Content area
Objectives
This study aimed to investigate the evolution of burnout levels and cardiovascular risk among healthcare professionals during the COVID-19 pandemic, identifying associated risk factors, with a particular focus on the impact of working hours, job roles and working units.
Design
A longitudinal, observational study was conducted.
Setting
The study was carried out in a medical centre in central Taiwan, encompassing various healthcare settings.
Participants
A total of 1502 healthcare workers participated, including nurses, medical technicians, resident doctors, attending physicians and administrative staff. Participants were selected based on consistent completion of a 4-year questionnaire, with exclusion criteria for those who did not complete.
Primary and secondary outcome measures
The primary outcome measured was burnout levels using the Chinese version of the Copenhagen Burnout Inventory. The secondary outcome was cardiovascular risk calculated from employees’ health check-up data using the Framingham Risk Score.
Results
Cardiovascular risk showed an upward trend over 4 years. Personal and work-related burnout significantly decreased from 2019 to 2020 but increased from 2020 to 2022, aligning with changes in weekly working hours. Nurses exhibited the most pronounced fluctuations, likely due to their younger average age, shorter professional tenure and frequent direct patient contact, which may heighten vulnerability to pandemic-related stressors. In contrast, attending physicians demonstrated age as a protective factor against burnout, as greater seniority, clinical experience and professional maturity may buffer stress and foster resilience. Participants who worked in COVID-related units generally had elevated burnout levels and working hours. During the initial outbreak in 2020, employees working in COVID-related units had reduced working hours but stable burnout levels, while employees in non-COVID-related units experienced decreased burnout.
Conclusions
This study highlights the critical impact of long working hours on burnout among healthcare professionals during the COVID-19 pandemic. Nurses emerged as a vulnerable group, sensitive to pandemic-induced changes, while attending physicians exhibited more resilience. COVID-related units face greater stress and are less likely to benefit from reductions in patient numbers and working hours during the pandemic. Our findings underscore the urgent need for tailored interventions, such as regulated work hours, flexible scheduling and enhanced organisational and peer support, to protect healthcare workers’ well-being. These strategies can strengthen workforce resilience and sustainability in future public health crises.
Full text
Correspondence to Dr Yu-Tse Tsan; [email protected]; Dr Wei-Min Chu; [email protected]
STRENGTHS AND LIMITATIONS OF THIS STUDY
Understand the longitudinal impact of working hours during the COVID-19 pandemic on burnout levels among healthcare professionals over 4 years.
Recognise the heightened vulnerability of nurses to burnout compared with other healthcare roles during the pandemic.
Identify the need for tailored support mechanisms to address burnout in healthcare workers, particularly in COVID-related units.
Introduction
The global COVID-19 pandemic from 2019 to 2022 has presented unprecedented challenges to healthcare systems, leading to significant physical and mental stress among healthcare workers. As the pandemic intensified, burnout prevalence among professionals surged.1 Risk factors repeatedly identified include long working hours, high work intensity, direct care of COVID-19 patients and disruption of personal or household activities.2 3
Few studies compared burnout at different time points during the COVID-19 pandemic. A systematic review and meta-analysis found alarmingly high rates of burnout among pulmonologists and respiratory therapists, with the prevalence spiking from 41.6% prepandemic to 68.4% during COVID-19 (p=0.01), likely due to increased workload and stress.4 Persistent burnout was further linked to depersonalisation, acute stress symptoms and adverse mental health outcomes, with epidemic-related stressors such as protective equipment discomfort and infection anxiety compounding the risks.5 6 Conversely, social and organisational support emerged as protective factors.6 This points out the importance of systemic interventions focused on reducing occupational stressors and enhancing support systems to safeguard the well-being of this essential workforce. Collectively, these findings reinforce the need for longitudinal research to capture dynamic burnout patterns over time and identify modifiable risk and protective factors.
Burnout among healthcare professionals may fluctuate in response to pandemic status and policies. Extended working hours, particularly beyond regular schedules, are consistently associated with physician burnout.7 However, during COVID-19, workforce shortages limited the feasibility of reducing working hours, further exacerbating risk.8 For example, in Australasia, nearly 75% of infectious disease physicians reported working the equivalent of at least one additional day per week, reflecting heightened vulnerability to burnout.9 Nurses have been consistently identified as a high-risk group for pandemic-related burnout,10 while age and experience appear protective.11–13 Frontline healthcare workers show higher rates of depression, anxiety, stress and post-traumatic stress than second-line workers.14 Building on this background, the present study specifically aims to track longitudinal changes in burnout among healthcare professionals from 2019 to 2022, and to test the hypothesis that job role, working hours and assignment to COVID-related units are key predictors of burnout during the pandemic.
However, existing studies primarily examine factors posing risks to healthcare professionals’ well-being during the COVID-19 pandemic.15–17 Most use cross-sectional designs, offering insights into immediate effects but lack longitudinal tracking of burnout over an extended period. Chu et al discovered that nursing staff, poor sleep quality, long working hours and lack of exercise were linked to higher levels of personal burnout, work-related burnout and depression among healthcare professionals.10 This is one of the few longitudinal studies on burnout during COVID-19. However, few studies have tracked the same cohort of healthcare workers across multiple years of the pandemic, limiting understanding of how burnout evolves under prolonged stress and shifting health policies. By conducting a 4-year longitudinal follow-up from 2019 to 2022, this study addresses this critical gap and offers novel insights into trends of personal burnout, work-related burnout and cardiovascular risk among healthcare professionals in a central Taiwan medical centre.
Methods
Data source and study group
Our study adopted a longitudinal approach to gather burnout data from healthcare professionals spanning 2019 to 2022, covering the entire COVID-19 pandemic period. It documents burnout status evolution, from prepandemic times through global and local outbreaks, providing comprehensive insights into healthcare professionals’ burnout throughout the pandemic.
Data for this study were sourced from the Overload Health Control System at a central Taiwan tertiary medical centre. Inclusion criteria encompassed all hospital staff who completed the online questionnaire across the study period, which collected information on job roles, working conditions, depressive symptoms, health check-up results and burnout levels. Exclusion criteria included participants who did not consistently complete the questionnaire across all 4 years (figure 1). Based on the number of participants who met the inclusion criteria and were retained across all analyses, the estimated study sample was 758. This sample size was determined according to OR-based calculations for burnout and cardiovascular risk, ensuring sufficient statistical power. To address concerns about selection bias, we compared the demographics and baseline characteristics of excluded participants with those retained in the final sample, and the comparison is presented in the supplementary material (online supplemental material 1). The aim was to investigate the COVID-19 pandemic’s influence on individuals and identify associated risk factors for burnout and mental stress. We explored correlations between burnout levels, working hours, COVID-related units, job titles and age using hospital-provided data.
Figure 1. Flow diagram of the enrolment of participants who completed the questionnaire for four consecutive years.
Independent variables
In 2019, Taiwan’s Ministry of Labour’s Occupational Safety and Health Administration introduced the Guideline for Prevention of Diseases Arising from Excessive Workload. It mandates an overwork assessment questionnaire covering sociodemographics (gender, age), working conditions (seniority, working hours, job titles, work units) and lifestyle factors (smoking, alcohol consumption, sleeping quality, exercising habits). Questionnaire items were selected by consensus of experts from the ministry. In the current study, we further incorporated these lifestyle factors into our analytical model to account for potential confounding variables related to personal life stressors, thereby strengthening the validity of our findings. COVID-related units were defined as participants working in the emergency department, medical intensive care units and negative-pressure isolation wards.
Dependent variables
We assessed employees’ burnout level using the Chinese version of the Copenhagen Burnout Inventory (online supplemental material 2)18 developed by Taiwan’s Ministry of Labour. It comprised six questions for personal burnout and seven for work-related burnout, each scored from 0 to 100. Scores of 0 to 49, 50 to 70 and >70 indicated mild, moderate and severe levels of personal burnout, and scores of 0 to 44, 45 to 60 and >60 indicated mild, moderate and severe levels of work-related burnout, respectively. Cardiovascular risk is the result calculated from employees’ health check-up data using the Framingham Risk Score.19 The rationale for including this measure is that the Framingham Risk Score is the standard tool recommended by Taiwan’s Occupational Safety and Health Administration in its official Guideline for Prevention of Diseases Arising from Excessive Workload20.
Statistical analysis
This study analysed the differences in cardiovascular disease risk, personal burnout, work-related burnout and weekly working hours over 4 years using the Friedman test. It further employed Spearman correlation to analyse the relationship between weekly working hours and cardiovascular disease risk, personal burnout and work-related burnout. Lastly, the generalised estimating equation (GEE) was utilised to explore the relationships among personal burnout, work-related burnout and various variables. We applied GEE to analyse longitudinal data, adopting an exchangeable working correlation structure. Variable selection was guided by clinical relevance and supported by prior literature. Regarding missing data, GEE treated missing observations as absent at the corresponding time points while retaining other observations from the same participant; this approach assumes data were missing at random, and no additional imputation was performed. Results were presented as effect sizes with 95% CIs rather than p values alone. In GEE, regression coefficients (estimates) represent effect sizes (eg, mean differences or log odds), SEs indicate the precision of estimates and 95% CIs were calculated as ‘estimate ± 1.96 × SE.’ The statistical analysis was performed using SAS 9.4 statistical software, and p values less than 0.05 were considered statistically significant.
Patient and public involvement statement
This study was designed in collaboration with experts from the Occupational Safety and Health Office of Taichung Veterans General Hospital. Input from hospital employees, who shared their challenges and concerns during the COVID-19 pandemic, was also incorporated. Their feedback helped refine the study focus and ensure its relevance to real workplace conditions.
Results
Over 4 years, from 2019 to 2022, 1502 participants completed the questionnaire every year. After excluding those with abnormal working hours (eg, more than 168 hours per week, likely due to data entry errors), a total of 1486 participants were included in the analysis. They included 358 administrative staff, 35 resident doctors, 81 visiting doctors, 675 nurses and 337 medical technicians. Demographic details are outlined in table 1. In 2019, participants averaged 43.19 years in age, with 16.01 years of seniority and a body mass index of 23.38. Average weekly working hours for all participants were 44.32.
Table 1
Demographics and characteristics of the participants from 2019 to 2022 (n=1486)
| N (%) | |
| Gender | |
| 303 (20.39%) | |
| 1183 (79.61%) | |
| Age | |
| 1077 (72.48%) | |
| 409 (27.52%) | |
| Seniority | |
| 544 (36.61%) | |
| 942 (63.39%) | |
| Weekly working hours | |
| 55 (3.7%) | |
| 1146 (77.12%) | |
| 208 (14%) | |
| 77 (5.18%) | |
| Job title | |
| 358 (24.09%) | |
| 81 (5.45%) | |
| 35 (2.36%) | |
| 675 (45.42%) | |
| 337 (22.68%) | |
| Unit | |
| 215 (14.47%) | |
| 1271 (85.53%) | |
| BMI | |
| 87 (5.85%) | |
| 910 (61.24%) | |
| 489 (32.91%) | |
| Waistline | |
| 1253 (84.32%) | |
| 233 (15.68%) | |
| Lifestyle | |
| 1474 (99.19%) | |
| 12 (0.81%) | |
| 1441 (96.97%) | |
| 45 (3.03%) | |
| 791 (53.23%) | |
| 695 (46.77%) | |
| 856 (57.60%) | |
| 630 (42.40%) | |
| 119 (8.01%) | |
| 449 (30.22%) | |
| 502 (33.78%) | |
| 416 (27.99%) | |
| Shift | |
| 1055 (71%) | |
| 142 (9.56%) | |
| 289 (19.45%) |
| Mean±SD | Median (IQR) | |
| Age | 43.19±10.17 | 42 (17) |
| Seniority | 16.01±10.71 | 13 (18) |
| BMI | 23.38±4.15 | 22.6 (5.2) |
| Waistline | 73.77±14.61 | 74 (13.55) |
| Daily working hours | 8.89±3.01 | 8 (1) |
| Weekly working hours | 44.32±7.04 | 40 (5) |
| CVD risk | 2.66±2.79 | 1 (2) |
| Personal burnout | 42.92±19.62 | 41.7 (25) |
| Work-related burnout | 42.11±17.29 | 42.9 (17.9) |
A, administrative staff; BMI, body mass index; CVD, cardiovascular disease; M, medical technician; N, nurse; R, resident doctor; V, visiting staff.
Table 2 illustrates changes in cardiovascular risk, personal burnout, work-related burnout and weekly working hours from 2019 to 2022 among all participants. Cardiovascular risk shows an upward trend over 4 years (M = +0.57). Personal and work-related burnout notably decreased from 2019 to 2020 (M=−1.93; M=−1.16) but then significantly increased from 2020 to 2022 (M = +0.92; M = +0.96). Weekly working hours vary correspondingly, with a significant decrease from 2019 to 2020 (M=−0.78) and a notable increase from 2020 to 2021 (M = +0.34).
Table 2
CVD risk, burnout levels and weekly working hours from 2019 to 2022
| 2019 | 2020 | 2021 | 2022 | 2022–2019 | P value | |
| CVD risk | <0.0001*** | |||||
| 2.36 | 2.58 | 2.77 | 2.93 | 0.57 | ||
| 2.37 | 2.74 | 2.94 | 3.03 | 2.14 | ||
| Personal burnout | <0.0001*** | |||||
| 44 | 42.07 | 42.63 | 42.99 | −1.01 | ||
| 19.4 | 19.39 | 19.81 | 19.84 | 17.35 | ||
| Work-related burnout | <0.0001*** | |||||
| 42.6 | 41.44 | 42 | 42.4 | −0.2 | ||
| 17.05 | 17.17 | 17.52 | 17.42 | 15.63 | ||
| Weekly working hours | <0.0001*** | |||||
| 45.04 | 44.26 | 44.6 | 44.32 | −0.73 | ||
| 7.99 | 7.35 | 7.54 | 7.04 | 7.15 |
Friedman test was used for statistical examination.
P value: *p < 0.05, **p < 0.01, ***p < 0.001.
CVD, cardiovascular disease.
Table 3 displays variations in working hours, cardiovascular risk and burnout levels across these categories. Resident doctors showed no significant differences in trends over 4 years. Attending physicians had significant changes in cardiovascular risk and working hours (p=0.0005; p=0.0145), with minimal fluctuations in personal and work-related burnout (p=0.4353; p=0.4721). Nurses exhibited significant trends in all measured aspects over the 4-year period.
Table 3
Differences in CVD risk, burnout levels and weekly working hours between job categories from 2019 to 2022
| 2019 | 2020 | 2021 | 2022 | 2022–2019 | P value | |
| A | ||||||
| 0.0013** | ||||||
| 3.18 | 3.64 | 3.77 | 3.84 | 0.66 | ||
| 3.06 | 3.47 | 3.77 | 3.64 | 2.67 | ||
| 0.0396* | ||||||
| 40.61 | 39.91 | 38.5 | 39.12 | −1.49 | ||
| 19 | 18.48 | 20.12 | 20.22 | 17.72 | ||
| 0.0568 | ||||||
| 38.86 | 38.82 | 38.77 | 38.6 | −0.26 | ||
| 17.29 | 17.26 | 18.48 | 17.99 | 16.17 | ||
| <0.0001*** | ||||||
| 42.49 | 42.33 | 42.44 | 42.44 | −0.04 | ||
| 5.09 | 5.69 | 5.62 | 5.66 | 4.71 | ||
| V | ||||||
| 0.0005*** | ||||||
| 4.65 | 4.68 | 4.81 | 4.85 | 0.2 | ||
| 2.93 | 2.93 | 3.18 | 3.8 | 2.6 | ||
| 0.4353 | ||||||
| 36.11 | 32.15 | 34.26 | 34.52 | −1.59 | ||
| 18.02 | 16.94 | 19.04 | 17.58 | 15.13 | ||
| 0.4721 | ||||||
| 36.69 | 32.72 | 33.69 | 33.86 | −2.83 | ||
| 16.63 | 13.79 | 14.98 | 14.46 | 14.19 | ||
| 0.0145* | ||||||
| 54.81 | 51.73 | 52.95 | 53.16 | −1.91 | ||
| 12.35 | 12.91 | 11.44 | 11.23 | 10.29 | ||
| R | ||||||
| 0.5161 | ||||||
| 2.06 | 2.11 | 2.17 | 2.43 | 0.37 | ||
| 1.08 | 1.28 | 1.25 | 1.38 | 0.77 | ||
| 0.1088 | ||||||
| 44.52 | 42.38 | 41.31 | 42.38 | −2.14 | ||
| 21.11 | 21.55 | 16.46 | 20.15 | 17.86 | ||
| 0.3173 | ||||||
| 44.79 | 42.75 | 41.22 | 43.05 | −1.74 | ||
| 17.92 | 19.63 | 12.68 | 17.26 | 15.09 | ||
| 0.1154 | ||||||
| 64.82 | 57.63 | 55.5 | 54.97 | −9.24 | ||
| 16.19 | 14.31 | 14.99 | 13.64 | 20.13 | ||
| N | ||||||
| <0.0001*** | ||||||
| 1.54 | 1.72 | 1.88 | 2.14 | 0.6 | ||
| 1.25 | 1.9 | 1.86 | 2.26 | 1.84 | ||
| 0.0085** | ||||||
| 46.83 | 44.3 | 45.39 | 46.09 | −0.74 | ||
| 20.01 | 20.26 | 19.66 | 19.82 | 17.7 | ||
| 0.0019** | ||||||
| 45.33 | 43.73 | 44.71 | 45.49 | 0.16 | ||
| 16.84 | 17.28 | 16.96 | 17.1 | 15.87 | ||
| <0.0001*** | ||||||
| 45.44 | 44.83 | 45.21 | 44.68 | −0.74 | ||
| 6.69 | 6.41 | 6.97 | 6.14 | 7.22 | ||
| M | ||||||
| 0.0155* | ||||||
| 2.6 | 2.74 | 3.07 | 3.12 | 0.52 | ||
| 2.5 | 2.69 | 3.14 | 3.1 | 2.03 | ||
| 0.0019** | ||||||
| 43.78 | 42.23 | 43.65 | 42.99 | −0.79 | ||
| 17.64 | 17.93 | 19.07 | 18.69 | 16.74 | ||
| 0.0404* | ||||||
| 42.31 | 41.58 | 42.1 | 42.24 | −0.07 | ||
| 16.16 | 16.32 | 17.43 | 16.71 | 14.94 | ||
| <0.0001*** | ||||||
| 42.48 | 42.01 | 42.54 | 42.33 | −0.25 | ||
| 4.79 | 4.31 | 5.23 | 4.58 | 4.68 |
Friedman test was used for statistical examination.
P value: *p < 0.05, **p < 0.01, ***p < 0.001.
A, administrative staff; CVD, cardiovascular disease; M, medical technician; N, nurse; R, resident doctor; V, visiting staff.
We aimed to compare work conditions, cardiovascular risk and burnout level between COVID-related and non-COVID-related units during the pandemic. table 4 shows changes in working hours, cardiovascular risk, personal burnout and work-related burnout for both unit types over 4 years. In 2020, working hours significantly decreased in COVID-related units (M=−0.61), but burnout levels did not correspondingly decrease. Conversely, non-COVID-related units had stable working hour cardiovascular diseases but saw significant decreases in both burnout levels (M = +2.1; M = +1.34). In 2021 and 2022, burnout levels and working hours remained stable in both unit types, with no significant changes.
Table 4
Differences in CVD risk, burnout levels and weekly working hours between different units from 2019 to 2022
| 2019 | 2020 | 2021 | 2022 | 2022–2019 | P value | |
| COVID-related unit | ||||||
| CVD risk | 0.4093 | |||||
| Mean | 1.52 | 1.61 | 1.77 | 2.15 | 0.63 | |
| SD | 1.21 | 1.58 | 1.71 | 2.19 | 1.54 | |
| Personal burnout | 0.7397 | |||||
| Mean | 44.79 | 43.86 | 43.68 | 44.21 | −0.58 | |
| SD | 19.51 | 19.62 | 20.34 | 20.96 | 17.34 | |
| Work-related burnout | 0.0491* | |||||
| Mean | 43.9 | 43.77 | 43.5 | 44.52 | 0.62 | |
| SD | 16.96 | 17.17 | 17.31 | 18.31 | 15.23 | |
| Weekly working hours | 0.0013** | |||||
| Mean | 45.74 | 45.13 | 45.49 | 44.97 | −0.76 | |
| SD | 7.11 | 6.66 | 7.31 | 6.19 | 7.65 | |
| Non-COVID-related unit | ||||||
| CVD risk | <0.0001*** | |||||
| Mean | 2.5 | 2.75 | 2.94 | 3.06 | 0.56 | |
| SD | 2.49 | 2.86 | 3.07 | 3.14 | 2.22 | |
| Personal burnout | <0.0001*** | |||||
| Mean | 43.87 | 41.77 | 42.45 | 42.78 | −1.09 | |
| SD | 19.39 | 19.34 | 19.72 | 19.64 | 17.36 | |
| Work-related burnout | <0.0001*** | |||||
| Mean | 42.38 | 41.04 | 41.75 | 42.04 | −0.34 | |
| SD | 17.06 | 17.14 | 17.55 | 17.25 | 15.7 | |
| Weekly working hours | <0.0001*** | |||||
| Mean | 44.92 | 44.12 | 44.45 | 44.22 | −0.73 | |
| SD | 8.12 | 7.45 | 7.57 | 7.17 | 7.04 |
Friedman test was used for statistical examination.
P value: *p < 0.05, **p < 0.01, ***p < 0.001.
CVD, cardiovascular disease.
Table 5 shows the results of multivariable analysis using GEE for personal and work-related burnout. Factors like age over 50, being an attending physician and exercising are protective, while being a medical technician, poor sleep and longer working hours are risk factors for both types of burnout.
Table 5
The correlation between multiple factors and burnout levels
| Personal burnout | Work-related burnout | |||||
| Estimate | 95% CI | P value | Estimate | 95% CI | P value | |
| Gender | ||||||
| REF | REF | |||||
| −1.2818 | (−4.6846, 2.1209) | 0.4603 | −0.846 | (−3.8692, 2.1773) | 0.5834 | |
| Age | ||||||
| REF | REF | |||||
| −3.112 | (−5.8738 to 0.3502) | 0.0272* | −3.4197 | (−5.8502 to 0.9893) | 0.0058** | |
| Seniority | ||||||
| REF | REF | |||||
| 2.2124 | (−0.6837, 5.1084) | 0.1343 | 1.4736 | (−1.0256, 3.9728) | 0.2478 | |
| BMI | ||||||
| 2.1115 | (−3.3598, 7.5828) | 0.4494 | −0.0372 | (−4.2379, 4.1635) | 0.9862 | |
| REF | REF | |||||
| −1.8586 | (−4.7798, 1.0627) | 0.2124 | −2.085 | (−4.6450, 0.4749) | 0.1104 | |
| Waistline | ||||||
| REF | REF | |||||
| 3.8172 | (0.4114, 7.2230) | 0.0280* | 3.1393 | (0.1264, 6.1521) | 0.0411* | |
| Job title | ||||||
| REF | REF | |||||
| −11.5522 | (-16.4493, to 6.6552) | <0.0001*** | −8.2373 | (-12.6664, to 3.8082) | 0.0003*** | |
| 3.0442 | (−22.5496, 28.6380) | 0.8157 | 4.0666 | (−18.9408, 27.0741) | 0.729 | |
| 2.5927 | (−0.9189, 6.1044) | 0.1479 | 1.6348 | (−1.3972, 4.6667) | 0.2906 | |
| 2.667 | (−0.6446, 5.9785) | 0.1145 | 2.8691 | (−0.0853, 5.8235) | 0.057 | |
| Unit | ||||||
| REF | REF | |||||
| −1.3369 | (−4.6477, 1.9738) | 0.4287 | −0.9456 | (−3.7942, 1.9031) | 0.5153 | |
| Lifestyle | ||||||
| REF | REF | |||||
| 7.479 | (−5.3803, 20.3384) | 0.2543 | 2.4514 | (−5.6354, 10.5383) | 0.5524 | |
| REF | REF | |||||
| 4.3513 | (−2.7628, 11.4653) | 0.2306 | 2.4435 | (−3.1959, 8.0830) | 0.3957 | |
| REF | REF | |||||
| 14.5447 | (12.2941, 16.7952) | <0.0001*** | 12.1623 | (10.2042, 14.1205) | <0.0001*** | |
| REF | REF | |||||
| −1.8054 | (−4.1054, 0.4946) | 0.1239 | −1.9525 | (−3.9247, 0.0197) | 0.0523 | |
| Shift | ||||||
| REF | REF | |||||
| −2.6716 | (−6.5061, 1.1628) | 0.1721 | 0.6953 | (−3.1146, 4.5052) | 0.7206 | |
| 0.977 | (−2.4004, 4.3544) | 0.5707 | 2.2618 | (−0.5200, 5.0436) | 0.111 | |
| Weekly working hours | ||||||
| 40–49 | REF | REF | ||||
| 50–59 | 8.4383 | (5.4135, 11.4630) | <0.0001*** | 8.1962 | (5.5902, 10.8022) | <0.0001*** |
| 60–69 | 12.8481 | (6.1008, 19.5954) | 0.0002** | 12.5151 | (7.2016, 17.8287) | <0.0001*** |
| ≥70 | 6.8557 | (−0.7844, 14.4958) | 0.0786 | 4.8453 | (−1.8855, 11.5762) | 0.1583 |
Generalised estimating equation (multivariate) was used for statistical examination.
P value: *p < 0.05, **p < 0.01, ***p < 0.001.
A, administrative staff; M, medical technician; N, nurse; R, resident doctor; V, visiting staff.
Discussion
Our study is the first-ever research to complete a continuous 4-year follow-up on burnout within the same subjects, covering every stage of the COVID-19 pandemic. This longitudinal study extends our team’s prior 2 year observational study published in 2022.10 Throughout this period, various factors including demographics (age, gender, education, financial status), social aspects (stigmatisation, family life), psychological conditions (stress, anxiety, depression), COVID-related factors (patient contact, being infected) and work organisation (workload, staffing, support) were found to influence burnout.21 Through continuous tracking of study participants, we aim to understand the long-term impact of the pandemic on burnout and to identify vulnerable groups during the pandemic to provide timely support and protection in similar future events.
Our study results partially align with prior research on burnout among healthcare professionals during COVID-19. A previous study showed increased job demands, workload, complexity, pressure and working hours elevate stress among nurses.22 Another study indicates that individuals who just commenced their professional careers during the pandemic experience higher stress levels.23 A multinational study finds clinical roles and redeployment to new areas increase burnout risk.24 Most studies are cross-sectional or short-term, spanning only a few months.25–28 There’s a notable absence of research on the long-term impact of COVID-19 on healthcare professionals’ burnout. This is crucial as pandemic severity and risk factors for burnout may change over time due to evolving circumstances and policies. We found a strong correlation between healthcare professionals’ overall working hours and the 4-year trend of burnout levels, underscoring the critical role of working hours in pandemic-related burnout. Subgroup analysis highlighted nurses as the most sensitive group, while attending physicians seemed less affected by fluctuations in working hours. Resident doctors, being in an intermediate role and specific age range, show the most stable cardiovascular risk, burnout levels and working hours. In COVID-related units, burnout levels remained consistently higher, despite fluctuations in patient visits during the pandemic’s onset. Non-COVID-related units had reduced burnout levels, which may be due to slight decreases in working hours amidst the pandemic.
Longer working hours are strongly correlated with job burnout, especially when there is a lack of time off for rest and rejuvenation.29 30 Our research shows that trends in personal and work-related burnout among healthcare professionals over 4 years align with changes in working hours. A multicentre prospective study conducted in Singapore in 2020 found that 40% of physicians and 43% of nurses worked longer hours than usual during the pandemic.30 Another study published in 2023 revealed that in India, the shortage of healthcare professionals in 2020 resulted in extended working hours for healthcare workers, leading to burnout among them.31 However, despite the WHO’s official declaration of the global COVID-19 pandemic on March 11, 2020, Taiwan promptly implemented rigorous epidemic control measures.32 33 As a result, Taiwan effectively maintained a low number of COVID-19 patients in 2020.34 Multiple studies across countries indicate a decrease in emergency department visits during the COVID-19 pandemic.35–37 Even though Taiwan has low COVID-19 cases, there’s a consistent trend of reduced emergency department visits, like countries with higher incidence.38 We suggest that the 2020 decrease in healthcare professionals’ working hours may be linked to reduced patient visits in outpatient, emergency, inpatient and surgical settings. This phenomenon may reflect changes in burnout among frontline medical staff. Importantly, cardiovascular risk showed a significant mean increase of +0.57 in the Framingham Risk Score over 4 years. This progressive rise is partly expected, as age is an inherent component of the Framingham calculation.39 Nevertheless, hospitals should proactively address the cardiovascular risk associated with workforce ageing, in addition to managing occupational stressors. Targeted health check-up programmes and preventive strategies should be integrated into employee health policies to mitigate both burnout and cardiovascular risk among healthcare professionals.
In 2021, Taiwan imported a significant quantity of COVID-19 vaccines from March to November, launching an extensive vaccination campaign.40 By the end of 2021, the first-dose vaccine coverage had reached 78.5% according to statistical data from the Taiwan Centres for Disease Control.41 The high demand for vaccine administration increased healthcare professionals’ working hours, contributing to higher burnout levels.42 Prior research showed that physicians involved in mass vaccination sites had the highest self-perceived burnout syndrome prevalence. In a study by Hrehova et al, only 52% expressed interest in future participation in mass vaccination sites.43 Consequently, in 2021, healthcare workers saw an upsurge in workload and extended hours, attributed to the spike in domestic COVID-19 cases and the demand for widespread vaccination. This resulted in an increasing trend in both personal and work-related burnout. Future study is warranted to explore the relationship with mass vaccination and burnout level among healthcare professionals in Taiwan.
Literature reviews and our prior research consistently emphasise that nurses face a relatively high risk of burnout compared with other healthcare professionals.10 44 Recent studies have confirmed this vulnerability, identifying younger age, high workload, extended hours in quarantine or isolation units, direct and frequent patient contact, inadequate resources, reduced social support and insufficient training as key drivers of burnout during the COVID-19 pandemic.22 45 In our study, nurses demonstrated the highest burnout levels and the greatest sensitivity to pandemic-related stressors, with marked fluctuations in working hours, cardiovascular risk, personal burnout and work-related burnout across the 4-year period. Additionally, nurses, on average, had lower age and seniority compared with physicians, and prior literature revealed that seniority was associated with burnout and intention to leave.11 Past study shows younger, less experienced nurses often experience higher fatigue and burnout levels.12 The 4-year trend in attending physicians supports our inference about age and seniority. A 2020 study identified being under 40 and having less than 10 years of professional seniority as common causes of burnout.13 Attending physicians, with the highest average age and seniority among occupational categories, showed no significant changes in both burnout levels despite fluctuations in working hours, indicating lower sensitivity to pandemic-induced working condition changes. Nurses are more of a frontline healthcare worker, and attending physicians primarily handle major medical decisions.46 47 Being a relatively intermediate role as a healthcare professional, resident doctors experience less pressure and maintain stable outcomes in four measurements.
Our results showed that frontline healthcare workers in COVID-related units, facing higher stress and burnout risks, align with prior studies.48 49 Our research shows these units generally have elevated levels of personal and work-related burnout and long working hours, except for lower cardiovascular risk attributed to the younger age of healthcare workers in our study. Notably, during the initial outbreak in 2020, employees working in COVID-related units experienced a significant decrease in working hours, with no significant changes in personal or work-related burnout. In contrast, employees working in non-COVID-related units showed no significant variation in working hours but experienced a significant decrease in both personal and work-related burnout in the same year. After reviewing the literature, we found a significant decrease in emergency department visits worldwide after the COVID-19 outbreak.35–37 As the emergency department falls under the COVID-related unit category, this decline may have led to reduced working hours. However, frontline healthcare workers, especially during the initial outbreak, experienced increased burnout risk and psychological stress.48 49 We hypothesise that being in a COVID-related unit increases burnout, while the protective effect of working hour reduction countered it, resulting in stable burnout levels. There was a stable working hour but improved burnout due to slight decreases in working hours in non-COVID-related units. People who worked there experienced less psychological stress compared with COVID-related units, leading to reduced burnout levels during the initial outbreak. Future study is needed to explore the balance between working atmosphere and working hours and their effect on burnout level.
Finally, we used GEE for multivariable analysis on personal and work-related burnout. Consistent with previous findings, poor sleep, lack of exercise and long working hours remained risk factors.10 Additionally, being over 50 and an attending physician emerged as protective factors against burnout, confirming our earlier observations in table 3. Being a medical technician has also emerged as a high-risk group for burnout. However, relevant studies on burnout among medical technicians during the pandemic were lacking. We speculate that this may be related to the surge in testing during the initial outbreak in Taiwan when home rapid test kits were not yet available, leading to a large influx of specimens for testing in medical laboratories and consequently increasing the risk of burnout among medical technicians. Moving forward, we aim to conduct more detailed research to explore the underlying reasons for this risk.
However, there are still limitations in our study. First, working hour data are self-reported by healthcare workers, leading to potential discrepancies due to individual interpretations. Additionally, there may be selection bias as participants were continuously employed and filled out the questionnaire for 4 years, potentially limiting the assessment of healthcare workers who experienced burnout and left their positions during the pandemic. The predominance of female participants in the sample may also introduce bias and affect the generalisability of the findings. Being conducted at a single institution, caution is also needed when generalising findings to other settings. Our findings should be interpreted in the context of Taiwan’s pandemic response, which featured rapid border control, strict prevention measures and early mass vaccination. These factors limited case numbers and may have mitigated healthcare worker burnout, so the results may not generalise to countries with higher infection burdens, repeated surges or different healthcare staffing conditions. In addition, to address concerns about selection bias, we compared the demographics and baseline characteristics of excluded participants with those retained in the final analysis (online supplemental material 1). Significant differences were observed between the two groups in variables such as gender, age, seniority, weekly working hours, job title, work unit, body mass index, waistline, frequency of eating out, shift work, cardiovascular risk and both personal and work-related burnout levels. Moreover, we found that the excluded group included a higher proportion of younger and less experienced staff, and they also exhibited higher levels of personal burnout and work-related burnout compared with those who remained in the study. These findings suggest that individuals excluded due to attrition over the 4-year period differed substantially from the final analysed cohort. Therefore, future research should focus on understanding the unique risk factors among healthcare workers who left during the pandemic, as this vulnerable group may require additional protection to prevent workforce loss and ensure sufficient staffing during public health crises. Moreover, adopting prospective study designs with strategies to minimise attrition and loss to follow-up would be valuable for generating more representative longitudinal data and reducing potential bias in future investigations.
Understanding burnout trends during the pandemic is crucial for grasping its long-term impact on healthcare professionals and predicting challenges in future outbreaks. Our research identifies occupational categories most vulnerable to pandemic effects and offers insights into major burnout risk factors. Based on these findings, we recommend implementing tailored interventions such as work-hour regulation and flexible scheduling, which have been shown to mitigate burnout risk.7 In addition, enhancing organisational and peer support systems can buffer psychological stress, particularly for frontline workers in high-risk units.50 These strategies can help governments and healthcare institutions strengthen resilience, safeguard medical professionals’ well-being and ensure workforce sustainability during future public health crises. Future studies could further explore how individual traits, temperaments and pre-existing mental health conditions influence vulnerability to burnout. Incorporating these psychological dimensions may provide a more comprehensive understanding of job satisfaction and resilience among healthcare professionals. In addition, future research should also consider organisational and psychosocial factors such as staffing ratios, systemic support and baseline burnout levels to address potential residual confounding. Longitudinal studies across multiple institutions and investigations of targeted interventions to mitigate burnout could provide valuable insights into effective strategies for supporting healthcare workers and enhancing workforce sustainability during public health crises.
Conclusion
Our study delves into the escalating burnout levels among healthcare professionals during the COVID-19 pandemic, emphasising the critical impact of long working hours on burnout. By incorporating objective measurements and continuous tracking over 4 years, our research highlights the dynamic nature of burnout and its correlation with workload fluctuations. Nurses emerge as a vulnerable group, sensitive to pandemic-induced changes, while attending physicians exhibit more resilience. COVID-related units are less likely to benefit from reductions in patient numbers and working hours during the pandemic, as they endure greater psychological and emotional stress. These findings underscore the urgent need for tailored support mechanisms to safeguard the well-being of healthcare workers in possible future pandemic outbreaks.
Correction notice This article has been updated since it was first published, to reflect change in author affiliations.
Contributors S-YL: Conceptualisation, writing—original draft and visualisation. W-TH: Formal analysis, data curation and visualisation. Y-TT and W-MC: Conceptualisation, supervision and writing—review and editing. W-CH, Y-LL, C-FL, C-LW, H-EH and P-KF: Review and editing. Y-TT and W-MC: Project administration. Y-TT is the guarantor.
Funding This work was supported by the Occupational Safety and Health Office of Taichung Veterans General Hospital, Taiwan.
Competing interests None declared.
Patient and public involvement Patients and/or the public were involved in the design, conduct, reporting or dissemination plans of this research. Refer to the Methods section for further details.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
1 Innstrand ST. Burnout among Health Care Professionals during COVID-19. Int J Environ Res Public Health 2022;19:11807. doi:10.3390/ijerph191811807
2 Liu X, Chen J, Wang D, et al. COVID-19 Outbreak Can Change the Job Burnout in Health Care Professionals. Front Psychiatry 2020;11:563781. doi:10.3389/fpsyt.2020.563781
3 Morgantini LA, Naha U, Wang H, et al. Factors contributing to healthcare professional burnout during the COVID-19 pandemic: A rapid turnaround global survey. PLoS One 2020;15:e0238217. doi:10.1371/journal.pone.0238217
4 Bai X, Wan Z, Tang J, et al. The prevalence of burnout among pulmonologists or respiratory therapists pre- and post-COVID-19: a systematic review and meta-analysis. Ann Med 2023;55:2234392. doi:10.1080/07853890.2023.2234392
5 Miguel-Puga JA, Cooper-Bribiesca D, Avelar-Garnica FJ, et al. Burnout, depersonalization, and anxiety contribute to post-traumatic stress in frontline health workers at COVID-19 patient care, a follow-up study. Brain Behav 2021;11:e02007. doi:10.1002/brb3.2007
6 Zhou T, Xu C, Wang C, et al. Burnout and well-being of healthcare workers in the post-pandemic period of COVID-19: a perspective from the job demands-resources model. BMC Health Serv Res 2022;22:284. doi:10.1186/s12913-022-07608-z
7 West CP, Dyrbye LN, Shanafelt TD. Physician burnout: contributors, consequences and solutions. J Intern Med 2018;283:516–29. doi:10.1111/joim.12752
8 Lin R-T, Lin Y-T, Hsia Y-F, et al. Long working hours and burnout in health care workers: Non-linear dose-response relationship and the effect mediated by sleeping hours-A cross-sectional study. J Occup Health 2021;63:e12228. doi:10.1002/1348-9585.12228
9 Foley DA, Chew R, Raby E, et al. COVID-19 in the pre-pandemic period: a survey of the time commitment and perceptions of infectious diseases physicians in Australia and New Zealand. Intern Med J 2020;50:924–30. doi:10.1111/imj.14941
10 Chu W-M, Ho H-E, Lin Y-L, et al. Risk Factors Surrounding an Increase in Burnout and Depression Among Health Care Professionals in Taiwan During the COVID-19 Pandemic. J Am Med Dir Assoc 2023;24:164–70. doi:10.1016/j.jamda.2022.12.010
11 Shapiro D, Duquette CE, Zangerle C, et al. The Seniority Swoop: Young Nurse Burnout, Violence, and Turnover Intention in an 11-Hospital Sample. Nurs Adm Q 2022;46:60–71. doi:10.1097/NAQ.0000000000000502
12 Sikaras C, Ilias I, Tselebis A, et al. Nursing staff fatigue and burnout during the COVID-19 pandemic in Greece. AIMS Public Health 2022;9:94–105. doi:10.3934/publichealth.2022008
13 Castañeda-Aguilera E, García de-Alba-García JE. Professional burnout syndrome in specialist surgeons: prevalence and risk factors. Cir Cir 2020;88:354–60. doi:10.24875/CIRU.19001502
14 Karlafti E, Benioudakis ES, Barouxi E, et al. Exhaustion and burnout in the healthcare system in Greece: A cross-sectional study among internists during the COVID-19 lockdown. Psychiatriki 2022;33:21–30. doi:10.22365/jpsych.2022.067
15 Mantri S, Song YK, Lawson JM, et al. Moral Injury and Burnout in Health Care Professionals During the COVID-19 Pandemic. J Nerv Ment Dis 2021;209:720–6. doi:10.1097/NMD.0000000000001367
16 Soto-Rubio A, Giménez-Espert MDC, Prado-Gascó V. Effect of Emotional Intelligence and Psychosocial Risks on Burnout, Job Satisfaction, and Nurses’ Health during the COVID-19 Pandemic. Int J Environ Res Public Health 2020;17:7998. doi:10.3390/ijerph17217998
17 Tan BYQ, Kanneganti A, Lim LJH, et al. Burnout and Associated Factors Among Health Care Workers in Singapore During the COVID-19 Pandemic. J Am Med Dir Assoc 2020;21:1751–8. doi:10.1016/j.jamda.2020.09.035
18 Yeh W-Y, Cheng Y, Chen C-J, et al. Psychometric properties of the Chinese version of Copenhagen burnout inventory among employees in two companies in Taiwan. Int J Behav Med 2007;14:126–33. doi:10.1007/BF03000183
19 Hemann BA, Bimson WF, Taylor AJ. The Framingham Risk Score: an appraisal of its benefits and limitations. Am Heart Hosp J 2007;5:91–6. doi:10.1111/j.1541-9215.2007.06350.x
20 Ministry of Labor OSaHAT. Guideline for prevention of diseases arising from excessive workload. Taiwan: Ministry of Labor, 2019.
21 Stodolska A, Wójcik G, Barańska I, et al. Prevalence of burnout among healthcare professionals during the COVID-19 pandemic and associated factors - a scoping review. Int J Occup Med Environ Health 2023;36:21–58. doi:10.13075/ijomeh.1896.02007
22 Galanis P, Vraka I, Fragkou D, et al. Nurses’ burnout and associated risk factors during the COVID-19 pandemic: A systematic review and meta-analysis. J Adv Nurs 2021;77:3286–302. doi:10.1111/jan.14839
23 Murat M, Köse S, Savaşer S. Determination of stress, depression and burnout levels of front-line nurses during the COVID-19 pandemic. Int J Ment Health Nurs 2021;30:533–43. doi:10.1111/inm.12818
24 Denning M, Goh ET, Tan B, et al. Determinants of burnout and other aspects of psychological well-being in healthcare workers during the Covid-19 pandemic: A multinational cross-sectional study. PLoS One 2021;16:e0238666. doi:10.1371/journal.pone.0238666
25 Siddiqui I, Gupta J, Collett G, et al. Perceived workplace support and mental health, well-being and burnout among health care professionals during the COVID-19 pandemic: a cohort analysis. CMAJ Open 2023;11:E191–200. doi:10.9778/cmajo.20220191
26 Valavani A, Garavela E, Papathanasiou IV, et al. Investigation of the Level of Burnout in Health Care Professionals in COVID-19 Pandemic Conditions. Adv Exp Med Biol 2023;1425:319–24. doi:10.1007/978-3-031-31986-0_30
27 Lee YJ, Yun J, Kim T. Stress- and Work-Related Burnout in Frontline Health-Care Professionals During the COVID-19 Pandemic. Disaster Med Public Health Prep 2021;17:e38. doi:10.1017/dmp.2021.279
28 Pereira-Lima K, Loureiro SR, Silveira ILM, et al. Workplace Protections and Burnout Among Brazilian Frontline Health Care Professionals During the COVID-19 Pandemic. Front Psychol 2022;13:880049. doi:10.3389/fpsyg.2022.880049
29 Matsuo T, Yoshioka T, Okubo R, et al. Burnout and its associated factors among healthcare workers and the general working population in Japan during the COVID-19 pandemic: a nationwide cross-sectional internet-based study. BMJ Open 2022;12:e064716. doi:10.1136/bmjopen-2022-064716
30 Teo I, Chay J, Cheung YB, et al. Healthcare worker stress, anxiety and burnout during the COVID-19 pandemic in Singapore: A 6-month multi-centre prospective study. PLoS One 2021;16:e0258866. doi:10.1371/journal.pone.0258866
31 Debbarma S. COVID-19 Pandemic and Challenges Faced by Healthcare Professionals in India. Cureus 2023;15:e38278. doi:10.7759/cureus.38278
32 Chen WC, Huang A-E. COVID-19 public health surveillance and response in Taiwan: From containment to mitigation. J Formos Med Assoc 2024;123 Suppl 1:S17–26. doi:10.1016/j.jfma.2023.08.016
33 Wang VY. Systemic resilience and COVID-19: lessons from Taiwan. Int J Qual Health Care 2022;34:mzac029. doi:10.1093/intqhc/mzac029
34 Chien LC, Beÿ CK, Koenig KL. Taiwan’s Successful COVID-19 Mitigation and Containment Strategy: Achieving Quasi Population Immunity Disaster Med Public Health Prep 2022;16:434–7. doi:10.1017/dmp.2020.357
35 Mahmassani D, Tamim H, Makki M, et al. The impact of COVID-19 lockdown measures on ED visits in Lebanon. Am J Emerg Med 2021;46:634–9. doi:10.1016/j.ajem.2020.11.067
36 Butt AA, Azad AM, Kartha AB, et al. Volume and Acuity of Emergency Department Visits Prior To and After COVID-19. J Emerg Med 2020;59:730–4. doi:10.1016/j.jemermed.2020.08.013
37 Westgard BC, Morgan MW, Vazquez-Benitez G, et al. An Analysis of Changes in Emergency Department Visits After a State Declaration During the Time of COVID-19. Ann Emerg Med 2020;76:595–601. doi:10.1016/j.annemergmed.2020.06.019
38 Lin PH, Su HY, Tsai IT, et al. Impact of COVID-19 Pandemic on Emergency Department Volume and Acuity in Low Incidence Area: Taiwan’s Experience in Three Hospitals. J Acute Med 2022;12:105–12. doi:10.6705/j.jacme.202209_12(3).0003
39 Imeshtari V, Imeshtari A, Torre G. Cardiovascular risk in young healthcare professionals: A cross-sectional study. Work 2024;79:1665–72. doi:10.3233/wor-220701
40 Hsiao SH, Huang SJ, Huang CY. Vaccination Strategies at a COVID-19 Mass Vaccination Site. Int J Health Policy Manag 2022;11:1981–2. doi:10.34172/ijhpm.2022.6881
41 Lubis R, Satria FB, Zaki RA, et al. Factors influencing the COVID-19 pandemic situation in Indonesia, Malaysia and Taiwan in 2021. Public Health Pract (Oxf) 2022;4:100311. doi:10.1016/j.puhip.2022.100311
42 Gu W, Liu Y, Lu Z, et al. Associated factors of burnout among Chinese vaccination staff during COVID-19 epidemic: A cross-sectional study. Front Public Health 2023;11:1086889. doi:10.3389/fpubh.2023.1086889
43 Hrehova L, Seifert B, de Sire A, et al. Physicians' perceptions of working in mass vaccination sites during COVID-19 pandemic. Bratisl Lek Listy 2022;123:470–4. doi:10.4149/BLL_2022_073
44 Shah MK, Gandrakota N, Cimiotti JP, et al. Prevalence of and Factors Associated With Nurse Burnout in the US. JAMA Netw Open 2021;4:e2036469. doi:10.1001/jamanetworkopen.2020.36469
45 Chigwedere OC, Sadath A, Kabir Z, et al. The Impact of Epidemics and Pandemics on the Mental Health of Healthcare Workers: A Systematic Review. Int J Environ Res Public Health 2021;18:6695. doi:10.3390/ijerph18136695
46 Fisher JF. Permanent resident. Med Educ Online 2016;21:31160. doi:10.3402/meo.v21.31160
47 Girdler SJ, Girdler JE, Tarpada SP, et al. Nonmaleficence in medical training: Balancing patient care and efficient education. Indian J Med Ethics 2019;4:129–33. doi:10.20529/IJME.2018.100
48 Danet Danet A. Psychological impact of COVID-19 pandemic in Western frontline healthcare professionals. A systematic review. Med Clin (Barc) 2021;156:449–58. doi:10.1016/j.medcli.2020.11.009
49 Sultana A, Sharma R, Hossain MM, et al. Burnout among healthcare providers during COVID-19: Challenges and evidence-based interventions. Indian J Med Ethics 2020;V:1–6. doi:10.20529/IJME.2020.73
50 Shanafelt T, Ripp J, Trockel M. Understanding and Addressing Sources of Anxiety Among Health Care Professionals During the COVID-19 Pandemic. JAMA 2020;323:2133. doi:10.1001/jama.2020.5893
© Author(s) (or their employer(s)) 2025. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ Group.. This work is licensed under the Creative Commons Attribution – Non-Commercial License https://creativecommons.org/licenses/by-nc/4.0/ (the "License"). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.