Determinants of burnout in Canadian health care workers during the COVID-19 pandemic

ABSTRACT

Background: Health care workers (HCWs) are among the most vulnerable groups to experience burnout during the coronavirus (COVID-19) pandemic. Understanding the risk and protective factors of burnout is crucial in guiding the development of interventions; however, the understanding of burnout determinants in the Canadian HCW population remains limited.

Objective: Identify risk and protective factors associated with burnout in Canadian HCWs during the COVID-19 pandemic and evaluate organizational factors as moderators in the relationship between COVID-19 contact and burnout.

Methods: Data were drawn from an online longitudinal survey of Canadian HCWs collected between 26 June 2020 and 31 December 2020. Participants completed questions pertaining to their well-being, burnout, workplace support and concerns relating to the COVID-19 pandemic. Baseline data from 1029 HCWs were included in the analysis. Independent samples t-tests and multiple linear regression were used to evaluate factors associated with burnout scores.

Results: HCWs in contact with COVID-19 patients showed significantly higher likelihood of probable burnout than HCWs not directly providing care to COVID-19 patients. Fewer years of work experience was associated with a higher likelihood of probable burnout, whereas stronger workplace support, organizational leadership, supervisory leadership, and a favourable ethical climate were associated with a decreased likelihood of probable burnout. Workplace support, organizational leadership, supervisory leadership, and ethical climate did not moderate the associations between contact with COVID-19 patients and burnout.

Conclusions: Our findings suggest that HCWs who worked directly with COVID-19 patients, had fewer years of work experience, and perceived poor workplace support, organizational leadership, supervisory leadership and ethical climate were at higher risk of burnout. Ensuring reasonable work hours, adequate support from management, and fostering an ethical work environment are potential organizational-level strategies to maintain HCWs’ well-being.

HIGHLIGHTS

• Canadian HCWs endorsed high levels of burnout during the COVID-19 pandemic.

• Having direct contact with COVID-19 patients and having fewer years of work experience were associated with a higher likelihood of probable burnout.

• Having stronger workplace support, greater perceived organizational and supervisory leadership, and a favourable ethical climate were associated with a lower likelihood of probable burnout.

Antecedentes: Los trabajadores de la salud (HCWs en su sigla en inglés) están entre los grupos más vulnerables a experimentar burnout durante la pandemia del coronavirus (COVID-19). Es crucial entender los factores de riesgo y protectores del burnout para guiar el desarrollo de intervenciones; sin embargo, el entendimiento de los determinantes del burnout en la población de HCW canadienses permanece limitado.

Objetivo: Identificar los factores de riesgo y protectores asociados con el burnout en los HCWs canadienses durante la pandemia del COVID-19 y evaluar los factores organizacionales como moderadores en la relación entre el contacto del COVID-19 y el burnout.

Métodos: Los datos fueron obtenidos de una encuesta longitudinal en línea de los HCWs canadienses recolectadas entre el 26 de junio de 2020 y el 31 de diciembre de 2020. Los participantes completaron preguntas relacionadas con su bienestar, burnout, apoyo en el lugar de trabajo y preocupaciones con respecto a la pandemia del COVID-19. Se incluyeron en el análisis los datos de tamizaje de 1029 HCWs. Se usaron prueba t para muestras independientes y regresión lineal múltiple para evaluar los factores asociados con los puntajes de burnout.

Resultados: Los HCWs en contacto con pacientes del COVID-19 mostraron significativamente probabilidades más altas de posible burnout que los HCWs que no proporcionaron cuidado directo a pacientes del COVID-19. Menos años de experiencia laboral se asoció con una probabilidad significativamente más alta de probable burnout, mientras que un apoyo más fuerte en el lugar de trabajo, liderazgo organizacional, liderazgo del supervisor, y un clima ético favorable se asociaron con una disminución de la probabilidad de posible burnout. El apoyo en el lugar de trabajo, liderazgo organizacional, liderazgo del supervisor, y clima ético no moderaron las asociaciones entre el contacto con pacientes del COVID-19 y burnout.

Conclusión: Nuestros hallazgos sugieren que los HCWs que trabajaron directamente con los pacientes del COVID-19, tenían menos años de experiencia laboral, y percibían un apoyo en el lugar de trabajo, un liderazgo organizacional, un liderazgo de supervisión, y clima ético deficientes, tenían un mayor riesgo de burnout. Garantizar horarios de trabajo razonables, apoyo adecuado de la administración, y promover un ambiente de trabajo ético son potenciales estrategias a nivel organizacional para mantener el bienestar de los HCWs.

KEYWORDS:

• Burnout
• COVID-19
• health care workers
• organizational environment
• ethical work environment

PALABRAS CLAVE:

• Burnout
• COVID-19
• trabajadores de la salud
• ambiente organización
• ambiente laboral ético

1. Introduction

The coronavirus (COVID-19) pandemic has placed health care systems under significant stress (Wilbiks et al., 2021). Frontline health care workers (HCWs) face challenges, including heavy workloads disrupting work-life balance, fear of infecting themselves and others, lack of personal protective equipment (PPE), and moral distress from being unable to provide adequate patient care (De Kock et al., 2021; Plouffe et al., 2021; Trumello et al., 2020). Together, these conditions place HCWs at risk of adverse mental health outcomes such as burnout (Saragih et al., 2021).

Burnout is a work-related syndrome caused by chronic job stress. Although burnout was already highly prevalent in HCWs before the COVID-19 pandemic, rates of burnout have only increased with the start of the pandemic. One US study showed that 62.8% of physicians had at least one manifestation of burnout in 2021, compared with 43.9% in 2017 (Shanafelt et al., 2022). Consequences of burnout in HCWs include impaired quality of care, lower patient satisfaction, medical errors, and increased turnover and shortages of HCWs (De Hert, 2020; Wilbiks et al., 2021). Furthermore, burnout can have a heavy psychological toll on HCWs, including increased anxiety, irritability, apathy, and alcohol and drug consumption (De Hert, 2020). Given the high prevalence of burnout among the HCW population, there is a need to identify risk and protective factors in order to guide health policymakers to develop interventions for vulnerable groups.

Research conducted during the COVID-19 pandemic has examined sociodemographic and work-related factors involved in the development of burnout. Among studies that have investigated sociodemographic risk factors, burnout was shown to be associated with being female (Evanoff et al., 2020; Jalili et al., 2021; Orru et al., 2021; Sung et al., 2021), being single, divorced, or separated (Aljabri et al., 2022), being younger (Evanoff et al., 2020; Jalili et al., 2021), and having fewer years of work experience (Aljabri et al., 2022). Burnout has also been reported to be associated with working with COVID-19 patients or in high-risk settings, such as emergency rooms or intensive care units (Bailey et al., 2021; Evanoff et al., 2020; Sung et al., 2021). We have learned about the impact of working closely with high-risk patients from the 2003 severe acute respiratory syndrome (SARS) outbreak, in which HCWs with direct contact with SARS patients reported higher levels of depression, anxiety, and stress than those who did not have contact with SARS patients (Maunder, 2004; McAlonan et al., 2007). Psychological distress in HCWs caring for infected patients may be due to vulnerability or loss of control, concerns about the spread of the virus to self, family, and others, heavy workloads and new demands at work, and being isolated (Evanoff et al., 2020; Sung et al., 2021; Wong et al., 2005). Having organizational and supervisory support during a pandemic may be instrumental to protecting HCWs from burnout, as burnout has been higher in those who experienced a lack of support, training, equipment, and information from management (Cyr et al., 2021; Evanoff et al., 2020; Fiabane et al., 2021; Pawlowicz-Szlarska et al., 2022). Specifically, one Canadian study showed that improved work-life quality, measured by factors such as work hours, appreciation of efforts, and fairness of work assignments, was associated with decreased burnout (Bailey et al., 2021). Having a favourable ethical climate, characterized by an atmosphere that increases ethical thoughts, mutual respect, and trust in the organization, also decreases burnout in the healthcare setting (Rivaz et al., 2020; Wlodarczyk & Lazarewicz, 2011).

Although some risk factors for burnout have been identified, including contact with COVID-19 patients (Bailey et al., 2021; Evanoff et al., 2020; Sung et al., 2021), it remains unclear whether specific facets of the work environment serve as buffers against the negative impacts of contact with COVID-19 patients. For example, among HCWs who specifically work with COVID-19 patients, their risk of burnout may be mitigated by protective workplace factors: having reasonable working hours to reduce exhaustion, leadership that provides HCWs with support, training, and confidence to reduce feelings of fear and vulnerability around the virus, and an environment that fosters safe discussions around ethical dilemmas. Examining these associations would better inform the implementation of organizational-level interventions to decrease the psychological burden of HCWs caring for COVID-19 patients.

Despite the growing volume of research exploring mental health during the COVID-19 pandemic, evidence of risk and protective factors for burnout in Canadian HCWs remains limited. Given that different countries vary in terms of their medical systems, availability of PPE, and employment and living conditions during the pandemic, it is important to evaluate factors contributing to burnout in individual countries and subsequently tailor guidelines to prioritize those at highest risk (Jahanshahi et al., 2020). For example, as Jahanshahi et al. found, age and education predicted distress in China but not in Iran. Therefore, sociodemographic and work-related factors contributing to burnout need to be elucidated specifically in Canadian HCWs. Past research has also not yet examined whether organizational factors moderate the associations between COVID-19 contact and burnout. If so, these specific protective factors may serve as a ‘buffer’ against burnout for individuals working with COVID-19 patients that might result from working in high-stress settings. Determining these relationships will allow policymakers to better allocate resources and interventions to groups at high risk of burnout.

1.1. Aims of the study

This study aimed to evaluate risk and protective factors for burnout in Canadian HCWs during the COVID-19 pandemic. Our specific objectives were to: (1) Evaluate differences in burnout between HCWs in direct contact with COVID-19 patients and those who are not; (2) Identify sociodemographic predictors of burnout, including age, sex, marital status, and years of work experience; and (3) Evaluate organizational factors, including workplace support, organizational leadership, supervisory leadership, and ethical climate as moderators in the relationship between COVID-19 contact and burnout.

2. Methods

2.1. Participants and procedure

Participants included 1029 English- or French-speaking HCWs employed across Canada. We defined HCWs as those who provide health treatment and advice based on formal training and experience, or who work to directly support those providers in a clinical setting. HCWs were eligible to participate if they were at least 18 years of age and employed as a HCW in Canada at the time of study onset. HCWs from across Canada were recruited via word of mouth, social media and online advertisements, participant recruitment websites, and media releases through Lawson Health Research Institute. This research was approved by the ethical review board at Western University (approval number 115894). Informed consent was obtained through a Letter of Information, presented to participants at the beginning of the survey which was offered in both English and French. Results presented here are based on baseline responses to a longitudinal survey, with initial responses in the cohort being collected between 26 June 2020 and 31 December 2020. Further details of our procedure are provided in our protocol paper (Liu et al., 2021).

2.2. Measures

2.2.1. Burnout

Burnout over the past month was assessed using the Expanded Well-Being Index (WBI; Dyrbye et al., 2016). The WBI consists of seven dichotomous (yes/no) items, with 1 point assigned to each ‘yes’ response, giving a summed score of 0–7. In addition, two items are used to add or subtract one point from participants’ initial scores. The first of these items evaluates how meaningful the individual’s work is to them (7-point scale from ‘very strongly disagree’ to ‘very strongly agree’), and the second item assesses work-life balance (5-point scale from ‘strongly agree’ to ‘strongly disagree’). Total scores range from −2 to 9, with higher scores indicating higher likelihood of probable burnout (Dyrbye et al., 2016). Given that no cutoff score for the presence of burnout has been suggested using the WBI, we used one of the dichotomous WBI questions on self-report of burnout to measure the prevalence of burnout among participants. The WBI has demonstrated strong reliability and validity in past research (Dyrbye et al., 2011; Dyrbye et al., 2014). In our current study, the WBI had a reliability of α = .67.

2.2.2. Organizational response to COVID-19

We assessed HCWs’ perceptions of their organization’s responses to the COVID-19 pandemic using an adapted version of the Pandemic Experiences and Perceptions Survey (PEPS; Leiter, 2020). For the purpose of our study, we collected data on workplace support (‘work-life impact’ subscale) and leadership domains of the PEPS. Workplace support included components such as work hours, fairness of work assignments, and social support, with response options ranging from 1 (‘strongly disagree’) to 5 (‘strongly agree’). The leadership domain was separated into organizational management and immediate supervisory management, with each component having response options ranging from 1 (‘not at all’) to 5 (‘frequently, if not always’). Mean scores on each subscale were calculated to create three separate scores, with higher scores indicating stronger workplace support, organizational leadership, or supervisory leadership. The validity of the PEPS has been supported in recent research (AlMulla, 2020). In our current study, workplace support, organizational leadership, and supervisory leadership had reliabilities of α = .83, .91, and .95, respectively.

2.2.3. Ethical climate

We used the 20-item Ethics Environment Questionnaire (EEQ; McDaniel, 1997) to measure HCWs’ perceptions of ethics in their organizations. Response options ranged from 1 (‘strongly disagree’) to 5 (‘strongly agree’). Mean scores were calculated, such that higher scores indicated a more ethical work environment. Past research has supported the reliability, validity, and unidimensional factor structure of the EEQ (e.g. α = .93; Corley et al., 2005; McDaniel, 1997). In our current study, the EEQ had a reliability of α = .93.

2.3. Data analysis

Analyses were performed using R statistical software, version 4.2.1 (R Core Team, 2022) and the ggcorrplot package (v0.1.4; Kassambara & Patil, 2022). Within-individual mode imputation was applied to individuals with 80% or higher response rates to address missing values. This was chosen as most questions were Likert-type, and mode imputation selects the response most similar to other activity by the individual to impute with (Tsai et al., 2018). We then calculated scale totals for those participants. Descriptive statistics were used to assess means, standard deviations, skewness, and kurtosis values for burnout and organizational variables. Bivariate correlations were also calculated between burnout and organizational variables. Prevalence of burnout was determined by the proportion of participants who answered ‘yes’ to the dichotomous WBI question on self-perception of burnout, and χ² test was used to compare prevalence among different occupations. To compare total WBI scores between HCWs in direct contact with COVID-19 patients versus those who were not, independent samples t-tests were used. Multiple linear regression was conducted to assess whether sociodemographic factors (age, sex, marital status, and years of work experience) and organizational factors (workplace support, organizational leadership, supervisory leadership, and ethical climate) were associated with total WBI scores. We conducted four additional multiple linear regression models to assess whether organizational factors moderated the relationship between COVID-19 contact and total WBI score. Organizational variables were separated to reduce potential multicollinearity between variables. For each of the four interaction models, we regressed total WBI scores on sociodemographic variables, contact with COVID-19 patients, the organizational variable, and the COVID-19 contact · organization variable interaction. All predictor variables were grand-mean centred. Statistical significance was set at an alpha level of 0.05 for all tests.

3. Results

3.1. Participant characteristics

Our sample consisted of 1029 participants across Canada (91.2% female; 8.8% male). A total of 608 (59.1%) participants were directly involved in the care of patients with elevated temperature or confirmed COVID-19, and 421 (40.9%) were not involved in the care of COVID-19 patients. The top five occupations among our analytical frame were nurse (45.6%), personal support worker (10.1%), physician (5.3%), paramedic (4.4%), and social worker (3.8%). Detailed sociodemographic characteristics of the participants are presented in Table 1.

Table 1. Participant characteristics.

Variables	n (%)
Age (years)
≤25	40 (3.9)
26–40	419 (40.7)
41–60	513 (49.9)
>60	57 (5.5)
Sex
Female	938 (91.2)
Male	91 (8.8)
Marital status
Married/common law	692 (67.2)
Single	211 (20.5)
Separated/divorced/widowed	126 (12.2)
Education
Less than college/university	127 (12.3)
Completed college/university	902 (87.7)
Number of years as a healthcare worker
≤5 years	200 (19.4)
6–10 years	214 (20.8)
≥11 years	615 (59.8)
Direct contact with COVID-19 patients
Yes	608 (59.1)
No	421 (40.9)
Primary job function
Direct client/patient care	896 (87.1)
Administration	79 (6.9)
Outreach	13 (1.1)
Research	13 (1.1)
Other	49 (4.3)

3.2. Descriptive statistics

Descriptive statistics, Cronbach’s alpha values, and bivariate correlations are reported in Table 2. Means and standard deviations of the WBI scale were comparable to previous studies using Canadian HCW samples (e.g. Plouffe et al., 2021). Bivariate correlations between study variables were all statistically significant (ps < .001), with effect sizes ranging in magnitude from small-to-large based on Cohen’s (1988) guidelines (Table 2). Among those who answered the WBI item reflecting perceptions of burnout, 698 (86.1%) of participants reported feeling burned out from their work. Nurses reported higher rates of burnout (89.5%), compared with 75.7% of physicians and 84.5% of all other health care professionals (χ² = 7.0, p = .03).

Table 2. Descriptive statistics and bivariate correlations.

Measures	Mean (SD)	α	Skewness	Kurtosis	1	2	3	4
1. WBI	3.72 (2.27)	.67	−.42	−.28
2. Workplace support	3.18 (0.85)	.83	−.06	−.24	−.47
3. Organizational leadership	3.02 (1.04)	.91	−.07	−.72	−.35	.64
4. Supervisory leadership	3.06 (1.21)	.95	−.13	−1.01	−.26	.57	.75
5. EEQ	2.98 (0.72)	.93	−.10	−.03	−.36	.58	.61	.54

Note. All correlations are significant at p < .001. WBI = Well-Being Index. EEQ = Ethical Environment Questionnaire. WBI α does not include items 8 and 9, which are used to add or subtract one point from participants’ initial scores (see WBI scoring information).

3.3. Mean differences in burnout and organizational variables by COVID-19 contact

Independent samples t-tests showed that mean scores on the WBI were significantly higher in HCWs working in direct contact with COVID-19 patients compared to those who were not (Table 3). Mean scores on the EEQ were also significantly lower in HCWs working with COVID-19 patients. There were no significant mean differences between groups on other organizational variables (Table 3).

Table 3. Scores for burnout and organizational variables in healthcare workers working or not working in contact with COVID-19 patients.

Variables	Total mean (SD)	COVID-19 contact mean (SD)	No COVID-19 contact mean (SD)	t(df)
WBI	3.72 (2.27)	4.04 (2.18)	3.27 (2.33)	−5.05(890)***
Workplace support	3.18 (0.85)	3.15 (0.82)	3.23 (0.88)	1.58(1143)
Organizational leadership	3.02 (1.04)	2.97 (1.04)	3.08 (1.05)	1.82(1137)
Supervisory leadership	3.06 (1.21)	3.03 (1.21)	3.10 (1.19)	1.06(1135)
EEQ	2.98 (0.72)	2.93 (0.70)	3.06 (0.73)	2.67(879)**

Note. WBI = Well-Being Index. EEQ = Ethical Environment Questionnaire. Total n = 1029; COVID-19 contact n = 608; no COVID-19 contact n = 421.

*p < .05; **p < .01; ***p < .001.

3.4. Predictors of burnout and interactions between COVID-19 contact and organizational variables

Multiple linear regression analyses were performed to identify sociodemographic and organizational predictors of burnout (Table 4). Overall, years of work experience, direct contact with COVID-19 patients, workplace support, organizational leadership, supervisory leadership, and ethical climate were significant predictors for burnout (F(13,787) = 21.83, R² = .25, p < .001). Having 5 or fewer years of work experience (β = 0.20, p = .038) and direct contact with COVID-19 patients (β = 0.25, p < .001) were associated with an increased likelihood of probable burnout (Table 4). Having stronger workplace support (β = −0.39, p < .001), stronger organizational leadership (β = −0.11, p = .027), stronger supervisory leadership (β = 0.09, p = .049), and a more ethical work environment (β = −0.09, p = .042) were negatively associated with likelihood of probable burnout (Table 4). It should be noted that despite supervisory leadership having a negative correlation with burnout as demonstrated in our preliminary analysis (Table 2), the regression coefficient for supervisory leadership was positive (Table 4). This was likely due to suppression effects considering the multicollinearity between the organizational variables.

Table 4. Burnout regressed on sociodemographic and organizational variables.

Variables	β	95% CI	p value
Age (years)
≤25	−0.16	−0.62, 0.30	.499
26–40	0.03	−0.27, 0.33	.840
41–60	0.01	−0.26, 0.29	.918
>60 (ref)
Sex ^a	0.17	−0.05, 0.38	.124
Marital status
Single	0.07	−0.09, 0.23	.376
Separated/divorced/widowed	−0.02	−0.21, 0.17	.807
Married/common law (ref)
Number of years as a healthcare worker
≤5 years	0.20	0.01, 0.39	.038
6–10 years	0.05	−0.12, 0.22	.536
≥11 years (ref)
Direct contact with COVID-19 patients ^b	0.25	0.12, 0.37	<.001
Workplace support	−0.39	−0.48, −0.31	<.001
Organizational leadership	−0.11	−0.22, −0.01	.027
Supervisory leadership	0.09	0.00, 0.18	.049
EEQ	−0.09	−0.17, 0.00	.042

Note. β = unstandardized coefficient. CI = confidence interval.

EEQ = Ethical Environment Questionnaire.

^a 0 = male; 1 = female.

^b 0 = no; 1 = yes.

We then conducted four additional linear regression models to assess whether workplace support, organizational leadership, supervisory leadership, and ethical climate moderated the relationships between COVID-19 contact and burnout. Across all models, direct contact with COVID-19 patients and organizational variables were significant predictors of burnout; however, none of the interaction terms were significant (Table 5). We also controlled for sociodemographic factors in each model (Tables S1–4, Supplemental Materials), which resulted in the same interpretation of findings for each organizational predictor variable.

Table 5. Burnout regressed on direct contact with COVID-19 patients, organizational variables, and COVID-19 contact × organizational variable interactions.

Variables	β	95% CI	p value
Model 1
Direct contact with COVID-19 patients ^a	0.27	0.14, 0.39	<.001
Workplace support	−0.49	−0.58, −0.40	<.001
COVID-19 contact × workplace support	0.05	−0.07, 0.17	.399
F(11,789) = 24.45, R^2 = 0.24, p < .001
Model 2
Direct contact with COVID-19 patients ^a	0.27	0.13, 0.40	<.001
Organizational leadership	−0.35	−0.45, −0.25	<.001
COVID-19 contact × organizational leadership	0.00	−0.13, 0.14	.971
F(11,789) = 13.40, R^2 = 0.15, p < .001
Model 3
Direct contact with COVID-19 patients ^a	0.29	0.16, 0.43	<.001
Supervisory leadership	−0.26	−0.37, −0.16	<.001
COVID-19 contact × supervisory leadership	0.01	−0.13, 0.15	.879
F(11,789) = 8.64, R^2 = 0.10, p < .001
Model 4
Direct contact with COVID-19 patients ^a	0.25	0.12, 0.38	<.001
EEQ	−0.28	−0.38, −0.18	<.001
COVID-19 contact × EEQ	−0.09	−0.22, 0.04	.188
F(11,789) = 12.87, R^2 = 0.14, p < .001

Note. β = unstandardized coefficient. CI = confidence interval.

EEQ = Ethical Environment Questionnaire.

^a 0 = no; 1 = yes.

4. Discussion

While previous research has documented risk and resilience factors associated with burnout, more research highlighting those considerations, particularly with national samples of HCWs, underscore the urgency and needs of HCWs as a systemic issue faced globally. Using a pan-Canadian sample, we highlighted the wide-spread impact of the pandemic on HCW wellbeing. We identified several factors associated with burnout, with a unique focus on how organizational variables may function as protective factors.

In our study, HCWs in direct contact with COVID-19 patients showed a higher likelihood of probable burnout than HCWs not working with COVID-19 patients. Our findings align with past research, which identified contact with COVID-19 patients as a risk factor for burnout among HCWs (Bailey et al., 2021; Evanoff et al., 2020; Sung et al., 2021). Previous research has shown that moral distress, defined as the emotional state that occurs when an individual is constrained to carry out an action that they perceive is morally or ethically inappropriate, is a predictor of burnout in HCWs (Plouffe et al., 2021). An example of moral distress in HCWs includes attending to patients without adequate PPE. Working with COVID-19 patients exposes HCWs to additional challenges, including increased risk of infection or passing infection to family members and friends. Frontline workers who are in contact with COVID-19 patients are also more likely to experience high workloads, be sleep deprived, and have disrupted social support during isolation or quarantine, which function to increase the risk of burnout (Sultana et al., 2020).

Our study also identified sociodemographic predictors of mental health outcomes among HCWs. Consistent with previous studies, we found that fewer years of work experience is associated with burnout. This is in line with another study which found that more years of work experience as a HCW was a protective factor against adverse mental health outcomes (Syamlan et al., 2022). HCWs with less experience may be more vulnerable to psychological distress due to feelings of incompetence when faced with critical patient cases. Research has shown that HCWs with prior public health emergency experience had lower rates of psychological conditions during the COVID-19 pandemic (Cai et al., 2020).

Contrary to previous studies, we found that sex was not significantly associated with burnout. Much of the existing literature has reported that being female is a risk factor of burnout (Evanoff et al., 2020; Jalili et al., 2021; Orru et al., 2021; Sung et al., 2021). Females may be likely to experience worsened mental health outcomes during COVID-19 due to a tendency to engage in internalizing behaviours (e.g. social withdrawal), the need to balance multiple duties at work and at home, and a potential role of sex hormones (Ayalew et al., 2021; Eaton et al., 2012; Li & Graham, 2017). Our results also indicate that age and marital status were not significant predictors of burnout. Past literature has demonstrated inconsistency in the impact of age on mental health outcomes (Ayalew et al., 2021; Elbay et al., 2020; Naser et al., 2020; Syamlan et al., 2022). Younger individuals may be protected because the COVID-19 virus posed a lower health threat than to their older colleagues (Kang & Jung, 2020), whereas older HCWs may be protected because of their experiences in facing critical events and the resilience and coping strategies they have developed throughout their lifetime. Being single and socially isolated have also been identified as important determinants of adverse psychological outcomes during current and past outbreaks (Aljabri et al., 2022; Kisely et al., 2020), but this did not align with our findings. Marriage could serve as a significant source of social support as individuals are able to talk about their experiences and challenges during work; however, it could simultaneously heighten concerns about potentially transmitting infections to household members. Even before the COVID-19 pandemic, the role of sex, age, and marital status were heterogeneous, with one systematic review outlining inconsistencies in their relationship with burnout (Bria et al., 2012).

All organizational variables in our study had a negative association with burnout. Similar to previous studies, having strong workplace support (Bailey et al., 2021), adequate leadership from organizational and supervisory management (Aljabri et al., 2022; Cyr et al., 2021; Evanoff et al., 2020; Fiabane et al., 2021; Pawlowicz-Szlarska et al., 2022), and a favourable ethical climate (Rivaz et al., 2020; Wlodarczyk & Lazarewicz, 2011) were protective factors against burnout. These factors are important in helping HCWs feel confident and safe in a high-pressure environment, maintaining trusting relationships with their leaders, and preventing physical and mental exhaustion. However, none of the organizational variables in our study moderated the association between contact with COVID-19 patients and burnout. This finding indicates that positive workplace support, strong organizational and supervisory leadership, and an ethical work environment alone may not have provided enough protection to HCWs against burnout from working with COVID-19 patients. The pandemic led to catastrophic failures in our healthcare system, and we were largely unprepared to take on the pandemic despite best efforts. For example, despite supervisors’ intention in providing their staff with PPE, there was simply inadequate supply to meet all HCWs’ needs.

Based on the findings from our study, we have several recommendations to reduce burnout in HCWs. First, psychological resources should be directed to HCWs working closely with COVID-19 patients or in settings at high risk of contact with COVID-19 such as respiratory, emergency, ICU and infectious disease departments. Second, more thorough training and guidance should be directed to HCWs with less experience to support them during difficult cases. In an over-taxed system, informal mentorship may be one of the first solutions to consider as it has led to increased confidence and job satisfaction in junior HCWs, which may ultimately increase their retention in the workforce (Esther et al., 2017). We also showed that organizational factors are important protective factors and propose more upstream solutions, such as supplying sufficient PPE, regulating work hours to prevent overexertion, and optimizing work efficiency through suitable job role distribution (Muller et al., 2020; Walton et al., 2020; West et al., 2018). In preparation for future pandemics, stocks of PPE should be maintained to prevent severe shortages like those faced during the COVID-19 pandemic. To prevent overworking, more personnel can be recruited including HCWs on leave, retired HCWs, medical and nursing students, and military personnel. Solutions may also require reaching outside of what has been offered thus far, highlighting a need for new initiatives and studies evaluating their effectiveness.

4.1. Limitations

Our study presents with some limitations. First, as our sampling was through an online self-selection snowball method, our sample may be limited to those who are willing to complete the survey, have internet access, or are technologically savvy. Second, females represented over 90% of our sample, whereas they constitute between 78% and 85% of HCWs in the Canadian HCW population (Shrma & Smith, 2021). Thus, the generalizability of our findings may be limited. Third, probable burnout in our sample was measured using the WBI, which does not easily allow us to compare our findings with other studies that used the Maslach Burnout Inventory (Bailey et al., 2021; Cyr et al., 2021; Fiabane et al., 2021; Shanafelt et al., 2022). We also questioned the validity of the PEPS scale considering its novelty, and therefore we suggest that our findings be interpreted with caution and that future research be done to evaluate its validity in assessing work-life and leadership experiences. Additionally, there may be other important predictors of burnout that we did not include in our analyses. Given the complexity of mental health, there are likely many factors outside of occupational and sociodemographic variables that are unaccounted for in our models. For example, we did not analyze other facets of mental health including depression, anxiety, PTSD, and distress, which likely all relate to burnout. As well, we are limited by our proxy of measurement for leadership and support. Our understanding of supervisory and organizational leadership has evolved since the pandemic, so questions may only tap into limited aspects of these constructs. For example, it may be worthwhile to break down support from leadership into instrumental and emotional support, as they have different implications for how HCWs’ well-being can be protected. Furthermore, there was no delineation made between leadership roles, which may potentially range from a nursing shift supervisor, for example, to the manager of the hospital. Finally, our cross-sectional study design does not allow the drawing of cause–effect relationships. A longitudinal assessment of HCWs may reveal whether the COVID-19 pandemic is responsible for declines in mental health.

5. Conclusion

Our study assessed burnout in Canadian HCWs during the COVID-19 pandemic, a time when frontline workers are particularly vulnerable to stressors and mental health deterioration. We demonstrated that HCWs working directly with COVID-19 patients were at increased risk of burnout. Years of work experience, workplace support, organizational and supervisory leadership, and ethical climate were significant determinants of burnout in our sample. Timely implementation of strategies to support individuals who fall under high-risk groups is critical to protect their mental well-being and ensure high-quality patient care.

Author contributions

NL, RP, JL, AN, and JR conceptualized the project. MN, PS, DG, and BD conducted the analyses. NL wrote the first draft of the manuscript. All authors contributed to manuscript revision, proofreading, and approved the submitted version.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Study data are not available on a repository to protect participant privacy. However, data are available from the corresponding author upon reasonable request.

Additional information

Funding

This project was funded through the support of the MacDonald Franklin OSI Research Centre by the St. Joseph’s Health Care Foundation (London, Ontario, Canada) and through a partnership with the Atlas Institute for Veterans and Families (Ottawa, Ontario, Canada).