flow nasal cannula oxygen as an AGP. In an EMS patient-encounter setting, CPR always involves both chest compressions and BVM ventilation, which constitutes an AGP. We identified AGP procedure usage from the EMS records by searching electronic text records for key phrases in the narratives or discrete electronic data elements that recorded AGP procedures. We evaluated the accuracy of this method to identify AGP by manually reviewing records of all EMS encounters with COVID-19 patients.

Classifying EMS Provider Person-Days at Risk. For each day of the study period, each EMS provider's day was classified into 1 of 4 mutually exclusive cohorts based on the time interval after COVID-19 patient encounters, if any, and whether or not AGPs were used. Person-days were classified into cohort 1 for COVID-19 patient encounters that involved ≥1 AGPs during the 2–14 d incubation period, cohort 2 for COVID-19 patient encounters that did not involve AGPs during the 2–14 d incubation period, cohort 3 for COVID-19 patient encounters before or after the 2–14 d incubation period, or cohort 4 if the provider had no COVID-19 patient encounters during the study period. Individual EMS providers could contribute discrete person-days to different cohorts, except for cohort 4.

We considered EMS providers at risk for transmission from a patient for 2–14 d after an encounter with a COVID-19 patient (Figure 1), because the biology of transmission and illness indicates that the COVID-19 incubation period is 2–14 d.[19] If an EMS provider tested positive for SARS-CoV-2 in the 2–14 d incubation period after treating a COVID-19–positive patient, the infection was attributed to the encounter. For classification, once an EMS provider completed the 14 d incubation period without SARS-CoV-2 infection, the provider's person-days for subsequent days would transition from cohort 1 or 2 to cohort 3 until the provider was involved with another patient with COVID-19.

Figure 1.

Examples of classification of EMS provider person-days at risk within 2–14 d after COVID-19 patient encounters, King County, Washington, February 16–July 31, 2020. The boxes correspond to the number of person-days an emergency medical services provider contributes to each mutually exclusive risk group. The first row (provider A) demonstrates a COVID-19 patient encounter without an AGP. The provider is classified at risk for COVID-19 transmission because of a patient treated without AGP within 2–14 d after encounter. After the incubation window ends, the EMS provider transitions back to person-days classification of COVID-19 patient outside the incubation period (cohort 3). The second row (provider B) demonstrates classification of person-days from COVID-19 patient without AGP and then with AGP. Person-days transitions from COVID-19 patient encounter without AGP (cohort 2) to patient encounter with AGP (cohort 1). The example illustrates the classification hierarchy that classified the patient into the AGP incubation period when a provider had overlap of person-days following distinct encounters caring for COVID-19 patients without an AGP and then with an AGP. After the incubation window, the EMS provider will transition back to person-days classification of COVID-19 patient outside the incubation period (group 3). AGP, aerosol-generating procedure; COVID-19, coronavirus disease; EMS, emergency medical service.

For days when a provider had multiple COVID-19 patient encounters and ≥1 involved an AGP, the provider's person-hours for that day were classified into cohort 1, given that AGP use is considered to possess greater intrinsic transmission risk. EMS providers could be diagnosed with COVID-19 on a person-day in any of the 4 cohorts. After a provider's first rRT-PCR–positive SARS-CoV-2 swab result, they were censored from the analysis and did not contribute additional person-days to any cohort. SARS-CoV-2 reinfection was not diagnosed in any provider.

Outcome Measures

We used COVID-19 infections among EMS providers as determined from the WDRS registry during February 15–August 14, 2020, as the primary outcome measure. We extended the period for assessing COVID-19 to August 14, two weeks beyond the final day for recording person-days, to ensure we captured infections identified ≥14 d after COVID-19 patient encounters within the study period.

As part of COVID-19 surveillance, EMS implemented a screening process for potential COVID-19 illness among EMS personnel at the outset of each shift comprising a temperature check and observation for symptoms of medical illness. EMS personnel were guided by a return-to-work algorithm that recommended COVID-19 rRT-PCR testing for any acute illness acquired on or off duty in an effort to limit the risk of provider-to-provider transmission and maintain workplace safety.[15]


We performed descriptive analyses at the encounter, patient, and EMS provider levels. We stratified provider encounters and classified person-days according to patient COVID-19 status and whether or not treatment included ≥1 AGPs. EMS providers were censored from the study on the date they were diagnosed with COVID-19 or at the end of the follow-up period (August 15, 2020) if never diagnosed with COVID-19. We then calculated the incidence of COVID-19 infection among EMS providers on the basis of person-days at risk from COVID-19 patient encounters. We calculated the incidence rate ratio using the collective person-days from cohort 3, the cohort including person-days before or after the 2–14 d incubation period of a COVID-19 patient encounter, as the referent group because this approach enabled providers to serve as their own controls when evaluating the risk attributable to COVID-19 patient encounters. In a post hoc analysis, we combined the person-days from cohorts 1 and 2 to evaluate the overall COVID-19 incidence among EMS providers attributed to a COVID-19 patient encounter regardless of AGP use.