Occurrences of post-traumatic stress disorder, anxiety, depression, and burnout syndrome in ICU staff workers after two-year of the COVID-19 pandemic: the international PSY-CO in ICU study

Purpose The present study aimed at assessing the prevalences of post-traumatic stress disorder (PTSD) (main objective), anxiety, depression, and burnout syndrome (BOS) and their associated factors in intensive care unit (ICU) staff workers in the second year of the COVID-19 pandemic. Materials and methods An international cross-sectional multicenter ICU-based online survey was carried out among the ICU staff workers in 20 ICUs across 3 continents. ICUs staff workers (both caregivers and non-caregivers) were invited to complete PCL-5, HADS, and MBI questionnaires for assessing PTSD, anxiety, depression, and the different components of BOS, respectively. A personal questionnaire was used to isolate independent associated factors with these disorders. Results PCL-5, HADS, and MBI questionnaires were completed by 585, 570, and 539 responders, respectively (525 completed all questionnaires). PTSD was diagnosed in 98/585 responders (16.8%). Changing familial environment, being a non-caregiver staff worker, having not being involved in a COVID-19 patient admission, having not been provided with COVID-19-related information were associated with PTSD. Anxiety was reported in 130/570 responders (22.8%). Working in a public hospital, being a woman, being financially impacted, being a non-clinical healthcare staff member, having no theoretical or practical training on individual preventive measures, and fear of managing COVID-19 patients were associated with anxiety. Depression was reported in 50/570 responders (8.8%). Comorbidity at risk of severe COVID-19, working in a public hospital, looking after a child, being a non-caregiver staff member, having no information, and a request for moving from the unit were associated with depression. Having received no information and no adequate training for COVID-19 patient management were associated with all 3 dimensions of BOS. Conclusion The present study confirmed that ICU staff workers, whether they treated COVID-19 patients or not, have a substantial prevalence of psychological disorders. Supplementary Information The online version contains supplementary material available at 10.1186/s12991-023-00488-5.

The COVID-19 pandemic presented with different intensities between countries.Therefore, some countries tried to fight and/or delay the start of the pandemic to reduce the peak infection rates of the disease.These actions aimed at reducing the overall pressure on national healthcare systems and was intended to decrease the COVID-19 mortality rate [9,10].
Based on the experience of previous pandemics, countries reacted by applying different transmission prevention strategies to prevent or delay the spread of the disease [9][10][11].Therefore, measures such as border closure, school closure, restricting social gatherings (even shutdown of workplaces), limiting population movements, and lockdowns at the scale of cities or regions were put into action.In public hospitals, several measures were implemented to concentrate care resources on the potential wave of admissions of patients with severe forms of COVID-19.For this reason, the number of available beds in the ICU was frequently increased by up to two-fold [12,13], and scheduled nonemergency surgical procedures were canceled.Frequently underutilized health care professionals (physicians such as anesthesiologists, and nurses of other units) were transferred to ICUs, and those of less busy units were transferred to busier ones.
All these measures lead to major daily-life changes that could be stressful to individuals.In the general population, it has been well documented that quarantine or confinement, or isolation may lead to the occurrence of post-traumatic stress disorder (PTSD) in about 30% of the exposed population [14].Importantly, high levels of depressive symptoms have been reported in up to 9% of hospital staff [15].Numerous symptoms, such as emotional disturbance, depression, stress, low mood, irritability, insomnia, and post-traumatic stress symptoms have been reported after quarantine or isolation [14].
In the ICU setting, it has been shown that the COVID-19 pandemic led to psychological consequences on caregivers.During the second wave in France (autumn 2020), Azoulay et al. reported symptoms of anxiety, depression, post-traumatic stress disorder, and burnout in 60.0%, 36.1%,28.4, and 45.1%, respectively, in 845 health care providers (66% nursing staff, 32% medical staff, 2% other professionals [16]).However, because the pandemic has continued over a prolonged period, with potentially different impacts on the population and healthcare systems, and varying in intensity according to the vaccination rate, the present study aimed at assessing the occurrence of PTSD, anxiety, depression, and burnout syndrome (BOS) in ICU staff workers in Australia (Queensland), France and Hong Kong after the first year of the COVID-19 pandemic.The primary objective was to assess the prevalence of PTSD in ICU staff workers.The secondary objectives were to identify potential associated factors to the occurrence of PTSD and to assess the prevalence of anxiety, depression, BOS, and their related associated factors in the same cohort.

Design
An international cross-sectional multicenter (20 centers) ICU-based online survey was carried out among ICU staff workers in Australia, France, and Hong Kong.
According to French law, this study does not involve patients and is considered a quality-of-care assessment [17].Therefore, the Institutional Review Board of the Nîmes University Hospital (# 20.05.08) and of the French Society of Anesthesia and Critical Care (IRB 00010254-2020-148) gave their approvals.This study was registered on ClinicalTrial.gov (NCT04511780 first posted on August 13, 2020) before the inclusion of the first participant.In Australia and Hong Kong (SBRE (226-20)), the local ethics committees of each institution gave study approval.
Around the time of the survey administration, in Hong Kong and France there were significant numbers of COVID related admissions to the ICUs, whereas at Royal Brisbane and Women's Hospital in Brisbane, Australia, COVID-19-related ICU admissions occurred post survey only.
The survey included 5 different questionnaires: 1) The center demographic questionnaire that focused on the nature and organization of the ICU: - 2) The individual demographic questionnaire that collected personal information: -Personal socio-demographic data and their changes during the pandemic; -Professional characteristics (job title, experience), their experience during the COVID-19 pandemic (feeling, family, and professional relationships); 3) Validated questionnaires for assessing PTSD (PCL-5) [18] 4) Hospital Anxiety Depression Scale (HADS) for assessing symptoms of anxiety and depression [19] 5) Maslach Burnout Inventory Human Services Survey for Medical Personnel (MBI-HSS-MP) for assessing BOS [20,21].

Study population
The principal investigators contacted ICUs in Australia, France and Hong Kong to participate.After center approval, all ICU staff workers (caregivers in contact with patients and non-caregivers) could participate in the present study.After having had the ability to read an information note about the study, responding to the questionnaire was considered to imply informed consent.The inclusion criteria were caregiver and non-caregiver staff working in the ICU during the COVID-19 outbreak and consent to complete the questionnaire.The recruitment was performed between February 25th, 2021 and June 8th, 2022.
The non-inclusion criteria were participation refusal and non-response to the questionnaire.Partially completed questionnaires were excluded.

Outcomes
The primary outcome was the prevalence of PTSD (defined by a PCL-5 score ≥ 32) and its 95% confident interval (95% CI).
The secondary outcomes were to identify potential associated factors with occurrence of PTSD and to the prevalences of anxiety and depression according to the HADS questionnaire, and burnout assessed by the MBI-HSS (MP) self-questionnaire.
Anxiety and depression were separately assessed by the HADS questionnaire according to the following rules: -0 to 7: absence of disorder; -8 to 10: suspected disorder; -11 to 21: proven disorder.
Burnout syndrome was assessed by the MBI-HSS (MP) in its 3 specific sub-scales allowing for the evaluation of emotional exhaustion, depersonalization, and personal accomplishment dimensions, respectively.However, many controversies remain unsolved for the global MBI assessment: [20,22] 1.Personal accomplishment is not always taken into account in the global MBI score; 2. In each subscale, the different thresholds are challenged.
Thus, we have analyzed the 3 sub-scores both separately and continuously.

Statistical analysis
The primary objective, i.e., to evaluate the prevalence of PTSD, was measured with the PCL-5 score and classified as probable PTSD versus no PTSD with PCL-5 scores of ≥ 32 versus < 32 with 95% confidence intervals (95% CI), respectively.The prevalence of PTSD was estimated in the total sample and in each country.
The associated factors with PTSD were searched as secondary objectives.For this purpose, we selected variables with univariate logistic regression to reduce the dimensionality of the model (relaxed alpha = 0.2) and then applied a multivariate logistic regression with backward selection (alpha = 0.05).First, the univariate analysis compared the dichotomous/categorical/nominal variables (expressed as numbers and percentages) according to PTSD occurrence by the chi-square test or the Fisher exact test when necessary.The links between the explanatory variables and PTSD variables were expressed by the odds ratios and their 95% CI by the Wald method.Covariates with a p-value ≤ 0.20 in the univariate analysis were pre-selected to perform a multivariate analysis and a backward selection strategy at the 5% threshold was applied.Adjusted Odds Ratio (AOR) was provided with 95% CI.Importantly, the prevalence of PTSD was assessed in all completed PCL-5 (n = 585) whereas the associated factors were searched in participants who completed PCL-5 AND personal life questionnaires (n = 525).
For the other secondary objectives, the same analysis strategy was applied to evaluate the prevalence of anxiety, depression on one hand, and the factors associated with these disorders on the other (using the same method used for PTSD and associated factors).A polytomous logistic regression with a proportional-odds cumulative logit model was used to search for factors associated with anxiety and depression classified in a 3-level ordinal variable.The scores of the emotional exhaustion, depersonalization, and personal accomplishment subscales were expressed as mean, standard deviation (SD), median and interquartile range (IQR).The associated factors to the 3 sub-scores were assessed with a multiple linear regression model.The same variable selection strategy was used for the previous models.Pearson correlation coefficients between PTSD, anxiety, depression, emotional exhaustion, depersonalization, and personal accomplishment scores are provided with their 95% CI.All statistical analyses used SAS statistical software, version 9.4 (SAS Institute Inc).
According to the multivariate analysis (including 525 participants who fully completed HADS and personal life questionnaires), comorbidity at risk of severe COVID-19, working in a public hospital, looking after a child, being a non-caregiver, having no information on the evolution of the pandemic, having requested a change of unit for not working in a COVID unit were associated with a greater occurrence of proven depressive disorder (Table 2).

Sub-scores of burnout Emotional exhaustion
The emotional exhaustion score in the total sample was 23.5 ± 13.7.
According to the multivariate analysis (including 525 participants who fully completed MBI and personal life questionnaires), usually living alone, being a noncaregiver, having no information on the evolution of the pandemic, not being adequately trained to manage a COVID-19 patient, not having accepted managing COVID-19 patients, and fear of managing a COVID-19 patient were independently associated with greater emotional exhaustion (Table 3).

Depersonalization
The depersonalization score in the total sample was 9.1 ± 7.0.
According to the multivariate analysis, having been infected with SARS-CoV-2 and having no information on the evolution of the pandemic were associated with a higher depersonalization score.An age > 50 years was associated with lower depersonalization (Table 3).

Personal accomplishment
The loss of personal accomplishment score in the total sample was 35.3 ± 7.9.
According to the multivariate analysis, comorbidity at risk of severe COVID-19, working in a public hospital, having no theoretical or practical training on individual preventive measures, and insufficient information about the management of COVID-19 patients were associated with lower personal accomplishment (Table 3).
Emotional exhaustion and Depersonalization scores were both correlated (r = 0.57, 95% CI [0.51-0.63],p < 0.0001), whereas the latter were negatively but less correlated with personal accomplishment (Additional file 1: Table S1).The position and dispersion parameters associated with each score are reported in Additional file 1: Table S1.

Discussion
In the present study performed in 20 centers in Australia, France, and Hong Kong, 525 ICU staff workers responded to the PCL-5, HADS, and MBI questionnaires.PTSD was present in 16.8% of participants with the highest prevalence in Hong Kong (30.8%).Anxiety and depressive disorders were reported in 22.8 and 8.8% of responders, respectively.The common associated factors with PTSD, anxiety, and depression were being a noncaregiver worker and not having been regularly informed of the COVID-19 progression during the pandemic.Concerning BOS, not having been regularly informed of the COVID-19 progression was associated with higher scores for emotional exhaustion, depersonalization, and the loss of personal accomplishment, respectively.The present study was performed during the second year of the COVID-19 pandemic in 3 different countries with different impacts of this pandemic, different strategies to prevent contamination, and different population vaccination rates.These factors could explain the different prevalences of PTSD, anxiety, and depression reported in previous studies that were essentially performed in European countries during the first and second waves.The FAMIREA group performed two studies in the first and second waves in 21 and 16 centers involving 845 (70% responders) and 1058 (67% responders healthcare professionals, respectively [16,23].The prevalences of PTSD were successively 32.0 and 28.4% with anxiety and depression reported in 50.4 to 60.0% and 30.4 to 36.1%, during the first and second waves, respectively.During the second wave, the authors reported a burnout syndrome in 45.1% using an overall score [23]. In January 2021, a single center study involving 136 healthcare workers (84 nurses, 52 physicians) in a temporary ICU during the pandemic in Milano Fiera, Lombardy reported 60% burnout syndrome, 53% anxiety (especially in nurses), and 45% depression [24].In June-July 2020, a cross-sectional study involving 709 healthcare providers from 9 English ICUs reported 40% PTSD, 11% severe anxiety, and 6% severe depression.In May 2020, a cross-sectional study involving 352 Swiss ICU healthcare workers reported 22% PTSD, 46% anxiety, and 46% depression [25].
The present study reports lower prevalences of PTSD, anxiety, and depression than the previous ones performed in the first two waves of the pandemic.Our findings could mean that the impact of COVID-19 pandemic has been blunted overtime.Indeed, the present findings are close to those observed at baseline prior to the COVID-19 pandemic [16,23,26].Another explanation could be related to different cultures, different impact of the pandemic and policies on restriction, lockdown, and vaccine strategies in Hong Kong Australia, and France [27][28][29].
The present study also reported that ICU staff workers in contact with COVID-19 patients are at lower risk of psychological consequences than those not in charge of these patients.This paradoxical phenomenon has been regularly reported in previous studies [14].Indeed, being far from the patients with no information and education about the disease could lead to fear, anxiety, stress, and other psychological consequences.The absence of information about local progression of the pandemic was also associated with BOS in its 3 dimensions (emotional exhaustion, depersonalization, and loss of personal accomplishment).
In contrast to the previous studies, a quantitative approach to BOS was performed.A threshold of MBI * According to the order of appearance of the survey forms, an imbalance in the completion rate was noted between the first questionnaire (Personal life questionnaire) and the last form (Personal and Professional questionnaire used to research the factors associated with the psychological disorders studied) (higher completion rate for the first questionnaire).To evaluate the prevalence associated with psychological disorders, all the answers filled in for each scale of evaluation of the latter were taken into account, although the questionnaire was not completed in full.For this reason, a difference in the numbers analyzed (between those for the prevalence of post-traumatic stress, anxiety, and depression and those for the analysis of associated factors) is observed (see Fig. 1).The search for factors associated with the occurrence of psychological disorders was carried out on 525 people (those who completed all the survey forms) ƚ Number of observations / total number of observations ǂ The results presented correspond to the pre-selection of variables at p-value < 20%.The second selection of variables was made at the 5% threshold and then integrated into the multivariate model § Adjusted odd ratio with a 95% confidence interval ** Not applicable   is classically used for diagnosing BOS.However, this dichotomous analysis has been challenged because MBI aggregate 3 different and independent part of the diagnosis.In 2016, the cut-off scores were removed by the MBI Manual 4th edition because they have no diagnostic validity [30].Even with this difference, the present study reported similar associated factors with the 3 different parts of BOS (lack of information about local progression of the pandemic and lack of theoretical or practical training on COVID-19 patient management).The present study highlighted several factors associated with PTSD, anxiety, depression, and symptoms of BOS.Moreover, it involved ICUs from different continents.Hong Kong was firstly impacted by the pandemic.France was also severely impacted by the first two waves with some ICU overwhelming episodes.Australia and particularly Queensland closed their borders and had limited transmission and cases in the early stages.Finally, the courses of vaccination covert were different according to the general health strategy against the COVID-19 pandemic.These differences could partly explain the heterogeneous findings of the present study.
We must acknowledge some limitations.First, the participation rate was only 16%, which is consistent with cross-sectional surveys.We did not send personal reminders to respect responder anonymity.Another reason may be the timing of our study (after the third wave, February-July 2021) that was perhaps too far from the start of the pandemic with participant weariness leading to a low response rate.The present study, therefore, likely reported the chronic states of stress, anxiety, depression, and BOS in ICU staff.Second, the cross-sectional survey design only led to isolating associated factors with PTSD, anxiety, depression, and BOS.For isolating risk factors of these psychological disorders, cohort or case-control designs might have been more appropriate.Third, the sample of the present study * According to the order of appearance of the survey forms, an imbalance in the completion rate was noted between the first questionnaire (Personal life questionnaire) and the last form (Personal and Professional questionnaire used to research the factors associated with the psychological disorders studied) (higher completion rate for the first questionnaire).To evaluate the prevalence associated with psychological disorders, all the answers filled in for each scale of evaluation of the latter were taken into account, although the questionnaire was not completed in full.For this reason, a difference in the numbers analyzed (between those for the prevalence of post-traumatic stress, anxiety, and depression and those for the analysis of associated factors) is observed (see Fig. 1).The search for factors associated with the occurrence of psychological disorders was carried out on 525 people (those who completed all the survey forms) ƚ Number of observations / Total number of observations ǂ The results presented correspond to the pre-selection of variables at p-value < 20%.The second selection of variables was made at the 5% threshold and then integrated into the multivariate model § Adjusted odd ratio with a 95% confidence interval ** Not Applicable    was not well balanced with a preponderance of French participation.Fourth, non-care giving staff was also underrepresented in this study.Finally, it is well known that the demands of working in ICUs could lead to psychological disorders such as PTSD, anxiety, depression and BOS.As no baseline assessment of these disorders was conducted before the pandemic, we cannot rule out the fact the present study reported only the baseline psychological state [26].

Conclusion
Our findings confirmed that ICU staff workers continue to suffer from psychological disorders.Even if some factors are linked to the COVID-19 pandemic (fear of managing COVID-19 patients), the lack of theoretical and practical training in the management of COVID-19 patients as well as the lack of information on the current status of the pandemic within the ICU were associated with a higher prevalence of PTSD, anxiety, depression, and BOS.These findings suggest the importance of good communication amongst staff in the ICU for staff wellbeing.* According to the order of appearance of the survey forms, an imbalance in the completion rate was noted between the first questionnaire (Personal life questionnaire) and the last form (Personal and Professional questionnaire used to research the factors associated with the psychological disorders studied) (higher completion rate for the first questionnaire).To evaluate the prevalence associated with psychological disorders, all the answers filled in for each scale of evaluation of the latter were taken into account, although the questionnaire was not completed in full.For this reason, a difference in the numbers analyzed (between those for the prevalence of post-traumatic stress, anxiety, and depression and those for the analysis of associated factors) is observed (see Fig. 1).The search for factors associated with the occurrence of psychological disorders was carried out on 525 people (those who completed all the survey forms) ƚ Number of observations / Total number of observations ǂ The results presented correspond to the pre-selection of variables at p-value < 20%.The second selection of variables was made at the 5% threshold and then integrated into the multivariate model § Adjusted odd ratio with a 95% confidence interval

Table 1
Associated factors with the presence of PTSD

Table 2
Associated factors with the anxiety and the depression

Table 3
Associated factors with the emotional exhaustion, depersonalization, and personal accomplishment scores