PHYSICIAN BURNOUT LEVELS AND ASSOCIATED FACTORS IN THE COVID-19 PANDEMIC

Introduction: Burnout, characterized by emotional exhaustion, depersonalization, and low personal accomplishment, is frequently observed in physicians. Aim: The purpose of this study was to determine burnout levels and associated factors in physicians during the pandemic. Material and Method: This cross-sectional study was performed online with 288 physicians from various fields in Erzurum . Sociodemographic questions and the Maslach Burnout Inventory represented the data collection tools. The data were collected online in May and June 2021. Statistical analysis: Descriptive statistics , Student's t and ANOVA tests And Multiple ordinal logistic regression analysis were used. p values <0.05 were regarded as significant. Analysis was performed on SPSS 22. Results: Physicians’ mean emotional exhaustion, depersonalization, and personal accomplishment component scores were 20.1±8.3, 6.7±4.5, and 21.1±4.5, respectively. Analysis showed that 49.7% of physicians exhibited moderate-high emotional exhaustion, 35.8% high-moderate depersonalization, and 69.8% signs of low personal accomplishment. Job title, regret concerning selecting the medical profession, satisfaction with the working environment, number of additional monthly out-of-hours shifts worked, regular sporting activity, and assessment of the physical conditions in the working environment emerged as factors affecting burnout components at regression analysis. Conclusion: The participants’ burnout levels were high. The planning of effective interventions addressing individual and work-related factors with a holistic approach is essential to halt this rapidly growing epidemic.

Maslach Burnout Inventory was developed by Maslach and Jackson (2) and adapted into Turkish by Ergin (11).This Likert-type scale (1= never, 2= very seldom, 3= sometimes, 4= usually, 5= always) consists of 22 items.Burnout is evaluated under three subdimensions; The emotional exhaustion (EE) dimension is evaluated in items 1, 2, 3, 6, 8, 13, 14, 16, and 20, depersonalization (DP) in items 5, 10, 11, 15, and 22, and personal accomplishment (PA) in items 4, 7, 9, 12, 17, 18, 19, and 21.The EE and DP dimensions consist of negative statements and the PA dimension of positive statements.High EE and DP scores and low PA scores are associated with higher levels of burnout.Scores obtained on the scale are associated with burnout at three different levels -EE, low ≤20, moderate 21-27, and high ≥28, DPlow ≤8, moderate 9-12, and high ≥13 and PAlow ≤23, moderate 24-27, and high ≥28 (12).The study data were collected in May-June 2021.Approval for the study was obtained from the Atatürk University Non-Interventional Clinical Research Ethical Committee.Participants completing the data collection tool sent to them electronically were regarded as consenting to take part.No personal information was collected within the scope of the study, and all data were kept secret.Data analysis was performed on Statistics Package for Social Sciences (version 22).Categorical variables were presented as numbers and percentages, and numerical variables as mean ± standard deviation.Normality of the distribution of numerical variables was evaluated using the Kolmogorov-Smirnov test, z values calculated for skewness and kurtosis, and chart methods.Student's t-test and One-Way ANOVA, the Kruskal Wallis and Mann Whitney U tests, and the Mann Whitney U test with Bonferroni correction at posthoc analyses were employed in the analysis of continuous variables, while chi-square tests were applied in the analysis of categorical variables.Spearman's rho correlation analysis was applied to investigate relationships between continuous variables.Ordinal logistic regression analysis was applied to evaluate the independent variables affecting the probability of inclusion of participants in the EE, DP, and PA groups determined based on defined cut-off points.p values <0.05 were regarded as significant for all analyses.

RESULTS
The mean age of the 288 physicians included in the study was 38.1±8.6, and 147 (51%) were women.The physicians; mean EE, DP, and PA dimension scores were 20.1±8.3, 6.7±4.5, and 21.1±4.5, respectively.The distribution of burnout dimensions of physicians according to cutoff points is presented in Table 1.Male and female participants exhibited similar mean EE, DP, and PA scores (p>0.05).Married physicians (73.3%) and those with children (63.9%) exhibited significantly lower DP scores (p=0.030 and p=0.012, respectively), while their PA scores were significantly higher (p=0.029 and p=0.002, respectively).There were significant differences between physicians' PA scores according to their income levels (p=0.001), and PA increased with income level.Physicians who had regular sporting activity(15.6%)registered significantly lower EE and DP scores and significantly higher PA scores (p=0.007,p=0.001, and p<0.001, respectively)(Table 2).It was observed that the score distributions of the physicians for all three dimensions of burnout were similar according to the branches (p>0.05 for all).EE and DP scores were significantly higher in physicians (51.7%) who were actively working in shifts (p<0.001).The distribution

O N L I N E F I R S T
of scores for all three dimensions of burnout differed significantly according to the physicians' regret about their choice of profession (p<0.001 for all) (Table 3).In addition, the score distributions for all three burnout dimensions were significantly different according to their satisfaction with the working environment and physical conditions, attending various training meetings, and following academic publications (p<0.001 for all) (Table 3).The results of multiple rank regression analysis applied to determine the factors affecting the burnout dimensions are presented in Table 4.In the EE dimension, female gender, low income, job title, regret about the choice of profession, and satisfaction with the working environment, were found to be significantly related factors (p<0.05 for all).On the other hand, the number of non-working shifts, regular sporting activity, and regret about profession choice were factors related to the DP dimension.Regular sporting activity, regret about the choice of profession, and physical working conditions were significantly associated with the ranking categories of the PA dimension (p<0.05 for all).

DISCUSSION
Burnout is a cognitive process emerging from the interaction of personal and work-related factors.This study produced a general framework in terms of burnout levels among physicians and related factors during the COVID-19 pandemic, which involved difficult working conditions for many physicians.The study also yielded results concerning the interactive role of factors associated with burnout.Moderate-high EE was observed in almost half the physicians in this study.Moderate-high DP was present in one-third of the physicians, and moderate-low PA levels in more than two-thirds.Our findings generally show high levels of burnout in physicians working in our region.High prevalences of burnout are reported for physicians in the majority of studies, although the results vary widely (3,4,10,13).The prevalence of general and subdimension burnout may vary in association with regional, national, and cultural effects, the health system, human factors, and patients' levels of education.The methodology and measurement tools employed may also affect the results.Mean age was negatively correlated with DP levels in this study and positively correlated with PA.However, no association was observed between EE categories and age.On the other hand, Koşan and Özkula reported low EE scores in the young physician group and high PA scores in the elderly physicians (10,14).In a study of 2576 Chinese physicians investigating the relationship between organizational and patient factors and burnout, Cheng et al. observed the highest level of burnout in the 35-44 age group (13).Ashraf's study of physicians from Pakistan reported similar findings (15).According to Maslach, age is the demographic variable most consistently associated with burnout, with younger physicians having a higher risk of burnout (6).Burnout component scores in the present study were similar between male and female physicians.However, the female gender was associated with higher EE levels in the regression model (OR=1.7).Our findings are compatible with the previous literature (6,10,14,15).However, it is clear that female doctors working in departments where workplace conflicts, workload, and stress are intense, are victims of burnout at higher levels than men (9,16,15).The gender differences observed in terms of burnout may be attributable to the greater

O N L I N E F I R S T
responsibilities assumed by women outside their working lives compared to men, and the variation in this depends on prevailing social structures.Married physicians registered significantly lower DP scores and higher PA scores in the present study.DP scores were lower among married physicians, but the difference was not statistically significant.These results are consistent with the local literature (14).In terms of marital status, non-married participants are known to be more disposed to burnout than married physicians (6,17).However, some studies have also reported no association between marital status and burnout (18).The inconsistency in the results for burnout may be due to factors such as the nature of society and the meaning that individuals attach to the institution of marriage.Physicians with children in this study registered lower DP scores and significantly higher PA scores.EE scores were lower among physicians with children, although the difference compared to those without children was not statistically significant.In a study of 324 general practitioners performed in 2019, Uyar et al. also reported significantly lower DP scores among participants with children (19).However, other studies have linked having children to high burnout levels.In their study of 1811 Chinese neurologists, Tian et al. reported a lower risk of burnout among participants with no children (17).Having children can sometimes affect burnout levels due to the psychological support it provides in terms of commitment to life and sometimes due to the various responsibilities it imposes on parents.PA scores in the present study were significantly higher in the high-income physician group, and PA scores were correlated with income.Unsatisfactory physician salaries are regarded as one of the determinants of work-related stress factors and thus, burnout (6,7).Although some studies have linked low wages to high burnout levels (13,15), others have found no link between pay and burnout (17).It appears that physicians' satisfaction with their pay varies between countries and exhibits show differing relationships with burnout levels.EE and DP scores were lower, while PA scores were higher, among physicians who engaged in regular sporting activity in this study.Similar findings have been reported in other studies from our region (10,20).Regular sporting activity may be a means of coping with stress and preventing burnout.No relationship was found in the present study between burnout and taking vacations, the presence of chronic disease, smoking, or alcohol use.However, Koşan reported a higher risk of burnout among physicians who did not take vacations and those who smoked (10).Burnout may increase the prevalence of alcohol use among troubled physicians (21-23).Significant differences were observed in this study in burnout component scores depending on physicians' job titles.Being on the lower rungs of the career ladder appears to be associated with higher burnout levels.Being a general practitioner was found to be a significant predictor of being in a higher EE category compared to being an academic (OR=7.9).Other risk factors increasing EE levels compared to academics were being a research assistant physician (OR=3.6)or a specialist (OR=2.8).Our findings are compatible with other studies from our region (10).In Dyrbye et al.'s study of 3574 general practitioners evaluating the factors associated with burnout symptoms and regret over career choice, burnout was reported in almost half of the participants (21).However, some studies have reported a low prevalence of burnout among general practitioners (3).The differences in burnout figures in career terms may derive from countries' different health systems.

O N L I N E F I R S T
Physicians who worked out-of-hours shifts registered significantly higher EE and DP scores in this study.A significant correlation was observed between the number of monthly out-of-hours shifts worked and component scores.Monthly out-of-hours shift numbers also emerged as an independent variable determining DP levels (OR= 1.2).Koşan also noted a high prevalence of burnout among physicians with extra shift duties and working more than six of these a month (10).Özkula reported significantly lower EE and DP scores among physicians not working outof-hours shifts but observed no variation in terms of the number of monthly out-of-hours shifts worked (14).There are results in the literature that support our findings (24).Intensive working hours also involving night shifts may exacerbate burnout levels due to loss of concentration, risk of error, and possibly also an increase in anxiety levels.Physicians who frequently experienced regret over their choice of career exhibited the lowest EE and DP scores and the lowest PA scores in this study.Physicians with no regrets concerning their choice of profession have a much lower likelihood of low EE levels.However, occasional feelings of regret over career choice appear to be significantly associated with the likelihood of being in the low DP category.Similar findings have also been reported in previous studies (17,21).In addition to personal factors, environmental factors are also associated with burnout (9,23).In the present study, physicians who were not satisfied with their working environment in general and found their physical conditions inadequate registered significantly higher EE and DP scores, and significantly lower PS scores.However, the likelihood of being in the higher DP category was significantly lower among physicians who were satisfied or slightly satisfied with their working environment.Our findings are consistent with those of Koşan (10).The positive effects of physician-and family-friendly institutional environments on physician wellbeing and burnout levels are important issues emphasized in the literature (25).EE and DP scores were significantly lower, while PA scores were significantly higher, among physicians who had taken part in such activities as training sessions, seminars, and congresses in the previous year and who read academic publications in their fields.In contrast, Cheng reported that regular clinical gatherings adversely impacted physicians' burnout levels (13).It appears that physicians regard occasional training activities as an opportunity to socialize and get away from the work environment and that these can thus positively affect burnout levels.

CONCLUSIONS
This study identified significant relationships between burnout levels and personal and workrelated factors.Burnout is damaging for physicians, patients, and applications performed.In conclusion, burnout is an important condition that must be carefully evaluated from all dimensions, and holistic approaches addressing individual and organizational factors are needed to combat it.There are several limitations to this study.In particular, problems deriving from the study's cross-sectional methodology need to be considered in terms of determining causality.This study focused more on individual determinants of burnout, and further studies are now needed to identify potential organizational and work environment-related determinants.In addition, the fact that the study data were collected online may have limited the representation capacity of the study population.

Table 1 .
Distribution of physicians' burnout dimensions by cut-off points.

Table 2 .
Distribution of some personal characteristics according to physicians' burnout dimension scores .

Table 4 .
Results of multiple ordinal logistic regression analysis evaluating factors determining the probability of physicians being in the different burnout dimension categories.Accepted papers are articles in press that have gone through due peer review process and have been accepted for publication by the Editorial Board of Sanamed.The final text of the article may be changed before the final publication.Accepted papers can already be cited using the year of online publication and the DOI, as follows: the author's last name and initial of the first name, article title, journal title, online first publication month and year, and the DOI.When the final article is assigned to volumes/issues of the journal, the Article in Press version will be removed and the final version will appear in the associated published volumes/issues of the journal.The date the article was made available online first will be carried over.Atatürk Mahallesi Atatürk Üniversitesi Tıp Fakültesi Morfoloji Binası A Blok Halk Sağlığı A.D.25240 Erzurum / Türkiye, email: dr_syilmaz@hotmail.comModel fitting: χ 2 = 63.172,p<0.001;Cox&Snell R 2 : 0.197, Nagelkerke R 2 : 0.249 EE: Emotional Exhaustion, DP: Depersonalization, PA: Personal Accomplishment, SE: Standart error, OR: Odds Ratio, CI: Confidence Interval O N L I N E F I R S T *