2021 Volume 3 Issue 1 Article ID: 2021-0014-BR
Objectives: The aim of the present study was to evaluate the prevalence of burnout among both medical and non-medical workers at a single coronavirus disease 2019 (COVID-19)-dedicated hospital and to find factors significantly associated with burnout. Methods: One hundred seventy-nine respondents (81.4% of the total employees) who answered a questionnaire were included. The questionnaire consisted of 50 questions, including the Maslach Burnout Inventory-General Survey. Multiple logistic regression analyses were performed to calculate the odds ratios for burnout using a non-burnout group as a reference, with adjustment for age, sex, medical or non-medical worker status, presence or absence of work in direct contact with COVID-19 patients, and presence or absence of work without direct contact but which was related to COVID-19 patients. Results: Burnout was equally observed in both medical and non-medical workers; the overall burnout rate was 8.9%. The factors significantly associated with burnout were anxiety in relation to infection with COVID-19, self-quarantine and stress behavior of patients, lack of sleep in comparison to the pre-COVID-19 period, and the desire for more days off, increased staff, hazard pay, and resources for coping with stress. Conclusion: System-level solutions for these factors may be effective for reducing dropout intention, burnout, and resignation of hospital workers.
The coronavirus disease 2019 (COVID-19) pandemic represents a serious public health threat throughout the world. As there is currently no prospect of completely controlling COVID-19, in addition to the physical protection of hospital workers and vaccination, it is considered important to prevent them from experiencing psychological distress in order to avoid dropout intention, burnout and resignation. Burnout is a syndrome characterized by exhaustion, cynicism, and professional inefficacy, which has been shown to influence quality of care, patient safety, staff turnover, and patient satisfaction1,2,3,4).
Tokai University Tokyo Hospital is located near Tokyo’s largest COVID-19 hotspot, Shinjuku. It previously had 99 beds and provided advanced medical treatment in various fields as one of the university hospitals of Tokai University. In mid-September 2020, we accepted a request from the Tokyo Metropolitan Governor and turned our hospital into the first COVID-19-dedicated hospital (60 beds) in Japan. Since then, in addition to medical professionals, people of various other occupations have been engaged directly or indirectly in work related to COVID-19.
There have been some studies on burnout among healthcare professionals, which have demonstrated the importance of mental healthcare in the face of the COVID-19 pandemic5,6,7,8,9). A systematic review and meta-analysis showed that women, rather than men, and nurses, rather than doctors, showed higher rates of generalized anxiety, stress disorder, depression, and insomnia9). Although medical workers are absolutely essential for the treatment of COVID-19, there are other hospital workers whose jobs are also necessary for hospitals to provide healthcare. In the present study, we focused on burnout and evaluated its prevalence among both medical and non-medical workers at a single COVID-19-dedicated hospital to identify factors significantly associated with burnout.
The original survey was conducted from October 16, 2020 to October 30, 2020 as part of the mental health care program for employees of Tokai University Tokyo Hospital by an industrial doctor. Two hundred four responses containing personally identifiable information were obtained from 220 employees. Then, study participants were recruited by an opt-out method; finally, 179 respondents (81.4% of the total employees) were included in the present study. The 179 employees were categorized into eight occupations (doctor, nurse, pharmacist, medical technologist, medical clerk, kitchen staff, ward staff, and other hospital staff). Doctor, nurse, pharmacist, and medical technologist were classified as medical workers, while other workers were classified as non-medical workers. All cases in the sample were valid for the analysis. Anonymized data were used for the analysis and the privacy of participants was completely protected using unlinkable anonymization. This study was approved by the Ethics Committee of Tokai University (20R-318) and was conducted in accordance with the Declaration of Helsinki.
Questionnaire surveyThe web-based questionnaire was generated using a Google Form. The questionnaire included general characteristics, such as age, sex, and job category. The questionnaire consisted of 50 questions. Sixteen of the 50 questions were items of Maslach Burnout Inventory-General Survey (MBI-GS)1,2,3,10,11,12); the other 34 questions asked about average working hours in direct or indirect contact with COVID-19 patients, dropout intention, types of anxiety in relation to self and home, types of anxiety in relation to hospital work, changes in lifestyle in comparison to those in the pre-COVID-19 period, and types of support needed. Burnout was evaluated using the validated Japanese version of the MBI-GS10,11,12), which has been recognized as the leading measure of burnout for the study of occupational burnout, including burnout among health professionals. The MBI-GS consists of 16 items and measures three dimensions of burnout: exhaustion, cynicism, and professional efficacy. Individuals with high levels of exhaustion (>4.0) plus either high cynicism (>2.6) or low professional efficacy (<1.5) were considered to have a high risk of occupational burnout11).
Statistical analysisThe results are shown as the median and interquartile range (IQR) for continuous variables and frequency and proportion for categorical variables. The significance of differences between two categorical variables was compared using the chi-squared test. Multiple logistic regression analyses were performed to calculate the odds ratios for burnout using the non-burnout group as a reference and with adjustment for age, sex, medical or non-medical worker status, presence or absence of work involving direct contact with COVID-19 patients, and presence or absence of work not involving direct contact but which was related to COVID-19 patients. Statistical analyses were performed using SPSS version 26.0 (SPSS Inc., Chicago, IL, USA). All p values were two-tailed, and p values of <0.05 were considered to indicate statistical significance.
The demographic characteristics of the participants are shown in Table 1. Among the 179 employees enrolled in the present study, 95 (53%) were medical workers and 84 (47%) were non-medical workers. Seventy percent of the respondents were women (68.4% of medical workers and 71.4% of non-medical workers, p=0.745), and the median age was 41 (IQR, 30–50) years. Doctors and nurses showed higher frequencies of direct contact with COVID-19 patients in comparison to other occupations (Table 1), and there was a statistically significant difference between medical (72.6%) and non-medical (44.4%) workers (p<0.001). However, as Tokai University Tokyo Hospital is a COVID-19-dedicated hospital, overall 87.1% (88.4% of medical workers and 85.7% of non-medical workers, p=0.658) of employees performed work that did not involve direct contact with COVID-19 patients, but which was related to COVID-19 patients more than once a week. All (100%) of the doctors and pharmacists performed some work related to COVID-19 patients. Regarding psychological distress, 34.1% of the employees showed dropout intention (answered “yes” to the question asking if they had ever wanted to quit their job since the hospital started admitting COVID-19 patients) (Table 1). The percentage of individuals with dropout intention was higher among non-medical workers (38.1%) than among medical workers (30.5%) (p=0.344). In the overall study population, 8.9% of the respondents were experiencing burnout (Table 1). When classified by occupation, doctors showed the highest burnout rate (20%), followed by medical clerks (16.1%). There was no significant difference between medical (9.5%) and non-medical (8.3%) workers (p=0.996).
| Medical workers | Non-medical workers | Overall | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Doctor | Nurse | Pharmacist | Medical technologist | Medical clerk | Kitchen staff | Ward staff | Other hospital staff | |||
| Total number | 15 | 48 | 8 | 24 | 31 | 10 | 12 | 31 | 179 | |
| Male | % | 73.3% | 10.4% | 50.0% | 41.7% | 22.6% | 30.0% | 33.3% | 32.3% | 30.2% |
| Female | % | 26.7% | 89.6% | 50.0% | 58.3% | 77.4% | 70.0% | 66.7% | 67.7% | 69.8% |
| Age, median (IQR), years | 50 (46–59) | 33.5 (25–45) | 35 (35–40.8) | 42.5 (31.3–54.5) | 42 (27–54) | 34.5 (28–51) | 44 (30–50) | 43 (34–56) | 41 (30–50) | |
| Experience, median (IQR), years | 25 (16–34) | 11 (3–22.8) | 10 (8.3–16.5) | 13.5 (8.3–28.8) | 6 (2–27) | 8.5 (3.4–12.3) | 2.5 (1.2–9) | 7 (3–12) | 10 (4–20) | |
| Work hours, median (IQR), hours | 9.5 (7.8–11.0) | 8.0 (6.0–8.0) | 8.0 (7.8–9.0) | 8.0 (7.0–8.0) | 8.8 (7.0–10.0) | 8.0 (8.0–8.6) | 7.0 (7.0–8.0) | 7.0 (7.0–8.0) | 8.0 (7.0–8.5) | |
| Sleep hours, median (IQR), hours | 6.0 (5.0–6.0) | 6.0 (5.5–7.0) | 6.0 (5.0–7.0) | 6.0 (5.0–6.0) | 6.0 (5.0–6.0) | 5.0 (5.0–6.1) | 6.0 (5.0–7.0) | 6.0 (5.0–7.8) | 6.0 (5.0–7.0) | |
| Direct contact with COVID-19 patients | ||||||||||
| 0 days/wk | % | 6.7% | 16.7% | 50.0% | 58.3% | 38.7% | 80.0% | 50.0% | 67.7% | 41.3% |
| 1–4 days/wk | % | 46.7% | 62.5% | 50.0% | 25.0% | 38.7% | 10.0% | 25.0% | 25.8% | 39.7% |
| >5 days/wk | % | 46.7% | 20.8% | 0.0% | 16.7% | 22.6% | 10.0% | 25.0% | 6.5% | 19.0% |
| Job without direct contact but related to COVID-19 patients | ||||||||||
| 0 days/wk | % | 0.0% | 8.3% | 0.0% | 29.2% | 9.7% | 20.0% | 16.7% | 16.1% | 12.9% |
| 1–4 days/wk | % | 26.7% | 52.1% | 37.5% | 29.2% | 29.0% | 10.0% | 8.3% | 38.7% | 34.6% |
| >5 days/wk | % | 73.3% | 39.6% | 62.5% | 41.7% | 61.3% | 70.0% | 75.0% | 45.2% | 52.5% |
| Psychological distress | ||||||||||
| Dropout intention | % | 33.3% | 37.5% | 12.5% | 20.8% | 51.6% | 50.0% | 0.0% | 35.5% | 34.1% |
| Burnout | % | 20.0% | 8.3% | 0.0% | 8.3% | 16.1% | 0.0% | 0.0% | 6.5% | 8.9% |
COVID-19, coronavirus disease 2019; IQR, interquartile range.
Next, multiple logistic regression analyses were performed to identify factors significantly associated with burnout, with adjustment for age, sex, medical or non-medical worker status, presence or absence of work involving direct contact with COVID-19 patients, and presence or absence of work not involving direct contact but which was related to COVID-19 patients. All of the results of the multiple logistic regression analyses that are reported in the present study were obtained by a forced entry method, and a forward-backward stepwise selection method also yielded the same results. Table 2 shows the variables selected by multiple logistic regression analyses. The full set of the independent variables and the ORs are shown in eTable 1. These analyses revealed that anxiety related to infection with COVID-19 and self-quarantine were factors that were significantly associated with burnout. Self-quarantine in this context means the avoidance of socializing (private eating, drinking, and other non-essential outings). Regarding anxiety in relation to hospital work, the stress behavior of patients was identified as a factor that was significantly associated with burnout (Table 2). Tokai University Tokyo Hospital admits mild (SpO2 ≥96%) to moderate Ι (SpO2 93% to <96%) COVID-19 patients according to the grade of severity defined by the Ministry of Health, Labour and Welfare, Japan12). Because the patients’ conditions are generally not severe, some patients take their frustration out on medical staff verbally during hospitalization. This was defined as a stress behavior of patients. Burnout was significantly more prevalent in employees with a lack of sleep in comparison to the pre-COVID-19 period (Table 2). Respondents tended to show decreased relaxation time in comparison to the pre-COVID-19 period; however, the result did not reach statistical significance. When types of support required were used as independent variables, the following types of support were selected: more days off, staff increase, hazard pay, and resources for coping with stress (Table 2).
| Independent variables | Regression coefficient | SE | Wald | OR (95% CI) | p |
|---|---|---|---|---|---|
| Types of anxiety in relation to self and home | |||||
| Infection with COVID-19 | 1.045 | 0.481 | 4.730 | 2.844 (1.109–7.294) | 0.030 |
| Self-quarantine | 0.970 | 0.465 | 4.361 | 2.639 (1.061–6.560) | 0.037 |
| Types of anxiety in relation to hospital work | |||||
| Stress behavior of patients | 1.015 | 0.326 | 9.676 | 2.758 (1.455–5.228) | 0.002 |
| Changes in lifestyle in comparison to the pre-COVID-19 period | |||||
| Lack of sleep | 1.251 | 0.502 | 6.220 | 3.494 (1.307–9.337) | 0.013 |
| Types of support required | |||||
| More days off | 1.281 | 0.461 | 7.723 | 3.601 (1.459–8.888) | 0.005 |
| Increased staff | 0.711 | 0.293 | 5.895 | 2.036 (1.147–3.615) | 0.015 |
| Hazard pay | 1.612 | 0.651 | 6.124 | 5.013 (1.398–17.969) | 0.013 |
| Resources for coping with stress | 1.264 | 0.636 | 3.948 | 3.539 (1.017–12.315) | 0.047 |
CI, confidence interval; COVID-19, coronavirus disease 2019; OR, odds ratio; SE, standard error.
In the present study, a single-center cross-sectional survey was conducted to evaluate the prevalence of burnout in both medical and non-medical workers and revealed that the overall burnout rate was 8.9%. The following factors were significantly associated with burnout: anxiety in relation to infection with COVID-19, self-quarantine, and stress behavior of patients; lack of sleep in comparison to the pre-COVID-19 period; and desire for more days off, increased staff, hazard pay, and resources for coping with stress.
The burnout rate in this study was 8.9%, and there was no significant difference between medical and non-medical workers. Our results showed some differences in comparison to previous studies. According to a report that evaluated medical staff dispatched from a certain tertiary hospital to temporary mobile hospitals to treat COVID-19 patients in Wuhan, China in March 2020, the burnout frequency was 13–39%14). Matsuo et al. reported that 31.4% of healthcare workers who worked in emergency departments and the intensive care unit at a tertiary hospital in Japan in April 2020 met the criteria for burnout12). In comparison to these results, the burnout rate in our study is smaller. One of the reasons may be the background of the surveys. The previous studies12,14) were performed in tertiary care centers in spring 2020, when COVID-19 started emerging, whereas our survey was conducted in October 2020 in a hospital designated to treat mild to moderate COVID-19 patients. Another Japanese study performed almost in the same season used another version of the MBI (the MBI-Human Services Survey) to evaluate doctors and nurses working in an emergency intensive care unit. The study showed that the prevalence of burnout was 50%15). Contrary to other studies showing higher burnout rates in nurses9,12), the rate in nurses was only 8.3% in our study. The bed occupancy rate in our hospital was only 16.3% during the study period, which might have resulted in a lower burnout rate, especially among nurses. Despite the fact that our hospital provides non-intensive care, it is noteworthy that 8.9% of the employees met the criteria for burnout. It is also noteworthy that non-medical workers experienced burnout at an equal rate to medical workers, even though their work involved less direct contact with COVID-19 patients. The percentage of people with dropout intention, which is considered to represent predisposition to burnout, among non-medical workers (38.1%) was even higher than that among medical workers (30.5%). As Wu et al. pointed out in their paper14), this could depend on the individual’s perceived degree of control in relation to their situation. From the perspective of preventing burnout and resignation, control in the workplace is thought to be a major driver of engagement and important for avoiding burnout4).
The factors significantly associated with burnout were explored, and the following factors were identified: anxiety in relation to infection with COVID-19, self-quarantine, and stress behavior of patients; lack of sleep in comparison to the pre-COVID-19 period; and desire for more days off, increased staff, hazard pay, and resources for coping with stress. There is no question about anxiety in relation to infection with COVID-19 among employees working at a COVID-19-dedicated hospital. With regard to self-quarantine, this factor is present because hospital workers are more restricted in comparison to ordinary people in terms of private eating, drinking and other non-essential outings, which may impose additional stress. The reason why anxiety in relation to stress behavior of patients was selected may be because the patients’ conditions were generally not severe and some patients took their frustration out on medical staff during hospitalization. Contrary to the results of an earlier study12), anxiety in relation to unfamiliarity with personal protective equipment (PPE) was not selected in our study. In our study population, burnout was significantly more prevalent among respondents who reported a lack of sleep in comparison to the pre-COVID-19 period, which was exactly consistent with the result presented by Matsuo et al12). Finally, the types of support needed were evaluated in order to suggest strategies that could be implemented to prevent burnout. We found that the respondents desired more days off, increased staff, hazard pay, and resources for coping with stress. Morishita et al. reported that 88.5% of the healthcare workers engaged in COVID-19-related work considered financial incentives as important for motivation16). The other factors suggested by a scoping review were adequate provision and training on the use of PPE, strict infection control practices, shorter shift length, and provision of mental health and support services6). The examples implemented in China were the release of an online self-help manual, 24-hour psychological hotlines, and online consultations8). Although the strategies for burnout prevention during the COVID-19 pandemic may vary among institutions, system-level solutions are necessary.
The present study had several limitations. First, the sample size was not large enough to draw conclusions regarding marginal p-values. Second, we cannot deny the possibility of selection bias. Concern about identification by other staff members may have caused the respondents to score questions low or give no response. Third, the cause-effect relationship of our study is unclear because of its cross-sectional nature. The survey was conducted just 1 month after our hospital became a COVID-19-dedicated hospital, and the present results may be considered to represent the baseline level of burnout among our staff. Thus, longitudinal tracking of the factors identified in the present study is now underway to compare changes in the prevalence of burnout, especially at the peak winter season when COVID-19 spread more drastically. Finally, it may not be appropriate to directly compare the prevalence of burnout with other studies that used different questionnaires and/or cutoff points. The important implication of our study is that non-medical workers without direct contact with COVID-19 patients experienced burnout equally, and not simply the burnout rate itself.
In conclusion, burnout was equally observed in both medical and non-medical workers, and the overall burnout rate in a COVID-19-dedicated hospital was 8.9%. The factors significantly associated with burnout were anxiety in relation to infection with COVID-19, self-quarantine, and stress behavior of patients; lack of sleep in comparison to the pre-COVID-19 period; and a desire for more days off, increased staff, hazard pay, and resources for coping with stress. System-level solutions for these factors may be effective for reducing dropout intention, burnout, and resignation of hospital workers.
We express deep appreciation to Dr. Marie Yoshikawa, Mr. Satoshi Iwadate, Ms. Junko Sasaki, Ms. Chiori Takamatsu, Ms. Asuka Kobayashi and Ms Masayo Hamada for their supports on implementation process of questionnaire survey.
The authors declare no conflict of interests.
C.Y. and N.K. conceived and designed the analysis of the obtained data. Y.M. performed the statistical analysis. Y.T. contributed to the analysis and interpretation of data. C.Y. wrote the first draft of the paper. I.K., O.C., A.E., K.S., and Y.N. contributed to critical revision of the manuscript for important intellectual content. Y.N. supervised the study. All authors read and approved the final manuscript.
This article contains supplementary material (Appendix), which is available in the online version (doi: 10.1539/eohp.2021-0014-BR)
