We used a data set from a recent cross-sectional survey by Statistics Canada on the impacts of COVID-19 on HCWs. Unlike other traditional Statistics Canada surveys, a random selection of participants was not used. Instead, Statistics Canada sent an email invitation to HCWs across Canada. Then, a snowball sampling procedure was used. The invitation included a link to a web-based survey that was available through Statistics Canada’s web page. Accordingly, 18,139 responses were collected between November 24 and December 13, 2020, across 7 provincial and territorial regions in Canada. No data were collected for this survey beyond these dates. Only responses by HCWs were included in the data set. The responses were completely anonymized by Statistics Canada.

The questionnaire asked HCWs for information related to the job environment, demographics, geography, and information on access to PPE as background information. Adaptive questioning was used. Our study was limited to these 3 factors categories only: demographics, occupational, and access to PPE based on the available data from Statistics Canada. In assessing mental health well-being, the survey asked HCWs: compared to before the COVID-19 pandemic, how would you say your mental health is now? HCWs self-reported their perceived mental health on a 5-point Likert scale: much better now, somewhat better now, about the same, somewhat worse now, much worse now.

Our study uniquely studied the association of various interesting factors, as depicted in Figure 1, with HCW mental health conditions as compared to before the start of the pandemic. To consider each factor separately, we used a multivariate ordinal logistic regression [17,18]. The model is defined in Multimedia Appendix 1. The dependent variable is the state of mental health of the respondents. We reduced the mental health state categories from the 5 mentioned above to 3 (improved, same, or worsened) to yield statistically significant model estimates. The independent variables include demographics such as age, sex, immigration status, and the province of the workplace. Occupational variables include work status, occupational group, and exposure to COVID-19 cases. The model also estimated the association of HCWs’ access to a variety of PPE, such as respirators, eye protection, and face shields. Access to PPE is also reported by the HCW on a 5-point Likert scale, which we reduced to 3 categories (always available, sometimes available, or never available) to yield statistically significant model estimates. All responses are self-reported. Analyses were conducted in RStudio (version 1.4.1717; Posit, PBC).

Data collected by Statistics Canada were reviewed based on the following principles: privacy, security, transparency, accountability, trust, sustainability, data quality, and fairness, as well as well vetted to be harmless to participants or the public. All ethical approvals were considered by Statistics Canada. No further ethical approval to use these data was required by the authors and the survey is made available to the public. This study is exempt from further ethical approval. Furthermore, the authors also did not have access to information related to the development and testing of the survey. The authors are also not aware of any compensation provided to survey participants.

From the full data set of 18,139 responses, we considered nonduplicate responses that provided valid answers to all questions of interest to this study (N=13,990). The remaining responses contained missing or invalid responses. For the considered population, Table 1 outlines the number of observations and percentages for different demographic factors such as age group, sex, and immigration status. The number of observations and percentages were also calculated for occupational factors including the province of the primary workplace, work status (full-time vs part-time), frontline work status, occupational group, and exposure to confirmed or suspected COVID-19 cases. Additionally, the last 3 rows of Table 1 outline PPE access levels for 3 different PPE: face shields, eye protection, and respirators. Overall 4261 (31%) HCWs were younger than 35 years. The majority were female participants (n=12,682, 91%). In total, 12,510 (89%) HCWs were nonimmigrant. HCW workers were distributed across 7 regions, with the largest group of 6626 (47%) working in Ontario. The majority (n=10,152, 73%) of HCWs worked full-time and 5511 (39%) were nonfrontline allied health professionals. A total of 8188 (59%) HCWs did not report exposures to confirmed or suspected COVID-19 cases. As for access to PPE, 10,758 (77%) HCWs always had access to face shields, 10,118 (72%) always had access to eye protection, and 6898 (49%) always had access to respirators.

In the following, we present the results for estimating the associations between the various factors and the likelihood of worsened mental health conditions for HCWs.

The statistical analysis in this study found that HCWs who are younger than 35 years old were found to be more likely to exhibit worsened mental health conditions than HCWs aged 35-44 years. Male HCWs were less likely to exhibit worsened mental health conditions than female HCWs. Immigrant HCWs were also less likely to exhibit worsened mental health conditions than nonimmigrant HCWs. In contrast, HCWs working in Alberta had a higher likelihood of worsened mental health conditions than HCWs working elsewhere (Atlantic provinces, Quebec, Manitoba, Saskatchewan, Ontario, British Columbia, and Northern Territories). Frontline workers were more likely to exhibit worsened mental health conditions than nonfrontline HCWs. Part-time HCWs were less likely to exhibit worsened mental health conditions than full-time HCWs. HCWs who reported encountering COVID-19 cases were more likely to exhibit worsened mental health conditions than HCWs who reported no contact with the disease. As for PPE, HCWs who never had access to respirators, eye protection, and face shields were more likely to exhibit worsened mental health conditions than those who always had access to the same PPE, respectively.

Previous research has shown that shock events can result in psychological trauma to HCWs [19-21]. To the best of our knowledge, our study is the first to assess the association of a comprehensive variety of factors (demographics, occupational factors, and access to PPE) independently with the mental health conditions of HCWs.

With regard to demographic factors, a study found sex to be associated with worsened mental health conditions of HCWs. Female medical staff exhibited a higher incidence of severe anxiety than their male counterparts [22]. Sex was also found as a predictor of increased anxiety and distress in a study that found female HCWs more vulnerable to such conditions than male HCWs [23]. Our study findings corroborate these previous findings. Similarly, for age, our results demonstrated that older HCWs exhibited less likelihood of worsened mental health conditions. This agrees with findings from a study in the Middle East that found older workers enjoyed better mental health conditions than younger workers [24]. Older HCWs with longer work experience seemed to have handled the pandemic better than their younger counterparts. Another study found that HCWs 40 years or older were less likely to report higher anxiety during the pandemic than younger HCWs [25]. Our results confirm such findings as well.

As for occupational factors, our findings align with a cross-sectional study conducted in Oman, which found frontline workers to be more likely to have anxiety and sleep problems [26]. This was attributed to frontline workers’ increased awareness of the mortality rate of COVID-19 and their fear of contracting the virus. Similar findings were reported by a study conducted in China as well [27]. Exposure to COVID-19 cases has been recognized as a risk factor associated with an increased likelihood of mental health issues in the literature. For instance, in France, female urologists working on the frontline were 1.41 times more likely to feel a degree of stress during their duties. Of this group, those who worked in a department where patients with COVID-19 are treated were 1.85 times more likely to report a degree of stress during work duties [28]. This was attributed to the workers’ fear of infection and the spread of the virus. Direct involvement with COVID-19 care is also found to be highly associated with fear, depression, and anxiety as compared with those working under lower risk conditions [29]. The authors attributed this to the workers’ fear of bringing the virus to their families at home.

As for the work setting, contrary to our findings, the literature reports higher levels of fear and anxiety among part-time HCWs than among full-time [30]. Another study in France found part-time HCWs to exhibit a greater association with distress [31]. Given the small number of part-time HCWs in our study, caution should be observed in interpreting the results. Regarding factors related to access to PPE, access to PPE was found to be associated with better health and less stress [24]. Lack of access to PPE was found to be a major source of HCW stress in a study that surveyed emergency physicians [32]. In Canada, a study found that inadequate PPE supply is associated with increased symptoms of anxiety and depression among HCWs [33]. Workers were concerned about the ability to access sufficient PPE during work hours. These findings are in line with our results. Compared to studies in the literature, our study is the first comprehensive study that assesses the association of a diverse pool of factors including demographics, occupational, and access to PPE, with HCWs’ mental health conditions.

Our comprehensive study sheds light on the association of various factors with HCWs’ mental health after the start of the pandemic. It helps in understanding the vulnerability of various HCW groups to mental health during such events. Certain groups were at substantially higher risk of exhibiting worsened mental health conditions after the pandemic, hence the need for a specialized support program to target this group. As a direct implication, the findings can be used to inform guidelines for mental health support for HCWs during future public health emergencies. Such mental health support may be directed more specifically to more vulnerable groups. Literature has pointed to the importance of an evidence-based approach to designing mental health support programs for HCWs [34,35]. For instance, a study in Alberta during the pandemic pointed to the importance of understanding HCWs’ occupational settings and mental health mitigation techniques [16]. Others have discussed the importance of support programs for HCWs during the pandemic based on need and background [36,37]. More importantly, many studies highlighted the importance of designing personalized mental health support programs based on various factors including demographic and occupational factors [38,39]. Some of these programs used mobile technology to offer mental health support to HCWs [35,40]. As such, our study aims to inform stakeholders of such factors associated with HCW groups most vulnerable to future events such as the pandemic and health crises.

This study has several limitations. First, the scope of the study was limited to assessing mental health in general without measuring different mental condition forms such as depression, anxiety, and fear. This is due to the limited data offered by the survey we used. Second, participants were asked, “Compared to before the COVID-19 pandemic, how would you say your mental health is now?” That is, HCWs had to objectively compare their mental health state from the time they received the survey to sometime in the past. There may be some variation in how HCWs perceive this comparison. Third, data collected were at a single point in time, which limits the ability to compare mental health assessments to later periods in the pandemic and during subsequent pandemic waves. Fourth, no information about HCW medical history is collected in this survey. In future research, we recommend collecting data at different points in the given event to allow for a time-series analysis and comparison at different crucial points of the event. We also suggest the collection of distress information and whether the conditions persisted beyond the initial shockwave. HCWs’ years of work experience and ethnicity information can also be useful in future studies. We also believe medical history, particularly a preexisting history of mental health disorders, is an important factor associated with the mental health conditions of HCWs and should be considered in future studies.

This study investigated the factors associated with worsened mental health conditions of HCWs in Canada during the pandemic. Our study suggested the association of various factors with the likelihood of HCWs exhibiting worsened mental health conditions as compared to feeling neutral and better. In agreement with the literature, our findings concluded that younger (vs older), female (vs male), nonimmigrant (vs immigrant), full-time (vs part-time), and frontline (vs nonfrontline) HCWs living in Alberta (vs other provinces) exhibited a higher likelihood of worsened mental health conditions than those who felt neutral or better. Those who reported concerns about access to PPE also exhibited the same trend. Such findings can guide the future development of health care programming and inform mental health support planning for HCWs. COVID-19 is a shocking event that introduced uncertainty to the health care system.