Discussion

Over the last 2 years, the SARS-CoV-2 pandemic has placed extraordinary and excessive demands on the healthcare system and HCPs from increased risk of infection, staff shortage, loss of productivity and/or financial burden.4 7 8 Our cohort illustrates the risk and impact of SARS-CoV-2 occupational exposure on HCP, particularly the risk posed by working alongside coworkers who may be working with an asymptomatic or presymptomatic infection themselves.

Similar to previously published studies,2 3 8 our cohort demonstrated the increased risk of developing COVID-19 infection after occupational exposure to a patient or coworker source with an overall post exposure infection rate of 4.5%. While the occupational risk of SARS-CoV-2 infection in HCP has been well defined in other studies,2 4 the difference in risk from occupational exposures to different source types has been inconsistent, with some studies reporting increased infection rates following exposures to patient sources9 10 while others have found an increased occupational exposure risk from infected coworkers.5 11 12 HCP in our study were three times more likely to test positive after an exposure to a coworker vs a patient, consistent with the previously reported result by Fell et al demonstrating a postexposure conversion rate of 3.8% when the source was a coworker compared with 1.3% for a patient source.11

In our analysis, we observed that more HCP were exposed when the source was a patient than a coworker (3.4 vs 2.2 HCP exposed per source, respectively), while the more significant exposures (high risk) were more common when the source was a coworker. This finding is consistent with the previously reported results by Emecen et al and Maltezou et al.5 8

The higher number of exposed HCP for each patient source is likely attributable to two common scenarios: lack of eyewear use by HCP caring for an inpatient not initially suspected of having COVID-19, and presence during or shortly after an AGP without the use of a respirator.13 Delayed and/or incidentally discovered infections, potentially due to a preadmission exposure or an initial false negative test, occurred throughout the study period, despite routine admission testing for COVID-19 starting in April 2020. Prolonged or continuous AGPs such as continuous positive airway pressure, on inpatients not clinically suspected of having COVID-19 infection, tended to expose large numbers of HCP if the patient was subsequently identified as having COVID-19 infection. The lower number of exposed HCP per coworker source may reflect general adherence to precautions such as masking and social distancing, with isolated close contact during breaks with one or two individuals. The number of coworkers exposed could also have been underestimated, given the contact tracing process largely relies on the infected worker’s recall of HCP with whom they had contact while communicable.

Most exposures to patient sources were medium risk while the vast majority of exposures to coworker were high risk. This discrepancy may be attributed to a difference in HCPs perceived risk when in close contact with patients compared with coworkers. HCP may use appropriate PPE when providing patient care but may be less diligent in PPE use in proximity to coworkers who are asymptomatic and appear healthy. In some work settings, masking and/or physical distancing may not be feasible such as in breakrooms and shared workspaces. Socialising with coworkers outside of work can also lead HCP to perceive coworkers as low risk at work, especially given the extended periods of time that HCP spend together.14 Additionally, the increased awareness of transmission risk factors and improvements in infection prevention and control measures implemented in healthcare setting reduced the intensity of exposure to COVID-19 from patient sources.12

The increased frequency of high-risk exposures to coworkers is especially concerning given the higher risk of post exposure infection demonstrated in our cohort. The institutional impact of high-risk exposures in the setting of staff shortages can be profound when considering HCP absenteeism due to exposure-related quarantine and/or isolation, in addition to the loss of productivity associated with mortality and years lost due to disability from COVID-19 infection and post-COVID-19 sequalae.7

Our study is subject to limitations. First, the observation period predated the availability of COVID-19 vaccines. Despite widespread vaccine requirements for HCP, the relationships between exposure source, exposure intensity and occupational infection risk are likely to remain relevant. Although vaccines are highly effective, the potential remains for waning immunity and the emergence of vaccine-resistant new variants. The perceived risk of encountering an infection from coworkers among HCPs, especially vaccinated coworkers, may be underappreciated by HCP as immunity wanes. Second, the more accessible identification of all HCP potentially exposed to a patient through available features in the electronic medical record, compared with HCP exposure to a coworker source which is subject to recall bias toward HCP with the most intense exposure. Third, the data regarding the exposure history and allocation to occupational vs nonoccupational exposure after a positive COVID-19 test was obtained from the OHS staff’s detailed interviews, however, we were not able to rule out non-occupational exposure that may have coincided with medium or high-risk exposures at work. The overall effect would be overestimating occupational exposure as the source of infection in healthcare workers. Fourth, during the first 6 weeks of the study, postexposure testing was only recommended for symptomatic individuals, which would limit the ability to identify infection in asymptomatic HCP in the initial 6 weeks of the study. Additionally, infected workers who are asymptomatic/presymptomatic and feel well enough to be at work are likely in the earliest, most infectious period of their illness, while hospitalised patients have primarily lower respiratory infection and may be less contagious.15 16 Finally, the continual emergence of new variants coupled with widespread vaccination limit the generalisability of the infection risk estimates observed.