Depression, Anxiety, and Stress in a Burned-Out Oncologist
Article information
Abstract
Objective
Considering the established intertwined relationships between burnout and other psychiatric disorders, this study aimed to clarify the existing levels of depression, anxiety, stress, emotional exhaustion, depersonalization, and personal accomplishment and their relationship among oncologists—a seemingly professional group at risk for mental health issues development.
Methods
A cross-sectional observational study was conducted in January 2024, involving 159 oncologists from Serbia and the Serbian Republic. Participants completed an online questionnaire assessing socio-demographic characteristics and working conditons, depression, anxiety, stress, and burnout using validated scales—the Depression Anxiety Stress Scale-21 and the Maslach Burnout Inventory-Human Services Survey.
Results
Results indicate that respondents were mildly depressed, moderately anxious, and mildly stressed overall. Regarding three dimensions of burnout respondents have moderate emotional exhaustion, depersonalization, and low personal accomplishment. No significant differences were found in depression, anxiety, and stress among medical, surgical, and radiation oncologists. Whether the respondents are seeing a psychiatrist and/or undergoing psychiatric therapy was predictive of depression, anxiety, and stress. Economic status was predictive for depression and stress, while an additional predictive factor of anxiety was the presence of somatic illness. Further analysis showed depression and stress levels can be predicted by all three dimensions of burnout, whereas anxiety levels can be predicted by the level of depersonalization.
Conclusion
These findings emphasize the complex interplay between burnout and other mental health disorders in oncologists, highlighting the need for targeted mental health interventions and support systems within the oncology field to mitigate the psychological toll on these physicians.
INTRODUCTION
In their educational book, the American Society of Clinical Oncology examined the topic of mental health challenges among oncologists. They observed that while burnout is widely acknowledged in the field, other stigmatized mental health concerns, like depression, are rarely discussed or documented [1]. Existing literature highlights significant mental health challenges among oncologists, with differences existing between countries and different time periods. In 2021, the European Society for Medical Oncology Resilience Task Force showed increased mental health vulnerability observed among women, younger individuals, and those facing changes in working hours [2]. Reported depression rates among oncologists during this period varied widely, ranging from 20% to 50% across different countries [3-6].
The population of oncologists seems to show a connection between anxiety and depression, considering that it was shown among US oncologists that in those with anxiety, approximately a third also had depression, while a striking 96.6% of those with depression also felt anxious [3]. A 2022 multinational study focusing on oncologists from the Middle East and North Africa revealed increased anxiety and stress, particularly among younger and female clinicians [7]. Another study from the same period, centered on caregivers for multiple myeloma patients, reported an anxiety prevalence rate of 44.1% [8].
Research shows that between 42% and 69% of oncologists experience workplace-related stress [9]. Oncology is widely recognized as one of the most demanding medical specialties, frequently leading to emotional exhaustion among physicians [10]. When it comes to oncologist burnout, similar rates are found across medical, surgical, and radiation oncologists [11]. However, the prevalence of burnout among oncologists demonstrates considerable variability depending on the country, the specific population being studied, and the time period in which the research was conducted. A 2017 meta-analysis revealed that approximately one-third of over 4,000 oncologists reported symptoms of burnout [9]. In a 2019 survey conducted in the USA, 12.2% of practicing hematologists and oncologists reported experiencing high levels of burnout. Among them, 7.8% had ongoing burnout symptoms, and 4.4% felt so overwhelmed that they considered seeking professional assistance [12]. In Eastern Europe, a study involving 637 oncologists from 19 countries found that 72% were at high risk of burnout, with higher vulnerability observed among female oncologists and those in medical and radiation oncology [13]. A recent Croatian study finds moderate to high burnout in 86% of the oncologists [14]. Burnout and occupational stress are also widespread among Korean medical oncologists [15], for example, and burnout is often linked to excessive workplace demands [16]. Besides sociodemographic and work-related factors, psychiatric disorders, such as depression, can contribute to the onset of burnout and result from burnout [17]. Connection between depression level and burnout has already been shown in the population of Serbian oncologists [18]. While burnout symptoms may diminish when work conditions improve, depression typically does not. Moreover, there is significant overlap between the two, as physicians prone to depression may also be more vulnerable to burnout [19].
Giving the data on the intertwined relationships between burnout and other psychiatric disorders, this study aimed for clarification by establishing the levels of depression, anxiety, stress, emotional exhaustion, depersonalization, and personal accomplishment as three dimensions of burnout, and also by evaluating the relationship between these disorders in the population of oncologists.
METHODS
In January of 2024, a cross-sectional observational study was conducted, involving oncologists from Serbia and the Serbian Republic who were members of the Society of Medical Oncologists of Serbia (UMOS) and the Society of Oncologists of the Serbian Republic, respectively.
To create a significant regression model, the necessary sample size was calculated using the program G*Power v 3.1 (Heinrich Heine University Düsseldorf). With an effect size f2=0.15 (indicating a medium effect size with predictors accounting for 13% of the variance in the outcome), a significance level of 5% (α=0.05), and a test power of 0.95, the required sample size was determined to be 153. The potential sample consisted of 176 physicians from UMOS and 21 from the Society of Oncologists of the Serbian Republic, focusing on oncologists in tertiary and secondary hospitals who completed the questionnaire fully. Those who were volunteer physicians or did not complete the questionnaires were not included. Calculating the response rate was unattainable due to the incentive for sharing the questionnaire to other oncologists.
Sampling was conducted via an online questionnaire that included sections on socio-demographic characteristics and working conditons, as well as the Depression Anxiety Stress Scale-21 (DASS-21) and the Maslach Burnout Inventory-Human Services Survey (MBI-HSS). The study complied with the 1964 Declaration of Helsinki and its amendments and received approval from the Ethics Committee of the University Clinical Center of Kragujevac, Serbia (protocol code 01/24-292, obtained on January 13, 2024). The responses were collected via an email distribution method. The email addresses list was compiled after obtaining permission from the Head Boards of the UMOS and the Society of Oncologists of the Serbian Republic and ethical clearance. The questionnaires were distributed as an anonymous Google Forms link attached in the emails that we have sent out. The survey was initially sent out, and a follow-up was conducted 15 days later to improve the response rate. Written consent from respondents was waived due to the anonymous nature of the online questionnaire. The introductory section of the questionnaire included information on the research methodology, assurance of respondent anonymity, and a clear explanation that by clicking the “Next” button to proceed and completing the questionnaire, participants were providing their informed consent to take part in the study.
The DASS-21, a 21-item self-report scale, evaluates depression, anxiety, and stress levels over the past week using a four-point scale (0 to 3). Each subscale is categorized by severity levels: depression (normal 0–9, mild 10–13, moderate 14–20, severe 21–27, extremely severe 28 or more), anxiety (normal 0–7, mild 8–9, moderate 10–14, severe 15–19, extremely severe 20 or more), and stress (normal 0–14, mild 15–18, moderate 19–25, severe 26–33, extremely severe 34 or more) [20-26]. The version of DASS-21 used in the questionnaire was in the native language and had previously been utilized in studies involving Serbian-speaking populations [23,24]. The calculated reliability was α=0.884 for depression, α=0.819 for anxiety, and α=0.858 for stress.
The MBI-HSS, used to assess burnout, includes 22 items evaluating three dimensions: emotional exhaustion (EE), depersonalization (DP), and personal accomplishment (PA). Responses are rated on a scale from “never” to “every day.” [27] Cutoff scores are applied to categorize burnout: high EE (≥27), high DP (≥10), low PA (<33). Burnout is identified if an individual has either high EE or DP scores, or high EE along with either high DP or low PA scores [28,29]. The version of MBI-HSS used in the questionnaire was in the native language and had previously been utilized in studies involving Serbian-speaking populations [30,31]. Reliability for the burnout dimensions was excellent for EE (α=0.917), and good for DP (α=0.787) and PA (α=0.890). The authors obtained the appropriate license to administer from the MBI-HSS in Serbian language from Mind Garden website.
Data analysis was performed using SPSS v27.0 software package (Statistical Package for Social Sciences; IBM Corp.), with a focus on measures of central tendency and variability. The Kolmogorov-Smirnov test assessed the normality of data distribution. Chi-square analysis was used to examine the relationship between demographic variables and specialization type. To evaluate differences in measured disorders based on specialization, socio-demographic characteristics, and working conditions, an analysis of variance for independent samples was employed. Linear regression analysis with categorical predictors was used to evaluate the prediction model of depression, anxiety, and stress levels based on demographic characteristics and working conditions. Additionally, linear regression was used to test models predicting depression, anxiety, and stress based on three dimensions of burnout. Statistical significance was set at p<0.05.
RESULTS
The sample consists of 159 oncologists, comprising 107 medical oncologists (MO) (67.3%), 30 surgical oncologists (SO) (18.9%), and 22 radiation oncologists (RO) (13.8%). Table 1 offers a detailed overview of their socio-demographic characteristics and working conditions.
Socio-demographic characteristics and working conditions of the three respondent groups
The mean score on the scale of depression of the entire sample showed that the respondents are mildly depressed. When it comes to the anxiety scale, the respondents were in the moderately anxious category. The mean score on the scale of stress indicated mild stress of the total sample. On average, respondents show moderate EE, DP on the border between low and moderate and low PA.
Depression levels are most pronounced among the MO, while the SO have the lowest scores on this scale; however, this difference is not statistically significant (F(2,156)=2.395, p=0.095, ηp2=0.030). Regarding anxiety, the SO show the lowest anxiety levels, while the MO are slightly less anxious than the RO, and these differences were not statistically significant (F(2,156)=2.516, p=0.084, ηp2=0.031). Under the highest stress levels are the MO, while the SO have the least pronounced stress levels. The analysis did not show statistical significance of the differences between the groups (F(2,156)=1.899, p=0.153, ηp2=0.024).
The EE is highest among the MO, while it is at a similar level among the SO and the RO. Regarding the testing of differences between groups, Levene’s homoscedasticity test was statistically significant (p=0.021). Since the condition for variance analysis with the F test was not met, a non-parametric analysis was performed. It showed that the groups of oncologists differ statistically significantly in the levels of EE (t(49.868)=5.834, p=0.005). The Games Howell post hoc test showed that the EE was more pronounced among the MO than the SO, and that the average difference in the average scores of these two groups was 7.49 and was statistically significant (p=0.010). Other group comparisons did not show statistical significance (SO-RO, p=0.999; MO–RO, p=0.057). The DP was also highest among the MO, slightly lower among the SO, and lowest among the RO. Also, these differences were not statistically significant (F(2,156)=0.175, p=0.840, ηp2=0.002). The PA is the third indicator of burnout syndrome and a higher value on this scale indicates a less pronounced burnout syndrome. The RO showed the highest PA levels, the MO slightly lower, and the SO the lowest. Again, these differences were not statistically significant (F(2,156)=0.822, p=0.442, ηp2=0.010). The frequencies of different levels of the examined mental health issues, detailed presentation of means and medians based on the scores for the total sample, as well as the differences between the three groups of oncologists are shown in Table 2.
Levels of symptom manifestation for the mental health disorders analyzed in the study sample
In order to test possible differences in the levels of depression, anxiety, stress depending on the different socio-demographic characteristics and working conditions of the sample, a series of 21 one-factor analysis of variance was initiated for each of the above-mentioned variables. In the case of some variables of working conditions, the assumption of homoscedasticity of analysis of variance was not fulfilled, so non-parametric analyses were performed, and instead of the F statistic, the table contains the t statistics—result of Welch’s robust test of equality of arithmetic means.
When it comes to depression, differences were registered according to economic status (p=0.044) and whether the respondents are seeing a psychiatrist (p<0.001) and/or undergoing psychiatric therapy (p=0.013). The level of depression in the sample depends on the economic status in such a way that the lower a person’s economic status, the higher the score on the depression scale. Respondents who are seeing a psychiatrist and those who undergo some form of psychiatric therapy have significantly higher depression scores (Table 3).
Differences in the severity of depression levels according to socio-demographic characteristics and working conditions
Differences in scores on the anxiety scale were registered depending on the presence of a somatic illness (p=0.041), seeing a psychiatrist (p<0.001), and/or undergoing psychiatric therapy (p<0.001). Subjects with somatic illness show significantly higher levels of anxiety. The same applies to seeing a psychiatrist and undergoing psychiatric therapy (Table 4).
Differences in the severity of anxiety levels according to socio-demographic characteristics and working conditions
Differences in the level of stress were registered according to the economic status—the lower the economic status, the higher the level of stress (p=0.021), then according to seeing a psychiatrist, undergoing psychiatric therapy, and a history of depression in the family (Table 5).
Differences in the severity of stress levels according to socio-demographic characteristics and working conditions
Bivariate correlation analysis was used to test the interrelationships between the measured mental health issues. The EE and the DP correlated with depression, anxiety, and stress levels with strong statistical significance (p<0.001), while the PA showed no correlations to depression, anxiety, or stress (p>0.05). Due to high correlations between some predictors (e.g., intercorrelation between EE and DP), a multicollinearity analysis was performed (Table 6).
Inter-correlation matrix of the measured mental health disorders
Predictive analysis: depression
Since multicollinearity assumption was not violated, scores on the scales of EE, DP, and PA were included as predictors in the linear regression analysis, and the depression score was used as a criterion. The analysis showed that the described model is significant (F(3,155)=17.143, p<0.001). The multiple correlation between predictors and criteria was R=0.502, and the percentage of variance in scores on the depression scale explained by the predictors in the model was 25.2% (R2=0.252). Significant predictors of depression were all three elements of burnout. More precisely, the depression level increases by 0.260 standard deviations and by 0.286 standard deviations when the EE and the DP levels increase by one standard deviation, respectively. Also, when the PA level decreases by one standard deviation, the depression score increases by 0.204 standard deviation.
The same analysis was performed separately for the three groups of oncologists. In the group of MO, no multicollinearity of predictors was registered, and the model was significant (F(3,103)=10.576, p<0.001). The multiple correlation between predictors and criteria was R=0.485, and the percentage of variance in depression scale scores explained by predictors was 23.5% (R2=0.235). Significant predictors of depression among the MO remained only the DP (p=0.002). The standardized regression coefficients had the expected direction - an increase in depression level can be predicted based on an increase in the DP level. A regression model applied only to the SO showed no significance, and was without the multicollinearity of predictors (F(3,26)=1.466, p=0.247). Finally, when it comes to the RO there was also no multicollinearity and the model was significant (F(3,18)=6.062, p<0.05). However, no individual predictors showed significance in this model (Table 7).
Standardized and unstandardized regression coefficients, SE, t statistic, statistical significance of regression coefficients, and 95% CI
Predictive analysis: anxiety
Given that in this case the assumption of multicollinearity was not violated, the scores on the scales of EE, PA, and DP were included as predictors in the linear regression analysis, and the score on the anxiety scale was used as a criterion. The analysis showed that the described model is significant (F(3,155)=15.952, p<0.001). The multiple correlation between predictors and criteria was R=0.486, and the percentage of variance in the scores on the anxiety scale explained by the above-mentioned predictors in the model was 23.6% (R2=0.236). Only the DP contributes to the model with statistical significance (p<0.001). That is, when the DP level increases by one standard deviation, the anxiety level increases by 0.471 standard deviations.
In the group of MO, the assumption of the existence of multicollinearity was not violated and the model was statistically significant (F(3,103)=10.352, p<0.001). The multiple correlation between predictors and criteria was R=0.481, and the percentage of variance in the anxiety scale scores explained by predictors was 23.2% (R2=0.232). Significant predictors of anxiety among the MO remained only the DP (p<0.001). The standardized regression coefficients had the expected direction - an increase in the anxiety level can be predicted based on an increase in the DP level. The model, which included the SO, showed no violation of multiciollienarity assumption and was not statistically significant (F(3,26)=1.798, p=0.172). Regarding the model which included the RO, the assumption of multicollinearity was not violated and the model was significant (F(3,18)=6.769, p=0.003). The multiple correlation between predictors and criteria was R=0.728, and the percentage of variance in the anxiety scale scores explained by predictors was 53.0% (R2=0.530); however, no individual predictors showed as statistically significant (Table 7).
Predictive analysis: stress
In our model, with the scores on the scales of EE, PA, and DP as predictors in the linear regression analysis, and the score on the stress scale as a criterion, the assumption of the existence of multicollinearity was not violated. The analysis showed that the described model is significant (F(3,155)=30.213, p<0.001). The multiple correlation between predictors and criteria was R=0.607, and the percentage of variance in the scores on the stress scale explained by the above-mentioned predictors in the model was 36.9% (R2=0.369). All three predictors contribute to the model with statistical significance (p<0.001), with the emphasis that according to the standardized Beta coefficient, the EE contributes the most to the model. That is, when the EE increases by one standard deviation, the stress level increases by 0.221 standard deviations. When the DP increases by one standard deviation, the stress level increases by 0.404 standard deviations. Finally, when the PA increases by one standard deviation, stress level decreases by 0.161 standard deviations.
In the group of the MO, the assumption of multicolliearity was not violated, and the model was significant (F(3,103)=18.227, p<0.001). The multiple correlation between predictors and criteria was R=0.589, and the percentage of variance in the stress scale scores explained by the predictors was 34.7% (R2=0.347). Significant predictors of stress were all three dimensions of burnout. Stress level increases by 0.187 standard deviations and by 0.399 standard deviations when the EE and the DP levels increase by one standard deviation, respectively. Also, when the PA decreases by one standard deviation, the stress score increases by 0.180 standard deviations. A regression model applied only to the SO showed statistical significance, and was without multicollinearity assumption violation (F(3,26)=4.430, p=0.012); however, no individual predictors showed the statistical significance. Finally, when it comes to the RO there was also no multicollinearity assumption violation and the model was significant (F(3,18)=7.347, p=0.002); however, no individual predictors showed the significance in this model (Table 7).
DISCUSSION
Even though our sample on average showed mild depression, moderate anxiety, and mild stress levels, the percentage of the respondents with severe and very severe levels of these disorders are 3.1% and 5.7%, 5.7% and 12.6%, 8.2% and 3.8%, respectively, which we considered alarming. Similar prevalence for higher depression levels is found in a recently published study on German oncologists4; however, the prevalence data for high anxiety in this study seems to be lower than ours. Some earlier data indicate an approximate prevalence of 20% for those with total registered anxiety, while our results of similar prevalence are for those with severe and very severe levels of anxiety [32]. In the stress domain, a study from 2021, evaluates the high stress levels in the 4.4% of the sample, which is again significantly lower than our sample [33]. In our sample, no statistically significant differences in the DASS-21 scores were found between the MO, SO, and RO, which is, to the best knowledge of the authors, previously an unexplored subject in the literature. Exploring this on a larger sample which would include more SO and RO would either confirm or debunk our results.
Regarding burnout, all respondents show the moderate EE, the DP on the border between low and moderate and the low PA; however, high levels of the EE and the DP are registered in 37.1% and 23.3% of the sample, respectively, while the low PA was registered in 49.1% of the sample. Kyoto Radiation Oncology Study Group (KROSG) study in this regard showed the high EE and DP with approximately two- and three-times lower prevalence values, while the low PA had higher prevalence values than ours [34]. Most recently published data on burnout showed significantly higher prevalence of the high EE and DP and similar prevalence of the low PA with no differences between the MO and RO which was also shown in our results [35]; however, we do note that in our study the MO had the higher EE than the SO. A multinational study would help clarify the disparities between different countries and their oncologists in regard to the burnout levels.
In our study, the economic status was shown to correlate to both depression and stress levels. A Brazilian study from 2018 found different results to ours regarding depression levels [32]. More recent data from a study in which the oncologists were a part of the sample did not show correlation between the income and the depression, anxiety, or stress [36]. Considering the differences between the countries and the time periods of these studies and ours, further research is necessary to clarify these connections. Seeing a psychiatrist or undergoing psychiatric therapy also showed the correlations to the depression, anxiety, and stress levels, which suggests that oncologists recognize their own mental health issues and are seeking professional help. This differs from data from 2023 obtained by Medscape in which approximately 90% of the oncologists did not seek professional help for their mental health issues [37]; certainly, further research in populations of the oncologists is necessary for obtaining the definitive conclusion. We also found the correlations between the anxiety level and the presence of the somatic illness as well as the correlation between the stress level and the depression in the family history. A study from 2021 showed that during the pandemic oncologists were worried about their own and their family health [38], which, in light of our results, suggests that oncologists’ own health and that of their families can be a source of elevated anxiety and stress. However, definitive comparisons and conclusions are not possible; further research is needed to shed more light on this subject.
Between the depression, anxiety and stress levels, mutual correlations were shown, as well as the correlation of measured levels of expression of these three mental health disorders with the EE and the DP levels, but not with the PA levels. The correlation between the burnout and the depression or the burnout and the anxiety has already been confirmed, both in healthcare population and in other occupational populations [39,40], confirming our results. Recently, in a report on physician burnout and depression from 2024, Medscape reported that more than 50% of physicians linked the burnout and depression to job-related stress, and they also identified the job burnout as a contributing factor to the depression [41], which is very similar to the data we have found. In oncologists, a study published in 2023, showed significant associations of the depression, anxiety and burnout to the moral distress [4]. In the gynecology oncologists, similar association between the burnout and the depressive symptoms was also shown [42]. In her PhD thesis, the psychologist Andrijic [18] examined the presence of the depression among the oncologists in Serbia using the Patient Depression Questionnaire (PHQ-9) questionnaire. The average PHQ-9 score was 8.4±6.1, indicating mild depression among participants. The study revealed that 14.1% of the respondents without depressive symptoms experienced burnout syndrome. This prevalence increased to 34% among those with mild depressive symptoms. Notably, 96.6% of the healthcare workers and associates with severe depressive episodes also suffered from the burnout syndrome. The data analysis further demonstrated that the frequency of burnout syndrome at work rose in tandem with the severity of the depression, with those experiencing the higher levels of burnout significantly more likely to also suffer from the depression.
Both the depression and stress levels on their respective scales can be predicted by all three measured dimensions of burnout. Anxiety levels can be predicted only by the DP levels. All three dimensions of burnout were predictive of the stress levels in the MO; however, for the SO and the RO, this was not shown. Only the DP was predictive of the depression and anxiety levels in the group of the MO, while regression models for the SO and the RO in this regard were not significant. Using a linear regression analysis and General Health Questionnaire–12 item. questionnaire for evaluation of respondents with a diagnosable psychiatric disorder or as a broad indicator of psychiatric well-being [43], the KROSG study on the RO demonstrated that a high level of the EE and a low level of the PA were significantly associated with an elevated level of the psychological morbidity [34]. Including both the MO and the RO in a convenience sample and using the linear regression, Granek et al. [44] showed that burnout can predict elevated emotional distress. These studies cannot be compared with our results due to the methodology disparity issue; however, it implies the existence of the relationship we found between depression, anxiety, and stress to the burnout. Further research with the similar or the same methodology as ours may clarify these correlations in the oncologist group.
When interpreting the results of this research, several limitations should be taken into account. First, as a cross-sectional study, the data were collected at a single point in time, offering a snapshot of the respondents’ current state and the characteristics. However, this design does not capture changes over time, which limits the ability to draw the definitive conclusions about causality. Repeating this research would provide a more detailed understanding of how the mental health issues among the oncologists develop and how they are associated with the variables examined over time. Additionally, variations between countries and the potential cultural effects should be considered in the future research planning. The findings of this study are based on the self-assessment questionnaires, which rely on the honesty of respondents in answering the questions. While this method has many advantages, the responses may be subject to the overestimation or the underestimation of subjective experiences, potentially leading to the distorted results. It is also important to note that comparing the results from different studies on the oncologists’ mental health is challenging due to the methodological differences and the heterogeneity of criteria used to define and assess samples, as well as the use of different approaches in scoring and interpreting results. Finally, it should be emphasized that this comprehensive research was the first of its kind conducted in Serbia, involving multiple institutions that deal with malignancies in their daily practice.
In conclusion, the study highlights the significant mental health challenges faced by the oncologists, with a notable prevalence of the elevated levels of depression, anxiety, and stress across the profession. Furthermore, the findings indicate that the MO exhibit higher levels of the EE compared to their surgical and radiation counterparts, suggesting that the demands of the medical oncology may contribute more to the burnout. However, the differences in the depression, anxiety, and stress levels among the different oncologist groups were not statistically significant, underscoring that these mental health issues are widespread across all the specialties. The study also reaffirms the complex relationship between the burnout and other mental health issues, emphasizing that while the burnout may fluctuate, underlying issues like depression and anxiety persist. These results underscore the need for the targeted mental health interventions and support systems within the oncology field to mitigate the psychological toll on these healthcare professionals.
Notes
Availability of Data and Material
The datasets created for the purpose of the study can be obtained from the corresponding author upon a reasonable request.
Conflicts of Interest
The authors have no potential conflicts of interest to disclose.
Author Contributions
Conceptualization: all authors. Data curation: Nevena Randjelovic, Aleksandar Djuric. Formal analysis: Nevena Randjelovic, Kristina Dugalic. Methodology: Nevena Randjelovic, Marija Peulic, Marina Petronijevic. Writing—original draft: Nevena Randjelovic. Writing—review & editing: all authors.
Funding Statement
None
Acknowledgments
We would like to express our deepest gratitude to Prof. Dragana Ignjatovic- Ristic for her invaluable guidance and support. Her encouragement has been instrumental in shaping both the content and direction of this research. Prof. Ignjatovic-Ristic’s dedication and her unwavering commitment to fostering the well-being of healthcare professionals have been a constant source of inspiration.