Introduction

No medical phenomenon in recent history has attracted such necessity for profound multilevel exploration of bio-psycho-social effects on humankind as the COVID-19 pandemic. The battle with the pandemic started with restrictive epidemiological measures which were similar to those worldwide1,2,3. However, in spite of their effects on virus spreading control, the aforementioned preventive measures, and especially quarantine and social isolation, may have had an extensive consequence on mental and physical health on general population as well on specific vulnerable groups4. Further on, the first year of pandemic was filled with several COVID-19 outbreaks, periods of different protection measures, and a lot of uncertainty and long-term adaption to disrupted daily life. A number of studies increased awareness on broad spectrum of emotional, psychological and behavioural alternation, particularly depression, anxiety, stress and stress-related disorders3,5,6,7,8,9,10. In Balkan region, research performed by Margetic et al.11 on a large sample of the population from Croatia, during the COVID-19 lockdown and after the country being hit by an earthquake found that 15.9% of the respondents experienced severe to extreme depression, 10.7% severe to extreme anxiety and 26.2% severe to extreme stress, while main predictors of emotional distress were identified to be lower scores of emotional stability as well as factors related to avoidant coping strategies. This could be of importance, not only in general mental health risks, but also in the context of the recent research findings that indicate anxiety and depression as potential predictors of long-COVID12. However, not everyone would experience depression, anxiety and stress symptoms in similar intensity, and this could depend on various factors. Some of the related factors could refer to socio-demographics, such as gender. The pandemic has raised some “old issues” related to gender disparities and added to an existing dual burden of “work-home overload” on the employed women. Thus, some differences were noted between genders, ranging from job-correlated circumstances, across academic achievements and opportunities, to the household activities13,14,15,16. Also, age, relationship status, living arrangements, child and elderly care, medical history, professional hierarchy and speciality, direct or indirect exposition to COVID-19 patients, could all be a source of additional psychological burden17,18,19,20. The study of Ajdukovic et al.21 emphasizes the importance of the current state of health, previous mental health diagnosis and psychological resilience for mental health outcomes. Other potentially relevant factors could be related to partner and family characteristics, employment, satisfaction with living circumstances, income or family relations, having someone close contracting COVID-19 or dying from COVID-1922,23,24,25,26,27,28. When it comes to professional factors, the group with very specific exposure to the pandemic were health care professionals (HCP), whose mental health functioning during this period was a target of various studies5,10,29,30,31,32,33,34,35,36, identifying HCP as a particularly vulnerable group faced with physically and psychologically challenging situation during the pandemic6,7. Among the HCP, mental healthcare professionals may represent an even more specific group. Bearing in mind the numerous psychological and psychosocial consequences during the pandemic, mental health professionals were needed more than ever to optimize the needs of the population in the domain of mental health and interventions related to stress response. However, small number of studies dealt with the mental health functioning of mental health professionals worldwide. Some studies from different countries draw attention to differences between psychiatrists and other medical specialities in the context of emotional and behavioural responses during the COVID-19 outbreak18,37, but a number of publications shedding light on the mental health of psychiatrists are still scarce38.

In Serbia, authorities decreed lockdown measures on March 15, 2020, which lasted until May 6, 20201 in order to prevent the spread of virus and „flatten the growth curve. “Restrictive epidemiological measures at the beginning of the pandemic were similar to those worldwide1,2,3. The life-threatening pandemic and the restrictive measures could be seen as yet another factor in the relatively recent history of traumatic exposures in Serbia (for example, massive floods in 201439, NATO bombing in 199940), which may affect the mental health functioning in a culturally sensitive way, and specifically among the HCP who are always „on the frontline “in such situations. Several papers were published in Serbia between the start of the pandemic and January 2022, increasing data on the mental health of healthcare providers, including anxiety, depression, sleep disorders, burnout, and changes in behaviour and environmental factors highlighting the importance of acknowledging their psychological difficulties17. Among these studies, none focused specifically on mental health professionals. When it comes to general population, a study from a nationally representative sample during the second year of pandemic in Serbia41, has found that 15.2% of the sample had a diagnosable psychiatric disorder (4.6% had mood disorders, 4.3% had anxiety disorders, and 8.0% had substance use disorders) not exceeding the range of pre-pandemic data from the literature. However, exploring the relevant predictors of the emotion-based mental health symptoms in the first year of the pandemic (during the adaptation period to the first COVID-19 outbreaks and restrictive measures, when the vaccination against the SARS-CoV-2 virus was not yet available) would be of specific interest in Serbia.

As the aim of this study, we tried to simultaneously explore a number of different factors potentially contributing to emotional distress during the first year of the pandemic in Serbia, specifically addressing the needs acknowledged by the literature (such as distress symptoms in psychiatrists). The key research questions of our study referred to identifying the predictors of emotional distress (depression, anxiety and stress symptom levels), among the sociodemographic factors, variables related to profession (with special focus on HCP status and psychiatric profession status), partner and family related variables, satisfaction with living circumstances, income and family relations, pandemic related circumstances, and pre-pandemic health-related variables during the first year of COVID-19 pandemic in Serbia.

Materials and methods

Study design and participants

This large cross-sectional study was conducted in Serbia during the first year of the pandemic. It started shortly after the end of the state of emergency, and relatively near to the onset of the pandemic in Serbia (the emergency state in Serbia ended on May 6, 2020, while the data collection started on May 10, 2020). The study is part of wider international multicentric study coordinated by the Italian Society of Social Psychiatry42. The study was approved by the Ethics Committee of the Clinical Centre of Serbia (No 502/3; May 6, 2020), performed in accordance with relevant guidelines/regulations and conducted in accordance with the Helsinki Declaration of 1989. The informed consent was obtained from all participants.

The study included 3328 participants. Participants were recruited using two-level chain-referral sampling method, as the main method of data collection, and convenience sampling (in the small part of the sample). The participants who provided electronic informed consent were included in the study with the following criteria: (1) being 18 years old or above, and (2) having resided in Serbia during the majority of the state of emergency period. Once the participant was included, there were no further criteria for exclusion.

The sociodemographic characteristics of participants are summarized in Table 1, whereas the work-related, pandemic-related, and health-related characteristics of participants are shown in Table 2.

Table 1 Sociodemographic characteristics of participants.
Full size table
Table 2 Work-related, Pandemic-related and Health-related characteristics of participants.
Full size table

Procedure and instruments

The research commenced following the conclusion of the state of emergency in Serbia, indicated by the conclusion of quarantine measures. The study was conducted through an anonymous web-based survey. Prior to filling in the online questionnaire, participants were provided with detailed information about the study’s objectives, eligibility criteria, procedural details, confidentiality, and the option to discontinue their participation at any point. The online form did not collect any identifiable information such as names or IP addresses, and all other gathered data was treated as strictly confidential. Subsequently, respondents were given the choice to either consent to participate in the study or decline. Those who agreed were permitted to proceed with the self-report questionnaire.

The online questionnaire consisted of the following instruments.

  1. (a)

    General questionnaire: This questionnaire was designed for the purpose of the study and included data on: (a) sociodemographic variables (age, gender, the highest level of educational achievement); (b) work-related variables (employment status, being a physician or a nurse, among those being a physician or a nurse who worked in the frontline during COVID pandemic, being a doctor, and among those being a psychiatrist); (c) social variables (relationship status – having a partner, living with a partner; family characteristics – household size (number of household members), being a parent, number of children, having at least 1 child under 18 years); (d) satisfaction with living circumstances, income, or family relations (participants were asked to rate their satisfaction with living circumstances, income and family relations (separately) on 7-item Likert type scale (1. It can’t be worse; 2. I am not satisfied at all; 3. I am mostly dissatisfied; 4. I am neither satisfied nor not; 5. I’m satisfied; 6. I’m very satisfied; 7. It couldn’t be better; further on this was presented as a dichotomous variable with those who answered 5,6, and 7 identified as satisfied, and the rest of the sample identified as dissatisfied); (e) pandemic-related circumstances (having a close contact who has contracted COVID-19, or died due to COVID-19); and (f) health-related variables (having a previously diagnosed somatic disorder and mental disorder).

  2. (b)

    The Depression Anxiety Stress Scales—Short Form (DASS-21)43: were used to measure the symptoms of depression, anxiety and stress. DASS-21 is a brief version of the DASS-4244, and it consists of three self-report scales designed to measure the emotional states of depression, anxiety, and stress. This instrument was widely used in both clinical and non-clinical settings. The Depression scale assesses dysphoria, hopelessness, devaluation of life, self-deprecation, lack of interest/involvement, anhedonia, and inertia. The Anxiety scale assesses autonomic arousal, skeletal muscle effects, situational anxiety, and subjective experience of anxious affect. The Stress scale is sensitive to levels of chronic non-specific arousal. It assesses difficulty relaxing, nervous arousal, and being easily upset/agitated, irritable/over-reactive, and impatient. Participants are asked to use 4-point severity/frequency scales to rate the extent to which they have experienced each state over the past week. Scores for Depression, Anxiety, and Stress are calculated by summing the scores for the relevant items. The DASS has been validated in a variety of populations, including Serbia45 where it has shown to be a reliable and valid measure of unpleasant emotional states.

Statistical methods

Description of categorical data was done using absolute and relative (percentages) numbers, while numerical data were presented as the arithmetic mean with standard deviation or as median with range, depending on the data distribution. Normal distribution was evaluated by mathematical (Shapiro–Wilk and Kolmogorov–Smirnov tests, skewness and kurtosis, coefficient of variation) and graphical (histogram and box-plot) methods. The influence of all evaluated factors on the DASS score (total, depression, stress, and anxiety) was assessed using linear regression modelling, first univariate, then multivariate. All variables that were statistically significant at the level of 0.05 from univariate, excluding those which were in multicollinearity evaluated by VIF and tolerance methods46, were included in multivariate models. Coefficient B, 95% confidence interval of B (95%CI B), and p-value are reported. All statistical methods were considered significant for the chosen level of 0.05%.

Results

DASS total score

The level of negative emotional states was evaluated through DASS Total, DASS-Depression, DASS-Stress, and DASS-Anxiety scores. The average DASS Total score was 15.33 ± 13.13 (median and range were12 (0–61)), DASS-Depression score was 4.62 ± 4.75 (median and range were, 3 (0–21)), DASS-Stress score was 7.28 ± 5.30 (median and range were, 7 (0–21)), and DASS-Anxiety score was 3.43 ± 4.39 (median and range were 2 (0–21)). Descriptive DASS scores are presented in Supplementary Material 1 (Table S1).

Factors that were univariately associated with higher DASS-Total score were younger age, female gender, lower educational level, not being a doctor, doctor non-psychiatrist, frontline doctor and/or nurse, living without a partner, smaller number of children, having at least one child under 18 years, having close contact ill from or died due to COVID-19, previous somatic and mental disorder, as well as, dissatisfaction with financial situation, family relations, and living conditions. After multivariate linear regression modelling, female gender, doctor non-psychiatrist, previous mental disorder, and dissatisfaction with financial situation, family relations, and living conditions were independently associated with higher level of DASS-total score (F = 5.862, p < 0.001, R2 = 19.4%, Adjusted R2 = 16.1%) (Table 3).

Table 3 Factors associated with DASS-Total score.
Full size table

DASS—Depression score

Factors that were univariately associated with higher DASS-Depression score were younger age, female gender, lower educational level, not being a doctor, non-psychiatrist, living without a partner, smaller number of children, having at least one child under 18 years, smaller number of family members, being a frontline doctor and/or nurse, having close contact contracted COVID-19/died due to COVID-19, previous physical and mental disorder, as well as, dissatisfaction with financial situation, family relations, and living conditions. After multivariate linear regression modelling, female gender, doctor non-psychiatrist, close contact contracted COVID-19/ died due to COVID-19, previous mental disorder, and dissatisfaction with financial situation, family relations, and living conditions were independently associated with higher level of DASS-Depression score (F = 5.551, p < 0.001, R2 = 19.8%, Adjusted R2 = 16.2%) (Table 4).

Table 4 Factors associated with DASS-D, DASS-S and DASS-A.
Full size table

DASS – Stress score

Factors that were univariately associated with higher DASS-Stress score were younger age, female gender, lower educational level, not being a doctor, doctor non-psychiatrist, living without a partner, without or with smaller number of children, having at least one child under 18 years, being frontline doctor and/or nurse, having close contact that contracted COVID-19 and /or died due to COVID-19, previous physical and mental disorder, as well as, dissatisfaction with financial situation, family relations, and living conditions. After multivariate linear regression modelling, female gender, doctor non-psychiatrist, living with a partner, previous mental disorder, and dissatisfaction with financial situation, family relations, and living conditions were independently associated with higher level of DASS-Stress score (F = 5.297, p < 0.001, R2 = 16.7%, Adjusted R2 = 13.6%) (Table 4).

DASS-Anxiety score

Factors that were univariately associated with higher DASS-Anxiety score were younger age, female gender, lower educational level, not being a doctor, being a doctor non-psychiatrist, living without a partner, with or without smaller number of children, having at least one child under 18 years, being frontline doctor and/or nurse, having close contact that had/died due to COVID-19, previous physical and mental disorder, as well as, dissatisfaction with financial situation, family relations, and living conditions. After multivariate linear regression modelling, female gender, previous mental disorder, and dissatisfaction with financial situation, family relations, and living conditions were independently associated with higher level of DASS-Anxiety score (F = 3.987, p < 0.001, R2 = 15.1%, Adjusted R2 = 11.3%) (Table 4).

Tables summarizing significant predictors of negative emotional statesare presented in Supplementary Material 2 (Table S2).

Discussion

Our study is one of the largest studies in the region that have investigated mental health issues during the first year of the COVID-19 pandemic outbreak. The study identified several predictors of emotional distress, in relation to sociodemographic, partnership, life satisfaction variables, being a healthcare professional, health-related and pandemic related factors. More importantly, the results indicate that previous psychiatric disorder, being a doctor but not a psychiatrist, dissatisfaction with various aspects of life, as well as female gender, were significant independent predictors of higher levels of depression, anxiety and stress taken together. Namely, female gender was frequently identified as one of the risk factors related to elevated levels of stress, anxiety and depression during COVID-19 pandemic, both in general and HCP population2,6. On the other hand, several studies underline the presence of gender disparities in the emotional distress, with HCP women being more vulnerable compared to their male colleagues, but this tendency is variable and dependent on socio-cultural context5,47,48. Overall, evidence shows that compared to men women were at higher risk for depression, anxiety and posttraumatic stress disorder (PTSD) during the pandemic, in general population49,50, as well as among HCP5. However, a study from Ireland observed that men were at higher risk for PTSD than women51, while a Greek study found no significant difference in stress levels between male and female nurses52. In our study female gender was a predictor of elevated scores of depression, anxiety and stress. Our sample consisted mostly of middle aged, highly educated women. These sociodemographic factors may have influenced the results, especially when considering studies exploring the way these aspects may contribute to the increased affective vulnerability such as stress, depression and anxiety8,9. According to the research, higher education, face to face jobs, reduced income during pandemic, and living alone or with under-aged children are associated with similar mental helath challenges4,49,53,54,55. Additionally, it is important to take into account that women score lower on emotional stability than men, with scores increasing with age56. Thus, research performed by Margetić et al.11, which shares similarities with ours, identified similar main predictors (less emotional stability, higher agreeableness, avoidant coping mechanism, and lack of active coping and perceived social support) of emotional distress during COVID-19 pandemic in Croatia. Furthermore, it should be kept in mind that female participants, apart from challenges caused by different professional obligations during the pandemic, have also been confronted with overwhelming personal duties (caring for others such as elderly members, children, etc.). This consequently led to having fewer opportunities to continue their professional and academic careers which, in turn, could have amplified psychological burden13,14,15,16. Studies also suggest a powerful influence of relationship status on emotional state, especially during health crises such as COVID-19 pandemic22,23. Mirroring these findings, in the current study, living with a partner was a strong predictor of elevated stress levels. Moreover, dissatisfaction with financial situation, family relations, or living conditions was associated with worsening of symptoms in all explored domains of emotional distress. These findings are in line with several studies supporting that relationship status, living arrangements and financial well-being could be associated with psychosocial problems such as anxiety, stress, depression, and loneliness9,22,23,26,57,58,59,60.

Furthermore, in our study, individuals who reported a previously diagnosed psychiatric disorder were at significant risk of having higher levels of depression, anxiety and stress measured by DASS scale. Previous studies further confirm this and report on the increased risk of psychological disturbances in those with previous somatic or psychological burden61,62. Although prior somatic illness was explored as potential factor possibly influencing the dimensions of affectivity, only previously diagnosed psychiatric disorder was confirmed as the significant predictor. Further research should more thoroughly explore the impact of pre-existing chronic, particularly somatic illnesses, and a history of specified mental disorders. This is also underlined by the studies showing that patients with chronic illnesses, as well as those with mental disorders are at higher risk for various mental outcomes and that they had the increased need for mental health service during the pandemic era21,63. It is necessary to take into account “time window” in which the research was conducted (from the end of pronounced containment measures such as obligatory quarantine throughout the first year of pandemic); the impact of limited social interactions and lack of social engagement (restrictive measures)that could have additionally provoked the feelings of loneliness, insecurity, uncertainty; and lastly, the individuals who had pre-existing mental health condition could have been particularly vulnerable to these measures59,64,65. It is also to note that the affective disturbances could be associated with COVID-19 related variables, similarly to our findings (close contact with contracted COVID-19 or death from COVID-19 were identified as strong predictors for increased depression score). A recent large study performed by Fountoulakis et al.66 reported findings in HCP similar to those reported earlier in the general population, although rates of depression and suicidal tendencies were lower. Moreover, aforementioned research shed light on importance of previous history of mental disorder in the context of clinical expressions. Consistent with these findings, Maric et al.41 reported that individuals with a prior diagnosis of a mental disorder experienced poorer mental health outcomes during the COVID-19 pandemic. Recent systematic review by Ahmed et al.67 reported diverse findings regarding changes in symptoms among people with pre-existing mental health conditions before, during and across different time points of pandemic. This comprehensive review identified no clear pattern of change in symptom severity with most studies showing either no significant change or varying outcomes depending on the specific condition with the absence of a clear worsening of most symptoms. Despite generally low certainty of evidence, the review revealed a significant worsening of PTSD symptoms and mixed outcomes for schizophrenia and bipolar disorder, particularly in the early stages of pandemic. Comparisons across time points revealed stability in depressive, schizophrenia, and bipolar symptoms, and mixed findings for OCD, anxiety, and eating disorders.

Mental health burden and overall functioning among HCPcontinues to attract attention10,29. Some evidence supports differences in mental health vulnerability among different groups of HCP and educational level, different job requirements (long shifts, workload, changeable information, insufficient personal protection equipment, etc.) job title (nurses vs. doctors), department category (frontline, second line, etc.), even different specialty among physicians could be associated with different mental health burden6,20,29,68,69,70,71. Despite the importance of observing intersectional differences, only a scarce number of papers studied the differences between psychiatrists and other specializations in relation to psychological alternations. In our sample, being a psychiatrist seems to have advantages when facing emotional distress during a health crisis. This is in line with previous studies18 which showed differences between psychiatrists in regard to COVID-19 anxiety scores. However, it is to be noted that the psychiatrists were more prone to risk of substance abuse. On the other hand, recent study performed by Tsionis et al.32 finds that mental health workers show lower rates of depression and anxiety compared to other healthcare workers and the general population speaking in favour of their possibly higher resilience. However, factors such as younger age, female gender, profession, work setting, fear of COVID-19, and workload were all linked to more psychological distress. In the study conducted in Germany, psychiatrists ranked midway between general practitioners and surgeons in terms of anxiety, while surgeons were considered to be the best informed and rated their resilience highly37. A recent study among psychiatrists in China38 found that they experienced significant levels of depression, anxiety, and burnout during COVID-19 pandemic. This research also identified other factors that can affect the results including income, hospital type (secondary) and occupational experience which elevated risk of depressive symptoms; while having a children elevated a risk for anxiety. A study by Mitkovic et al.72 revealed that psychiatrists, unlike neurologists, showed a reduced external LOC, indicating a lesser tendency to attribute control to external agents. Additionally, the uniformity of LOC among psychiatrists suggests a higher degree of consistency within this specialty, highlighting a distinct correlation between this psychological phenomenon and the psychiatric profession.

Additionally, other studies observed45, that psychiatrists were identified to be less anxious than other physicians, without differences in levels of depression and burnout73. In our study, being a doctor but not a psychiatrist was a solid predictor of elevated stress and depression. Even though psychiatrists composed only a third of doctors in our sample, the findings still indicated that being a psychiatrist has its positive sides in case of catastrophic events. Alternatively, this may reflect different coping mechanisms and psychological flexibility which previously was singled out as a more effective coping strategy in the context of well-being and individual capacity to cope with acute and long-term challenges74,75,76,77.On the other hand, other specialists could have been more involved in direct contact with patients that suffered from more severe clinical forms of COVID-19 infection and were more often required to make challenging decisions, relying on limited resources and dealing with large patient populations which could have contributed to psychological and psychiatric symptoms78.

Limitation and strengths

Although our findings support previous studies on the psychological burden of COVID-19, a few limitations are to be considered. Firstly, our study used cross-sectional approach, non-random sampling method and an online survey taking more exploratory rather than confirmatory line. Also, the psychometric instruments used in the study offer the information regarding severity of particular psychological symptoms and cannot be considered diagnostic tools. However, it is to note that we used self-rating psychometric instruments as direct, diagnostic evaluation by an independent rater was not possible due to epidemiological reasons. Also, detailed information about pre-existing mental and somatic health are lacking, for the same aforementioned epidemiological reasons. Our sample was comprised of predominantly female participants, included individuals who were educated and recruited only participants with access to the Internet which could have affected the characteristics of the participants and paused a selection bias.

The study has some significant strengths that should also be taken into account. They include a large sample sizeand the multivariate exploration of the various sociodemographic, social, work, health and COVID-19-related predictors in general population. Moreover, the study provided additional data about vulnerable individuals among HCP and contributed to the scanty number of studies that considered advantages of being a psychiatrist during catastrophic occasions regard to other medical specialists, especially in the context of increasing need for mental health services in the post-COVID era. Despite its limitations, recent study fulfilled regional literature in abovementioned background.

Conclusion

In spite of the limitation, our findings indicate several predictors of emotional distress during first year of COVID-19 outbreak in Serbia, such as female gender, previous psychiatric disorder, being a doctor but not psychiatrists, and dissatisfaction with various aspect of life. Accordingly, findings may be of theoretical and practical importance. From theoretical viewpoint, it is clear that there is a great need to consider mental health functioning in the context of some specific vulnerable population groups in order to create effective strategies to prevent further psychological burden. From practical perspective, our research provides a basic platform for more detailed and follow-up analyses of the impact of the pandemic on mental health in the region. Findings suggests that vulnerable categories such as women, those dissatisfied with various aspects of everyday life, with accompanying previous psychiatric condition, represent populations at elevated risk for depression, stress and anxiety symptoms. On the other hand, special attention should be paid to gender responsive crisis awareness policy in order to achieve already challenging work–life balance in female population. Special attention should also be payed to medical professionals, especially those who have worked in direct contact with patients (frontline workers) and who are not psychiatrists. Our results aim to prioritize the role of psychiatrists both through timely appropriate psychosocial interventions in the community setting, but also among colleagues in terms of expanding the repertoire of healthy coping mechanisms. Equally important is the strategic planning and recruitment of psychiatric personnel, as well as the education and support of other specialists.

Further research should carefully distinguish changes in mental health functioning directly connected to pandemic from long term consequences of pandemic. More refine studies of groups at risk both in general population and among people who suffer from pre-existing mental or somatic health problems could also be appreciated. The studies shouldalso include the low income and middle income regions, to provide more detailed overview of how the pandemic affected mental and physical health, as well as to evaluate presence or gaps in available psychosocial support in the community, in crisis.