Comparison of the Psychological Impact of COVID-19 on Healthcare Workers between 2022 and 2023 in a Romanian COVID-19 Hub Hospital

: We assessed depression, anxiety, and stress in healthcare workers (HCWs) in 2023 and the evolution of depression in 2023 compared with 2022. In September–November 2023, 181 HCWs from the Infectious Diseases Hospital, Cluj-Napoca, Romania, completed the Depression, Anxiety, and Stress 21 Scale (DASS-21 R), the Patient Health Questionnaire 9 (PHQ-9), and the Hamilton Anxiety Rating Scale (HAM-A). The prevalence of moderate-to-severe levels was 13.8% for DASS-Depression, 18.8% for DASS-Anxiety, 16.1% for DASS-Stress, 19.9% for PHQ-9, and 18.2% for HAM-A. Moderate-to-severe PHQ-9 depression was significantly lower in the 2023 study group (19.9%, N = 181) compared with 2022 (30.6%, N = 114) ( p = 0.04), also within the 2022–2023 follow-up participants (N = 88) ( p = 0.03). We did not find significant statistical differences between those infected vs. non-infected, vaccinated vs. non-vaccinated, or working with vs. not working with SARS-CoV-2 infected patients. Males were significantly more stressed ( p = 0.02) and anxious ( p = 0.03), and physicians in training had the highest prevalence of moderate-to-severe depression (31.6%), followed by physicians (25.6%). In 2023, depression, anxiety, and stress symptoms decreased but remained within a concerning range. By addressing these psychological issues, we may reduce the possibility of medical errors and prevent professional crises in the healthcare system.


Introduction
When it comes to their mental health, healthcare workers (HCWs) all over the world have found the COVID-19 epidemic to be problematic.Several studies found that HCWs were more likely to experience negative psychological symptoms such as depression, anxiety, fear, stress, burnout, post-traumatic stress disorder (PTSD), and stigmatization [1][2][3][4][5].
Despite HCWs' strong social and professional identities that are based on experience and preparation to deal with intense emotional and cognitive demands (resilient personality or hardiness), the lack of knowledge, preparedness, and consistent protocols of action has led to demotivation and mental health issues such as increased stress, anxiety, or depression [2].
At the onset of the pandemic, HCWs were more concerned with occupational protection, social support, and getting more rest [6,7].Following the first shock, numerous studies were published indicating an increased prevalence of burnout, depression, anxiety, and stress in frontline HCWs directly involved in the diagnosis, treatment, and care of COVID-19 patients [8][9][10][11][12][13], with an increased risk of medical errors and near-miss events, especially in pediatric wards [14].
There have been studies providing strong evidence of a high burden of anxiety and depression in HCWs during the pandemic, which have been summarized in systematic reviews and meta-analyses.Sahebi et al. (2021) showed that the overall prevalence of anxiety and depression among healthcare workers during the COVID-19 pandemic from the beginning of January to the end of October 2020 was 24.94% (95% CI: 21.83-28.05)and 24.83% (95% CI: 21. 41-28.25),respectively [18].Fernandez et al. (2021), in a systematic umbrella review conducted in accordance with the Joanna Briggs Institute (JBI) methodology on 169,157 HCWs from 35 countries in March 2021, found that the prevalence of anxiety ranged from 22.2% to 33.0% and the prevalence of depression ranged from 17.9% to 36% [19].
There have been few longitudinal studies worldwide on the evolution of the psychological impact of the COVID-19 pandemic on HCWs.Using the Depression, Anxiety, and Stress 21 (DASS-21) scale, a study from Turkey found that HCWs experiencing the pandemic waves had significantly higher levels of depression, anxiety, and stress during the second peak (November-December 2020) compared with the first peak (April-May 2020) (p < 0.001) [15].
Another study comparing the effects of the first wave (April 2020) and the second wave (December 2020) on 152 intensivists from a COVID-19 hub hospital in Rome found that the prevalence of workers manifesting symptoms of depression on the Goldberg Anxiety and Depression Scale (GADS) significantly increased from 49.4% to 63.6% (p = 0.028), while the number of workers manifesting symptoms of GADS anxiety increased, but not significantly, from 49.4% to 63.6% (p = 0.442) [16].
Following the confirmation of the first positive case on 26 February 2020, the Romanian government implemented a number of formal preventive measures.Thus, two years after the pandemic started, Romania had 3,043,205 reported cases of COVID-19 infection, including 65,906 (2.2%) deaths [20].Despite the previously implemented procedures, which proved effective in the first three waves, the fourth (September-November 2021) and fifth (January-March 2022) waves severely impacted Romania.Our World in Data reported in the Coronavirus Pandemic Country Profile for Romania close to 35,000 daily infections, with more than 1000 daily deaths, by mid-December 2021 [21].
Previous research has indicated that COVID-19 generated a substantial level of distress among emergency department (ED) staff, with three major factors contributing to the distress level: loss of control or vulnerability, worry for self-health, and virus spread [22].It was assumed that COVID-19 would have a greater psychological impact on medical personnel than on the general population, but unlike other countries, there have been few studies in the Romanian general population.Vancea and Apostol (2021) published a report about mental health levels at the end of the first wave of the pandemic from 9 May to 15 May 2020 (t1) and amidst the third wave from 8 November to 14 November 2020 (t2).At t1, 23.8% of participants were anxious and 19.2% were depressed, and at t2, 25.7% of participants were anxious and 17.8% were depressed, with no statistical differences between t1 and t2 [23].
There have been several studies on the prevalence of mental problems among HCWs in Romania, although most of them have been cross-sectional rather than longitudinal.Burnout was the most investigated topic during the early phases of Romania's COVID-19 pandemic.Accordingly, a study released in 2022 on the prevalence of burnout syndrome in an infectious diseases hospital in Constanta, Romania, following the first year of the pandemic revealed that 69.2% suffered moderate to severe Maslach burnout [24].In addition to burnout, there have been some reports regarding depression, anxiety, and stress in the late stages of the COVID-19 pandemic.At the Teaching Hospital of Infectious Diseases in Cluj-Napoca, Romania, another institution of the same type, our team reported a 71.4% incidence of moderate to high Maslach burnout and a 63.1% prevalence of mild to severe depression after two years of the pandemic (May-June 2022), between the fifth and sixth pandemic waves [25].Also, our team investigated which psychological characteristics were more strongly connected with burnout and depression in COVID-19 healthcare professionals.We discovered that emotional exhaustion was a significant predictor of both personal and work-related burnout and depression, implying overlapping interference, and that reduction in personal achievement was only a significant predictor of depression, not of burnout, emphasizing its importance for depression [26].
In the same period (spring of 2022), it was reported that there were lower levels of anxiety, depression, and stress in different types of public hospitals in Timisoara, Romania, among 300 healthcare professionals who were assessed using the Beck Anxiety Inventory (BAI), the Beck Depression Inventory (BDI), and the Perceived Stress Scale (PSS-10) (47.7% prevalence of mild anxiety, 33% prevalence of mild depression, and 65.3% prevalence of moderate stress).The severity of anxious, depressive, and stress symptoms differed considerably depending on the professional degree, the frequency of on-call duties per month, the type of medical unit where the participants worked, and the number of SARS-CoV-2 patients treated [27].
In the previous study, taking place between the fifth (January-March 2022) and sixth (July-September 2022) waves of the COVID-19 pandemic in Romania, when the daily numbers of confirmed cases decreased significantly to under 5000, we found a high prevalence of mild to severe depression (63.1% of HCWs) [25] Five months after the last wave, in September 2023, we initiated a study to find out whether the psychological problems of the medical staff had improved.

General Aim of This Study
The aim of this study is to investigate depression, anxiety, and stress in HCWs three years after the COVID-19 outbreak and to examine the course of depressive symptoms between 2022 and 2023 among HCWs in the Clinical Hospital of Infectious Diseases, Cluj-Napoca, Romania, a tertiary mono-specialty hospital that provided medical services for patients with COVID-19 from Cluj County and neighboring counties.We chose to assess depression, anxiety, and stress because they can have an impact on professional activity, possibly resulting in medical errors or bad outcomes that may create emotional scars, known as "second victim" syndrome [28].

Main Hypotheses
Three years following the initial outbreak of COVID-19, we anticipate that (1) the prevalence of moderate-to-severe depression, anxiety, and stress among HCWs would be much lower than reported levels in the first two years of the pandemic, and (2) the prevalence of moderate-to-severe depression would be significantly lower in 2023 compared with 2022 in the same hospital.

Study Design, Participants, and Original Data Collection
Repeated cross-sectional surveys on HCWs were performed in May-June 2022 regarding depression (2022 survey) and in September-November 2023 (2023 survey) regarding depression, anxiety, and stress.
We found a number of subjects who had participated in both surveys, allowing us to assess the evolution of their depression (2022-2023 follow-up study).
HCWs from the Teaching Hospital of Infectious Diseases in Cluj-Napoca, Romania, who voluntarily agreed to participate in the study responded to anonymous online surveys.The recruitment method was simple, with an e-mail invitation after a short presentation of the study's objectives in an online meeting explaining the purpose and the names of the investigators.
Three questionnaires were converted into online questionnaires via Google Forms to be completed: the Depression, Anxiety, and Stress 21-item scale-Romanian version (DASS-21 R), the Adult Depression Severity Scale adapted from the Patient Health Questionnaire 9 (PHQ-9), and the Hamilton Anxiety Rating Scale (HAM-A).Forms that did not meet the inclusion criteria and were not filled out properly were excluded from the study.The forms were evaluated and scored according to their instructions, recorded in the data set, and statistically analyzed.We centralized the data in an Excel file, and the results were calculated by the same investigator.
The DASS-21 is a set of three self-rating scales designed to assess negative emotional states including depression, anxiety, and stress.It was developed to capture non-clinical syndromes and ensure maximum discrimination between them [29].The Romanian version of DASS-21, standardized for the Romanian population, has 21 items divided into three scales: anxiety, depression, and stress [30].The DASS-21 R has internal consistency coefficients higher than 0.8, and each item was scored on a basic numeric scale of 0 ("does not apply to me") to 3 ("applies to me a lot or almost all the time").The DASS manual provides a detailed account of scale development, normative data, and research applications [29,30].
The PHQ-9 is a clinical depression scale with nine items that assess the severity of DSM-5 symptoms.It evaluates symptoms such as low interest, feeling down, difficulty falling or staying asleep, fatigue, poor appetite, self-doubt, and difficulty concentrating.The answers are rated on a Likert scale in four steps from 0 to 3 ("not at all" = 0, "on some days" = 1, "half of the days" = 2, "almost every day" = 3) [31].The PHQ-9 is valid and reliable, with high internal consistency and good sensitivity and specificity [32,33].It has been reported to have a sensitivity of 0.80 and a specificity of 0.92 [34].
The Hamilton Anxiety Scale (HAM-A) is a 14-item instrument developed over 40 years ago to assess anxiety symptoms and severity [35].It is the most commonly used outcome measure in clinical trials and can be pooled for meta-analyses [36].The HAM-A has demonstrated adequate reliability, validity, and sensitivity, with internal reliability coefficients of 0.89 [35] and 0.74 [37] and a high internal consistency (alpha = 0.85) [38].It consists of 14 symptom-defined elements relating to both psychological and somatic symptoms, including anxious mood, tension, fears, insomnia, intellectual effects, depressed mood, somatic symptoms, sensory, cardiovascular, respiratory, gastroenterological, genitourinary, and autonomic features, and observed behavior at interview.HAM-A can be used as a self-report online questionnaire and is available at www.mdcalc.com[39].It is still in use today, being considered one of the most widely used rating scales [40].Each item was scored on a basic numeric scale of 0 (not present) to 4 (severe) [40].A validation of a computer-administered version of the HAM-A was realized by Kobak et al. in 1993 [41].
The data collection process was performed over the Internet through an online questionnaire program (Google Forms).

Ethical Considerations
The responses to the online questionnaires were recognized as implicit consent.The participants agreed to share their sociodemographic data (age, sex, profession, history of COVID-19, SARS-CoV-2 vaccination, and working with COVID-19 patients).

Data Analysis
Demographic and questionnaire data were analyzed using various statistical techniques, with the assistance of IBM SPSS Statistics 23.0 (IBM, Armonk, NY, USA).Variables measured at a nominal scale were represented using proportions (%), and scale variables were presented as the median (M) and interquartile range (IQR).The normality of the variables was tested using the Shapiro-Wilk test.To highlight the difference in the number of participants between the different groups, we used a t-test with bootstrapping.For comparing the continuous variables between non-parametric groups, we used the Mann-Whitney test (between two groups) and Scheffe Multiple Comparisons with One-Way Anova (between more than two groups).
A p < 0.05 value was considered statistically significant.
When we analyzed the prevalence and the range values, we found a prevalence of 13.8% for moderate-to-severe/extremely severe DASS-depression, 18.8% for moderateto-severe/extremely severe DASS-anxiety, and 16.1% for moderate-to-severe/extremely severe DASS-stress.We also found a prevalence of 19.9% for moderate-to-severe/extremely severe clinical depression (PHQ-9), and 18.2% for moderate-to-severe/extremely severe clinical anxiety (HAM-A).The total median (IQR) scores for each scale and the prevalence of each clinical category are presented in Table 2.When we looked at the prevalence and range values of depression, anxiety, and stress severity for different professions in the 2023 study, we found that moderateto-severe/extremely severe values were present among physicians in training, infectious diseases physicians, and nurses.The most relevant values were for moderate-tosevere/extremely severe depression on the PHQ-9 scale in physicians in training (31.6%), infectious diseases physicians (25.6%), and nurses (18.2%).In terms of moderate-to-severe anxiety on the HAM-A scale, we found a higher level in physicians in training (23.7%).The prevalence of depression, anxiety, and stress and the range values for different professions in the 2023 study are presented in Table 3.When we compared DASS-depression, DASS-anxiety, DASS-stress, PHQ-9, and HAM-A between the female and male groups, SARS-CoV-2 history of infection vs. noninfection groups, SARS-CoV-2 vaccination vs. non-vaccination groups, working vs. not working with COVID-19 patients groups, and between various professionals, we found significantly higher DASS-stress and HAM-A in men than in women, significantly higher DASS-stress in physicians in training vs. nurses, and also significantly higher PHQ-9 values in physicians in training vs. professional caregivers.We used the Mann-Whitney test (between two groups) to compare continuous variables between non-parametric groups, as well as Scheffe Multiple Comparisons using One-Way Anova.Table 4 shows comparative DASS-21 R, PHQ-9, and HAM-A scores among different groups of the 2023 participants (N = 181).In bold-results with p-value ≤ 0.05.M, median value; IQR, interquartile range; X 2 , Chi-squared test used to compare differences between categorical variables; p, significance level; a -comparison of proportions; b -Mann-Whitney non-parametric test used to compare two groups; c -Scheffe Multiple Comparisons with One-Way Anova.
There were no significant differences between men and women.The PHQ-9 scores (IQR) and the comparative analyses among 2023 (N = 181), 2022 (N = 114), and 2022-2023 follow-up participants (N = 88) are presented in Table 5.In bold-results with p-value ≤ 0.05.F, female; M, male; M, median value; IQR, interquartile range; X 2 , Chisquared test used to compare differences between categorical variables; p, significance level; a -comparison of proportions; b -Man Whitney non-parametric test used to compare two groups.

Discussion
The long duration of the COVID-19 pandemic has exposed frontline HCWs to an unprecedented strain [16].From a psychopathological perspective, the epidemic is a relatively new type of stressor or trauma for mental health practitioners [8], comparable to natural disasters such as earthquakes or tsunamis [42].
Many studies reported the prolonged negative effects of the COVID-19 pandemic on HCWs' mental health in the first year [15][16][17].
There has been a lack of longitudinal research examining the long-term stress impacts of the COVID-19 epidemic on the mental health of HCWs.Most of these studies were undertaken in the United States [43,44].Shanafelt et al. (2022) reported that the levels of psychological distress in the second year of the pandemic (2021) continued to be elevated among HCWs, according to a survey of US physicians between December 2021 and January 2022, approximately 21 months after the epidemic began.In 2021, 62.8% of physicians reported burnout symptoms, up from 38.2% in 2020 (p < 0.001).Satisfaction with work-life integration decreased from 46.1% in 2020 to 30.2% in 2021 (p < 0.001), and depression scores increased by 6.1% (2020 mean, 49.54; 2021 mean, 52.59; p < 0.001) [44].
Furthermore, in Romania, few studies have followed the evolution of depression, anxiety, and stress among healthcare workers due to COVID-19 work-related stress [27].
Therefore, the purpose of our study was to analyze the prevalence of depression, anxiety, and stress after three years of the COVID-19 outbreak and to examine the course of depressive symptoms between 2022 and 2023 among HCWs in the Clinical Hospital of Infectious Diseases, Cluj-Napoca, Romania, a tertiary mono-specialty hospital that provided medical services for patients with COVID-19 from Cluj County and neighboring counties.
In terms of depression, 13.8% of respondents reported moderate to severe DASSdepression, while 19.9% reported moderate to severe PHQ-9 clinical depression (Table 2), which was close to the lowest prevalence found in a systematic umbrella review of global evidence carried out by Fernandez et al. (2021) [19].They found that the prevalence of depression among all HCWs ranged from 17.9% to 36% in ten systematic reviews (100 unique studies) comprising 169,157 HCWs from 35 countries.
Our study suggests that three years after the COVID-19 outbreak, the prevalence of HCWs with moderate-to-severe depression symptoms is decreasing but remains within the concerning range reported in the umbrella review.
Since the most concerning mental health problems are severe-to-extremely severe depression, anxiety, and stress, we focused on physicians in training and infectious disease physicians.Using the DASS-21 scale for physicians in training, we found the following prevalences of severe-to-extremely severe depression, anxiety, and stress: 10.5% for DASSdepression, 5.2% for DASS-anxiety, and 7.7% for DASS-stress, respectively.For infectious disease physicians, we found prevalences of 2.6%, 2.6%, and 12.8%, respectively (Table 3).In the pre-pandemic period, Hayes et al. (2017), using the same scale, found in a great number of physicians (N = 1749) a higher prevalence of severe-to-extremely severe symptoms of depression (7.2%) and anxiety (6.1%) but lower levels of stress (9.5%) [45].The disparities can be explained by the participants' professional profiles, as physicians in Hayes' study represented all hospital specialties.
We found near-similar values for M (IQR) for DASS-depression and PHQ-9 clinical depression (Table 2); the insignificance of the difference may be explained by the correctness of the answers and the good reliability of both scales [29,30,[32][33][34].The PHQ-9 scale has become increasingly popular over the past decade for detecting major depressive disorders in various clinical settings [25,[46][47][48][49].
Given a global maximum of 10% for 12-month prevalence of depression in the general population [50,51], our 2023 findings revealed a 1.3 (DASS) and 2 (PHQ-9)-fold higher prevalence of depression in HCWs, even three years after the pandemic began.
Regarding anxiety, 18.8% of respondents indicated moderate-to-severe DASS-anxiety, and 18.2% expressed moderate-to-severe HAM-A clinical anxiety (Table 2).Our results showed lower values than those reported by Fernandez et al. (2021), who found 22.2-33.0%prevalence of anxiety [19].According to the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, 2013, the reported prevalence of general anxiety disorder in Europe was 0.4-3.6% in the general population [51].Using the HAM-A scale, which mostly evaluates symptoms of general anxiety disorder [38], we found the prevalence of anxiety to be 5-6-fold greater than in the general population.
Regarding stress symptoms, 16.1% of the respondents reported moderate-tosevere/extremely severe DASS-stress manifestations (Table 2) as a result of psychological or physical demands that could not be reasonably managed.Symptoms of stress meant that the respondents did not prectice good stress management, as reported by Kushal et al. (2018) [52].Stress factors in HCWs have been well documented, mostly during the COVID-19 pandemic [53][54][55].In comparison with a 41.2% prevalence of moderate-to-severe stress manifestations reported by Elbay et al. (2020) for 442 physicians in Türkiye, using the same DASS scale, our results showed a 2.5-fold lower value [56].
The median M (IQR) values for DASS-depression, DASS-anxiety, DASS-stress, PHQ-9, and HAM-A did not indicate psycho-emotional disorders for the entire group; however, we did find variances amongst professional groups.Infectious disease physicians and physicians in training reported the greatest degree of DASS-depression and PHQ-9 clinical depression (Table 4); physicians in training exhibited the highest levels of DASS-stress and HAM-A anxiety, with significant differences between physicians in training vs. nurses for DASS-stress (p = 0.03) and between physicians in training vs. professional caregivers for DASS-stress (p = 0.01) and PHQ-9 (p = 0.03).These results were consistent with studies that emphasized the emotional impact of COVID-19 on physicians in training with less work experience [25,56,57].In a systematic review and meta-analysis regarding the prevalence of depression and depressive symptoms among resident physicians, Mata et al. (2015) found lower prevalence of moderate-to-severe PHQ-9 depression (20.9%) in 2015, compared with the 31.6%prevalence observed in our study (2023).The explanation could be the increased and ongoing stress caused by the COVID-19 pandemic [58].
Males presented much more DASS-stress (p = 0.02) and much more HAM-A (p = 0.03) in comparison with females.Even though most relevant studies have reported considerably more psychological issues in females [56,[59][60][61], our findings might be explained by the higher participation of male physicians in training with perceived psychological issues.Accordingly, this aspect requires further inquiry.
The decreasing trend of moderate-to-severe PHQ-9 clinical depression measured with PHQ-9 between the 2022 and 2023 surveys (p = 0.04) and also among the 2022-2023 followup participants (p = 0.03) (Table 5) can be explained by the reduced level of stress in the third year of the pandemic or by the coping strategies [53] and posttraumatic growth and resilience at the individual and organizational levels [62].Many studies have demonstrated that resilience development is associated with lower levels of depression, anxiety, and burnout [62][63][64][65][66].

Strengths and Limitations
Our study's strength is that we used three separate scales to examine HCWs' psychological responses to the pandemic three years after its beginning, as well as repeated cross-sectional investigations to determine the progression of the COVID-19 pandemic's psychological impact on the most exposed HCWs during the pandemic.
There were several limits to our study.Firstly, and foremost, it was conducted in a single center, with participants exclusively from a hospital in Cluj-Napoca; therefore, the results might not actually depict the entire situation for HCWs in Romania.Secondly, our sample size was rather small (181 out of 818 hospital employees-22.1%),which is a classical issue that was found in another study [48] with the same design.The non-response rate and how selection bias affected the survey findings are classical difficulties that might have led to underestimation, which may be explained by 'pandemic fatigue', according to the WHO Regional Office [67], and by the lack of organizational culture regarding psychological aspects.However, compared with other studies on Romanian HCWs during the COVID-19 pandemic [55,65], the participation of front-line COVID-19 HCWs in this psychological study was not significantly different.Thirdly, the study variables were measured by a self-reported questionnaire, which could have had an impact on the results as a consequence of common method variance bias and also because the cross-sectional methodology excluded inferences about causal relationships.

Future Directions
Further studies are needed to determine whether personal and organizational resilience played a role in the psychological improvement associated with the negative impact of the COVID-19 pandemic.Organizations should implement interventions to promote resilience and adaptive coping, such as counseling services, social connection initiatives, and focused training, with the aim of developing a positive mindset.

Conclusions
Three years after the COVID-19 outbreak, the prevalence of depression, anxiety, and stress symptoms decreased but remained within a concerning range, mainly among physicians in training, followed by physicians, with males having much more stress and anxiety in comparison with females.Factors like history of COVID-19 illness, SARS-CoV-2 vaccination, and work with COVID-19 patients did not have significant influences on psychological health.
By addressing these psychological issues, we may reduce the possibility of medical errors and professional crises in the healthcare system.
. Prior to September 2023, Our World in Data, in its Coronavirus Pandemic Country Profile for Romania (available online), reported another two peaks: one on 28 August 2022, with 40.000 new daily confirmed COVID-19 cases, and another on 16 April 2023, with 6400 new daily confirmed COVID-19 cases [21].

Table 1 .
Demographic and working characteristics of the 2023 and 2022 surveys and 2022-2023 follow-up participants.

Table 3 .
Prevalence of depression, anxiety, and stress for different professions in the 2023 study (N = 181).