The Effect of Burnout, Stress, Depression, and Anxiety on the General Efficacy of Health Service Provision Among Chinese Nurses During Coronavirus disease 2019 outbreak: A Single-Center Survey.


 Background: In-person caring for coronavirus disease patients amid the outbreak and shortage of medical supplies create burnout and fear of getting infected among healthcare workers. These deem questioning service provision’s efficacy by healthcare workers. This study aims to explore the effect of burnout, stress, depression, and anxiety on the efficacy of health service provision among nurses during the Coronavirus disease outbreak in China, by multiple linear regressionMethods: Our Cross-sectional study utilized three Chinese versions of validated questionnaires; Maslach Burnout Inventory, The General Self-Efficacy Scale, and Depression Anxiety Stress Scales to assess burnout, efficacy, and anxiety among Chinese nurses caring for Coronavirus disease patients during the outbreak. We performed multiple linear regression with burnout, depression, anxiety, and stress scores as independent variables versus general self-efficacy scores as the dependent variable. We used SPSS at a 95% level of significance in the data analysis, and the STROBE tool in the write-up.Results: There was a total of 408 study participants. The majority showed high levels of depersonalization, emotional exhaustion, low levels of personal achievements, and extremely high levels of depression, anxiety, and stress. The general self-efficacy was below the international average. Emotional exhaustion and stress demonstrated a small negative correlation with general self-efficacy. The effect of anxiety and depression on the general self-efficacy was not statistically significant.Conclusions: Coronavirus disease outbreak in China, overwhelmed healthcare facilities creating burnout, anxiety, stress, and depression among nurses working in the hard-hit areas. These may explain an observed lower than average general self-efficacy. However, only stress and emotional exhaustion were associated with reduced general self-efficacy among nurses.


Background
Believed to have started from a sh market in a twelve million city of Wuhan in China, coronavirus disease (COVID-19) grew from a localized epidemic to a global pandemic in about three months since its initial outbreak in late 2019. It had infected more than 450,000 people worldwide, as of 25th March 2020 (1,2). The viral strain was new to the human host. The virus name evolved from uno cial Wuhan coronavirus, 2019-nCOV to the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) (1,3,4). The COVID-19 affects all age groups at an estimated mortality rate of 3.4%; however, older and comorbid people are more vulnerable to severe disease. (5).
With 2-14 days estimated incubation period (6), COVID-19 has negatively impacted governments' budgets, production, and the world's supply chain, threatening to plummet the world economy (7). With no licensed vaccine or de nitive cure, governments are taking drastic measures to increase public awareness on prevention through social-distancing, as well as strengthening healthcare systems in terms of screening, testing, and management of con rmed patients (8). However, it is evident that governments with weaker healthcare systems could face more enormous challenges as compared to governments with more reliable healthcare systems (9).
China being the epicenter of the outbreak, reported the highest proportion of cases worldwide with a new number of cases reaching 15,000 cases per day during the peak of the disease in mid-February 2020. The situation overwhelmed health care facilities (10). The ooding of patients attending healthcare facilities not only led to fatigue and burnout among healthcare workers but also increased the risk of infection among healthcare workers. As of March 4th, more than 3,300 healthcare workers had contracted the virus in China, and at least 18 had died (10). The situation also caused an increased risk of nosocomial transmission among patients.
Despite these challenges, healthcare facilities across the globe developed a variety of mitigation measures to ensure the quality of service was not affected. Amid the outbreak, the Chinese government saw the need to mobilize health workers countrywide (11). This effort served to ease the burden that healthcare facilities faced in the hardest-hit areas, particularly Wuhan. In the US and Europe, on the other hand, several healthcare facilities employed the use of telemedical innovations in caring for COVID-19 patients. This aimed at reducing the risk of nosocomial infections resulting from in-person patients care. The method had proven to be conducive for quarantined individuals and protective to patients and health care workers as compared to in-person care (12). This being recently utilized in managing COVID-19 patients, it was not used in China amid the peak of the outbreak.
From the healthcare workers' perspective, in-person caring for COVID-19 patients during the outbreak creates burnout due to a large number of in owing patients, as well as emotional disturbances, including fear of getting infected at work and depression. The situation led to the concern of the e cacy of health services provided by exhausted and anxious healthcare workers. The aim of this study was, therefore, to explore the effect of burnout, stress, depression, and anxiety on the e cacy of health service provision among drafted nurses returning from Wuhan.

Design
Our study was a cross-sectional study involving 408 (100% response rate) drafted nurses, returning to our institution from Wuhan. These nurses were directly involved in the treatment of COVID-19 patients during the outbreak in Wuhan. We maintained communications with participants through a social-network mobile application, WeChat (13), that all drafted nurses from our institution shared.

Data collection
Three validated questionnaires, the Maslach Burnout Inventory (MBI), the General Self-E cacy Scale (GSES), and Depression Anxiety Stress Scales (DASS), were utilized for data collection. Socio-demographic characteristics, including age, sex, marital status, and education level, were collected. We collected data for a week, from March 16, 2020. We "electronically," prepared, shared, and received all dully-lled questionnaires from participants to reduce the risk of contamination by "paper-based" questionnaires as fomites.

Maslach Burnout Inventory (MBI)
MBI is a validated tool for measuring burnout among workers (14). In this study, a validated Chinese version of the MBI tool for human service survey was utilized (15). The tool contains three subscales; depersonalization (6 items), emotional exhaustion (9 items), and reduction in personal accomplishment (10 items), each of which corresponds with a normal score range of 6-9, 19-26, and 34-39, respectively. Scores below 6 and above 9 on the depersonalization subscale indicate higher and lower levels of burnout in terms of depersonalization, respectively. Scores below 19 and above 26 on the emotional exhaustion subscale indicate lower and higher levels of burnout in terms of emotional exhaustion, respectively. Scores below 34 and above 39 on the personal achievement subscale indicate lower and higher levels of burnout in terms of reduction in personal achievement, respectively. The internal consistency indices (Cronbach's α) for each of the three parts are 0.842, 0.912, and 0.793 for depersonalization, emotional exhaustion, and personal accomplishment, respectively (16). We sought permission to use the tool from Mind Garden Inc. available at https://www.mindgarden.com/ The General Self-E cacy Scale (GSES) GSES is a validated scale to measure the feeling of self-e cacy in an individual's deed (17). In this study, we used the validated Chinese version of the tool (18) to measure nurses' self-e cacy in managing patients amid the COVID-19 outbreak. The scale consists of ten Likert-scored questions summing to a minimum of 10 and a maximum of 40. A higher score indicates higher self-e cacy. The international average score is 29.55 (17). The internal consistency (Cronbach's alpha) for the scale is 0.92 (18). Since this scale is a public domain, no permission to use was needed.

Depression Anxiety Stress Scales (DASS)-21
DASS-21 is a validated scale used to measure three negative emotional states of depression, anxiety, and stress. In this study, the short version of the validated simpli ed Chinese scale was used (19). The scale consists of 21 Likert-scored questions summing to the minimum and maximum of zero and 21, respectively, for each of the three scales. The higher score indicates a higher level of depression, anxiety, or stress.
The internal consistencies (Cronbach's alpha) for the scale are 0.83, 0.80, and 0.82 for the depression, anxiety, and stress scales, respectively. The Cronbach's alpha for the overall DASS-21 is 0.92. Since this scale is a public domain, no permission to use is needed (20).

Calculation of sample size
We calculated the sample size by using the following formula (Daniel, 1999); n = (z) 2 p(1-p)/d 2 . Where, n = the estimated sample size, z = the standard normal deviation at 95% con dence level set as 1.96, p = the estimated proportion of problem 25%(21), q = 1-p and d = the precision error set as 5%.

Bias
We assessed detection, information, and self-selection biases. To minimize reporting biases, a tool, Strengthening the Reporting of Observational Studies in Epidemiology (Supplementary File 1) customized for the cross-sectional study was used in the write-up.

Ethical consideration
We sought written consent from the participants. We sought the approval to conduct this study from the Institution's review board (IRB) of behavioral and nursing research in School of Nursing of Central South University (CSU). The study's approval number was 2018035.

Analysis
We performed multiple linear regression to assess the effect of burnout and anxiety towards general self-e cacy. All scores were continuous variables. The three burnout sub-scales from MBI were analyzed separately as depersonalization, emotional exhaustion, and personal achievement. DASS-21 questionnaire provided depression, stress, and anxiety scores. The GSES questionnaire provided general self-e cacy scores. DASS-21 and MBI-HSS scores were our independent variables, while GSES scores served as dependent variables. We used the computer software, SPSS for analysis at a 95% level of signi cance. We performed forward-selection and backward-elimination multiple linear regression as additional analyses. A correlation coe cient of more than 0.7 de ned the presence of multicollinearity between predictor variables.

Results
A total of 408 nurses (Mean age: 36.67 ± 11.21) returned from Wuhan, where they were directly involved in treating COVID-19 patients. All drafted nurses (100%) participated in our study. Table 1 illustrates the participants' demographic characteristics.     Counterintuitively, depersonalization showed a small positive correlation with general self-e cacy. From the regression, a one-unit increase in the level of depersonalization was associated with a 0.025 increase in the general self-e cacy. However, this unexpected nding did not reach statistical signi cance (p-Value = 0.075).

Additional analysis
We additionally conducted a forward-selection and backward-eliminations multiple linear regressions. A forward selection method systematically identi ed only variables with less than 0.05 level of signi cance, Table 4, while backward elimination methods systematically eliminated all variables with more than 0.10 level of signi cance, Table 5.   Regarding self-e cacy, our study reported that all participants (100%) had levels that were below average (i.e., 14.28 ± 2.12). This is in contrast to the result by Ratnayake et al. (2016)(28) during the Ebola virus disease (EVD) outbreak where general self-e cacy among nurses caring for the sick was above average de ned by median of 4 (i.e., strongly agree) in the general self-e cacy Likert scale. In another study by Ko et al. (2004) (29), the mean general self-e cacy among nurses caring for SARS patients in Taiwan was 11.38 ± 12.98, which is lower than the one reported from our results. The study by Ratnayake et al. was conducted 2 years after the o cial announcement of the EVD outbreak in Sierra-Leone. After the epidemic had passed its peak, healthcare workers had already developed psychological acceptance, and they had already received thorough safety training courses, adequate medical supplies, and personal protective equipment (PPE) from WHO and partners. This is in contrast to our study, which was conducted about 2 months after the o cial announcement of the outbreak when the knowledge and experience of the disease were limited, the public was in a panic, the outbreak had not yet reached its peak, and PPE was inadequate. Time differences in conducting the studies could explain the differences in the self-e cacy results reported. We call upon for studies comparing trends of self-e cacies during different time points in the infection curve.
From our study, emotional exhaustion (Beta=-0.016, p-Value = 0.034) and stress (Beta=-0.138, p-Value = 0.004) were found to negatively affect general self-e cacy signi cantly. Anxiety, personal achievement, depersonalization, and depression did not show a statistically signi cant relationship with general self-e cacy. The results for anxiety contradict those reported by Ho et al. (2005) (30). They reported that increased anxiety signi cantly lowered self-e cacy among nurses during the SARS outbreak in 2002. In another study by Ng et al. (2006) (31) demonstrated that lower levels of depression among nurses caring for SARS patients were signi cantly associated with increased general-self e cacy. SARS had about 4 times higher fatality rate as compared to COVID-19 (32). This could mean that SARS created higher anxiety and depression levels in healthcare works as compared to COVID-19. We unexpectedly found a small positive correlation between depersonalization and general self-e cacy. Authors believe this was due to the encountered biases in our study. We call upon for lesser biased studies such as systematic reviews and meta-analyses on the topic.
In an attempt to stall the spread of COVID-19 amid the world's effort to develop new treatments and manufacture dire needed medical equipment, WHO o cials during the World Economic Forum's brie ng, recommended countries to practice "Isolate, Test, Treat and Trace." We support the slogan. However, we believe that the slogan is disproportionately favoring patients than healthcare workers. We, therefore, recommend to healthcare systems and partners to provide psychological support, protective training, and adequate PPE to healthcare workers. We also promote paying tributes to healthcare workers and other public service front liners across the world through cheering and encouragement posts in social media.
Limitations and strengths of the study Despite promising results, authors urge readers to interpret these results with caution due to several encountered biases. To mitigate non-response biases, adequate time (i.e., 1 week) to respond was given, con dentiality was guaranteed, and electronically accessed questionnaires were previously tested for compatibility to different electronic devices. Correspondence with participants was made possible through a joint WeChat group; therefore, mitigating attrition biases. The STROBE tool was utilized in the write-up to alleviate reporting and publication biases.
On the other hand, the extensive popular and nearly a "must-have" mobile application in China, WeChat, made it possible to reach all participants effectively. This reliable platform for correspondence may explain our study's strength achieving a 100% participants' response rate. Moreover, the participants had received a "hero-welcoming" upon their return from Wuhan. From an otherwise a weaker basis, such welcoming might also have motivated the participants to take part in the interviews.

Conclusion
COVID-19 disease outbreak in China, overwhelmed healthcare facilities creating burnout, anxiety, stress, and depression among nurses working in the hard-hit areas. These may explain an observed lower than average general self-e cacy. However, only stress and emotional exhaustion were associated with reduced general self-e cacy among nurses. Decreased nurses' general self-e cacy may increase medical errors, thus reducing patients' safety. We, therefore, recommend healthcare systems and partners to provide more support to healthcare workers in terms of psychological support, provision of protective training, and supply of adequate PPE. We also promote paying tributes to healthcare workers and other public service front liners across the world through cheering and encouragement posts in social media.