The vulnerability of healthcare workers to COVID-19 is enhanced by the fast spreading of the coronavirus (the reproduction rate of SARS-CoV-2 is around 2.5 compared with 0.9 for MERS-CoV) (Petersen et al, 2020) and unexpected transmission by air travel, because many countries had limited or no time to prepare well. As of 3 June 2021, there were over 210 countries and territories suffering from COVID-19, with around 171 million confirmed cases and 3.5 million deaths (European Centre for Disease Prevention and Control, 2021). Among the COVID-19 cases reported to the World Health Organization (WHO), 14% are among health workers in which the loss of lives accounts for thousands (WHO, 2020). Similarly, the observational cohort study conducted by Nguyen et al (2020) presented that frontline healthcare workers in the UK and USA had at least a threefold increased risk of COVID-19 infection compared with the general community.
Besides the physical threats from COVID-19, all healthcare workers have to deal with psychological challenges that come from the excessive workload, fear of infection, negative feelings from isolation, and lack of clinical experience (Spoorthy et al, 2020; Zhang et al, 2020a). A previous systematic review showed that post-traumatic stress disorder (PTSD), anxiety, depression and burnout are the most frequently reported clinically significant mental health symptoms among healthcare workers attributed to infectious outbreaks (Cabello et al, 2020). A recent meta-analysis of the impact on mental health in healthcare workers due to COVID-19 reported that the pooled prevalence of depression and anxiety among this workforce was 25% (95% CI: 17-33%) and 26% (95% CI: 18-34%) respectively (Luo et al, 2020). Insomnia among healthcare workers during COVID-19 was estimated by another meta-analysis at 34% (95% CI: 27%-42%) (Spoorthy et al, 2020).
At the cornerstone of healthcare systems and caring for the sickest patients, intensive care units (ICUs) crystallise strains and tensions. The rate at which ICU nurses tested positive for PTSD symptoms was 10% higher than among general nurses (P=0.03) (Mealer et al, 2007). In the context of the COVID-19 pandemic, estimates of 28-day in-hospital mortality in ICU of 26–45% have been reported, exposing staff to repeated suffering, which in turn could be felt as a failure (Intensive Care National Audit and Research Centre, 2021). Surges of patients exceeding normal capacity have pushed ICUs to expand into areas such as operating theatres and conventional wards. Many hospitals have had to increase the ICU workforce via staff redeployment, replacing familiar clinical environments with a completely new cohort of patients, with staff required to undertake duties that require different skills from their norm (Bhargava et al, 2020). Three studies reported that working in ICU during the COVID-19 pandemic contributed to a higher risk of having anxiety disorders compared with working in other departments, and anxiety is likely to be the cause and comorbidity of insomnia and depression (Ohayon et al, 1998; Liang et al, 2020; Zhu et al, 2020).
This study aimed to address the lack of current literature concerning ICU workers' mental health and evaluate the global mental health burden of ICU staff during the COVID-19 pandemic, by disseminating a survey including validated screening methods to establish the prevalence of indices of depression, insomnia and PTSD. The authors' hypothesis is that critical care staff will exhibit adverse mental health across all examined areas, as widely reported among frontline healthcare workers during the COVID-19 pandemic.
Methods
Study design
This study is a cross-sectional cohort survey study using three validated scales. It was conducted during the COVID-19 pandemic from 18 May 2020 to 29 May 2020 and was approved by the Imperial College Research Ethics Committee (reference: 20IC5991). Study participants included all health professionals (physicians, nurses, and allied health personnel) working in ICU during the pandemic, regardless of whether this was their usual place of work (referred to as ‘non-redeployed’ staff) or whether they had been redeployed to ICU (‘redeployed’ staff). Participants who filled the questionnaire but were not working in ICU were excluded.
The survey was conducted in the UK, France, Italy, Taiwan, and Mainland China. Participants were recruited via verbal dissemination, social media platforms (eg Facebook, Twitter, LinkedIn, WeChat), and email communication. The online survey, written in English, was administered via Google Docs, with additional translated versions in French, Italian, simplified and traditional Chinese (administered via www.typeform.com) by respective fluent medical professionals. All data were anonymised. Participant consent was implied by them participating in the survey, as stated in the promotional material and on the survey front page. The research is reported according to the STROBE statement for cohort studies (see tinyurl.com/strobe-checklist).
Materials
The study examined themes from existing COVID-19 literature in China relating to the psychological impact on healthcare workers, such as fear, worries, compassion fatigue and burnout (Chen et al, 2020; Wu et al, 2020). The authors adopted three brief screening instruments, validated for the evaluation of the psychological impact on healthcare workers during the COVID-19 pandemic or previous crises (De Stefano et al, 2020; Qi et al, 2020; Zhang et al, 2020b):
- The two-item Patient Health Questionnaire-2 (PHQ-2) (Kroenke et al, 2003), which comprises the first two items of the nine-item depression scale PHQ-9, aims to grade depression severity and assess depression change over time by a four-point response format for each of its two items (Löwe et al, 2005). Each of the two items is scored from 0 to 3, giving a total score range of 0 to 6. The original validation study of the PHQ-2 recommended a cut-off point of ≥3 corresponding to a sensitivity of 0.83 and specificity of 0.90. As an initial first step assessment in the sign of depressive symptoms for this study's purpose, PHQ-2 is adequate with good acceptability in translated versions, including Chinese, German, Japanese (Manea et al, 2016).
- The 8-item Athens Insomnia Scale-8 (AIS-8) (Soldatos et al, 2000) is a scale for use in a large variety of clinical and research settings where the quantification of sleep problems is required and was suggested by previous reviews as an appropriate instrument for screening insomnia in terms of the comparison of scales characteristics for diagnostic properties, sleep domains and feasibility (Chiu et al, 2016). Each item is rated from 0 to 3, where 0 corresponds to ‘no problem at all’ and 3 corresponds to ‘very serious problem’; the total score ranges from 0 (denoting absence of any sleep-related problem) to 24 (representing the most severe degree of insomnia).
- The 10-item Trauma Screening Questionnaire (TSQ) (Brewin et al, 2002) consists of 10 arousal and re-experiencing items (Foa et al, 1993) simplified to provide binary response options only, and performs particularly well for predicting PTSD within 1 year of a traumatic event (Brewin, 2005). Respondents reveal whether or not they have experienced each symptom at least twice in the past week. An optimal cut-off point was found to be a ‘yes’ response to at least 6 items in any combination. (Six or more positive responses indicate that a respondent is at risk.)
The authors recorded demographic data, including decade of age, gender, ethnicity, job title, pre-existing mental health conditions, redeployment status, departments of exercise, and average daily hours spent wearing personal protective equipment (PPE). They retrieved the date of peak infection from the 2019 Novel Coronavirus Visual Dashboard operated by the Johns Hopkins University Center for Systems Science and Engineering (Dong et al, 2020).
Statistical analyses
The statistical analyses were performed using Matlab (version R2020a, The Mathworks Inc, Natick, Massachusetts, USA). Continuous variables were expressed as mean and standard deviation (SD) or median and interquartile range (IQR), as appropriate, and examined by the non-parametric Kruskall-Wallis (KW) one-way analysis of variance test (Theodorsson-Norheim, 1986). Multivariable linear regression was used to evaluate the association between the participants' characteristics and the scores on the scales used in the survey. A P-value of <0.05 was considered as statistical significance.
Results
Cohort description
The main elements of the cohort demographics are summarised in Table 1. The authors initially collected 670 responses, of which they excluded 120 from Mainland China and 35 from Taiwan because they did not correspond to staff deployed to ICU. The final cohort comprised 515 respondents from 7 countries: 309 from France (60%), 73 from the UK (14.2%), 68 from Mainland China (13.2%), 31 from Italy (6%), 29 from Taiwan (5.6%), 3 from Egypt (0.6%) and 2 from Belgium (0.4%). Most respondents were female (376/515, 73%), predominantly of white ethnicity (376/515, 73%), with Asian as the second largest group (113/515, 21.9%). The most common age bands were 31–40 years old (223/515, 43.3%) and 41–50 (125/515, 24.3%), with the other groups accounting for about 30% of the responders. Of those who responded, 11/190 (5.8%) had pre-existing mental health conditions.
Table 1. Cohort demographics
| Item | n(%) |
|---|---|
| Gender | |
| ▪ Male | 139 (27) |
| ▪ Female | 376 (73) |
| Age range | |
| ▪ 20–30 | 110 (21.3) |
| ▪ 31–40 | 223 (43.3) |
| ▪ 41–50 | 125 (24.3) |
| ▪ 51–60 | 50 (9.7) |
| ▪ 61–70 | 7 (1.3) |
| Country | |
| ▪ France | 309 (60.0) |
| ▪ UK | 73 (14.2) |
| ▪ Mainland China | 68 (13.2) |
| ▪ Italy | 31 (6.0) |
| ▪ Taiwan | 29 (5.6) |
| ▪ Egypt | 3 (0.6) |
| ▪ Belgium | 2 (0.4) |
| Ethnicity | |
| ▪ White/Caucasian | 376 (73) |
| ▪ Asian/Asian British | 113 (22) |
| ▪ Black/African/Caribbean/Black British | 12 (2.3) |
| ▪ Mixed/Multiple ethnic groups | 5 (1.0) |
| ▪ Other | 3 (0.6) |
| ▪ Did not wish to disclose ethnicity | 6 (1.2) |
| Occupation | |
| ▪ Senior physician | 71 (13.8) |
| ▪ Senior resident/specialist trainee/fellow | 46 (8.9) |
| ▪ Junior resident/specialist trainee | 33 (6.4) |
| ▪ Staff nurse/specialised nurse | 270 (52.4) |
| ▪ Nurse in charge/head nurse | 36 (7.0) |
| ▪ Healthcare support worker | 25 (4.8) |
| ▪ Physiotherapist/Occupational therapist | 15 (2.9) |
| ▪ Other | 19 (3.7) |
| Redeployed to intensive care? | |
| ▪ Yes | 310 (60.2) |
| ▪ No | 205 (39.8) |
| Specialty or department of usual practice | |
| ▪ Intensive care | 174 (33.8) |
| ▪ Anaesthetics | 170 (33.0) |
| ▪ Medical ward | 37 (7.2) |
| ▪ Emergency department | 31 (6.0) |
| ▪ General practice/primary care | 20 (3.9) |
| ▪ Surgery/operating theatres | 16 (3.1) |
| ▪ Psychiatry | 12 (2.3) |
| ▪ Radiology | 11 (2.1) |
| ▪ Oncology | 10 (1.9) |
| ▪ Other | 34 (6.6) |
| Pre-existing mental health condition?* | |
| ▪ Yes | 11 (5.8) |
| ▪ No | 179 (94.2) |
The main occupation was nurse (270, 52.4%), and the authors also collected 36 replies from senior nurses or nurses in charge (7%). Among physicians, 71 responses came from senior doctors/attendings/consultants (13.8%), 46 from senior residents or fellows (8.9%), and 33 from junior residents (6.4%). The rest of the cohort consisted of healthcare support workers (25, 4.8%), physiotherapists/occupational therapists (15, 2.9%) and others (19, 3.7%).
Although all respondents were working in intensive care at the time of answering the questionnaire, the majority had been redeployed to intensive care from their usual specialty (310/515, 60.2%). Their most common place of work was intensive care (174/515, 33.8%), followed by anaesthetic departments (170/515, 33%), medical wards (37/515, 7.2%), emergency units (31/509, 6%), general practice or primary care (20/509, 3.9%), and others. There were no missing data.
General wellbeing questions
The authors started by asking the participants a few general questions around their wellbeing and anxieties (Table 2). The feeling of support from the general public was strong, with 74% of respondents replying that they felt ‘very or extremely supported’. However, 57% of staff declared feeling very or extremely worried about getting infected, while 21% were very or extremely worried about dying after catching the virus. Overall, 78% of respondents reported feeling at least a little numb about the loss of human lives, and most (88.9%) thought about it outside of work. Finally, 455/515 (88.3%) agreed that healthcare staff should be offered dedicated time off work once it was clinically safe to do so, and 439/515 (85.2%) were in favour of offering individual counselling sessions to healthcare workers.
Table 2. General wellbeing questions
| n (%) | ||||
|---|---|---|---|---|
| No | A little | Very | Extremely | |
| During the outbreak, did you feel supported by the public? | 17 (3.3) | 117 (22.7) | 234 (45.4) | 147 (28.5) |
| To what extent did you worry about getting infected? | 26 (5) | 196 (38) | 176 (34.1) | 117 (22.7) |
| To what extent did you worry about dying? | 178 (34.6) | 229 (44.5) | 80 (15.5) | 28 (5.4) |
| Regarding the extra loss of lives due to the outbreak, do you feel kind of numb? | 112 (21.7) | 221 (43) | 140 (27.2) | 42 (8.1) |
| Did you think about it outside of work? | 57 (11.1) | 184 (35.7) | 172 (33.4) | 102 (19.8) |
| Did you avoid dealing with your feelings about it? | 174 (33.8) | 185 (35.9) | 122 (23.7) | 34 (6.6) |
Scores main results
The median score for the PHQ-2 was 2 (IQR 1-3), for the AIS-8 it was 10 (IQR 6-14) and for the TSQ it was 3 (IQR 1-6). Overall, 192 (37.3%), 405 (78.6%) and 143 (27.7%) of the participants screened positive for the PHQ-2, AIS-8 and TSQ, respectively, using the recognised test thresholds (Figure 1a). The authors observed a negative correlation between the time elapsed since the peak infection and the rate of positive screening for all tests (Figure 1b), although the correlation was only significant for PHQ-2 and TSQ: R2=0.83 for PHQ-2 (P=0.03); R2=0.56 for AIS-8 (P=0.15), and R2=0.76 for TSQ (P=0.05).
Figure 1. Violin plot of the distribution of scores for the three screening tests (a). The median score for the PHQ-2 was 2 (IQR 1–3), for the AIS-8 it was 10 (IQR 6–14) and for the TSQ it was 3 (IQR 1–6). Relationship between the time elapsed since peak infection and mean score for three screening tests (b). Correlation for the 3 tests were: R2=0.83 for PHQ-2 (P=0.03); R2=0.56 for AIS-8 (P=0.15), and R2=0.76 for TSQ (P=0.05)
Scores subgroup analyses
Redeployment status and gender
The AIS-8 score was higher in redeployed staff (mean (SD): 10.4 (4.8)) versus non-redeployed (9.3 (5), KW P=0.008), while PHQ-2 and TSQ scores were not statistically different (PHQ-2: 2.1 (1.7) in redeployed versus 2.2 (1.7) in non-redeployed, KW P=0.58 and TSQ: 3.8 (2.8) versus 3.6 (2.8), P=0.44) (Figure 2a). Women scored higher than men in all 3 tests: PHQ-2: 2.3 (1.7) versus 1.8 (1.6); AIS-8: 10.5 (4.7) versus 8.4 (5); TSQ: 4 (2.7) versus 3 (2.9), KW P<0.005 for all tests (Figure 2b).
Figure 2. Test results by redeployment status (a) and by gender (b). The AIS-8 score was higher in redeployed staff (mean (SD): 10.4 (4.8)) versus non-redeployed (9.3 (5), KW P=0.008), while PHQ-2 and TSQ scores were not statistically different. Women scored higher in all 3 tests: PHQ-2: 2.3 (1.7) versus 1.8 (1.6); AIS-8: 10.5 (4.7) versus 8.4 (5); TSQ: 4 (2.7) versus 3 (2.9), KW P<0.005 for all tests
Country
Subgroup analyses by country identified variations in the rate of participants screening positive for a given test (Figure 3). The UK scored the highest for PHQ-2 (mean 2.6 (1.8)), France for AIS-8 (11.1 (4.5)) and Italy for TSQ (4.1 (2.8)). The proportion of participants screened positive across the different countries was 16–49% for PHQ-2, 60–86% for AIS-8 and 17–35% for TSQ. The differences across countries were positive for all tests (KW P<0.01).
Figure 3. Test results by country. The United Kingdom (UK), France (FR), Italy (IT), Mainland China (CN), Taiwan (TW). The UK scored the highest for PHQ-2 (mean (SD) 2.6 (1.8)), France for AIS-8 (11.1 (4.5)) and Italy for TSQ (4.1 (2.8)). The proportion of participants screened positive across the different countries were 16 to 49% for PHQ-2, 60 to 86% for AIS-8 and 17 to 35% for TSQ. The differences across countries were positive for all tests (KW P<0.01) (NB Egypt and Belgium not included due to insufficient data)
Time in PPE
The authors identified an increase in the scores with longer time spent in PPE (Figure 4). Between those who did not spend any time in PPE and those who spent more than 6 hours per day in PPE, there was a 40% increase in PHQ-2 score, 24% increase in AIS-8 and in TSQ. The differences across groups were significant for all three tests (KW P<0.01).
Figure 4. Test results by time spent in PPE. We identified an increase in the scores with longer time spent in PPE. Between those who did not spend any time in PPE and those who spent more than 6 hours per day in PPE, there was a 40% increase in PHQ-2 score, 24% increase in AIS-8 and in TSQ. The differences across groups were significant for all three tests (KW P<0.01)
Occupation
The authors analysed the scores for all three tests by classifying the participants into three main occupation categories: doctors, nurses and support workers (Figure 5). For all three tests, doctors scored lower than the other two categories (KW P<0.005).
Figure 5. Test results by occupation category: doctors, nurses and support workers. For all three tests, doctors scored lower than the other 2 categories (KW P<0.005)
Linear regression
Multiple linear regression models were computed to test the association between the demographics data (all converted to categorical features) and the scores on the three scales. A significant regression equation was found for all scores: PHQ-2: F(31,477)=2.2 (P<0.001), with an R2 of 0.13; AIS-8: F(31,477)=4.06 (P<0.001) with an R2 of 0.21; TSQ: F(31,477)=2.1 with an R2 of 0.12 (P<0.001). The most important features for each model with P<0.1 are presented in Table 3. After adjustment, the features that appear positively associated with the scores include the role of senior and junior resident, advancing age and the redeployed status. Conversely, the gender male, respondents from China, Taiwan, and from specialties such as radiology, psychiatry, medicine or oncology were negatively associated with the scores.
Table 3. Most important features in the linear regression models
| Test | Feature | Estimate | SD | P value |
|---|---|---|---|---|
| PHQ-2 | Country/China | -1.36 | 0.48 | 0.004 |
| Role/Senior Resident | 0.58 | 0.21 | 0.005 | |
| Age/41–50 | 0.65 | 0.24 | 0.008 | |
| Role/Junior Resident | 0.77 | 0.31 | 0.014 | |
| Country/Taiwan | -1.13 | 0.58 | 0.051 | |
| Specialty/Radiology | -1.26 | 0.67 | 0.061 | |
| Specialty/General Practice | 0.67 | 0.41 | 0.098 | |
| AIS-8 | Role/Senior Resident | 2.23 | 0.58 | <0.001 |
| Role/Junior Resident | 3.10 | 0.86 | <0.001 | |
| Specialty/Psychiatry | -3.46 | 1.40 | 0.014 | |
| Specialty/Medicine | -2.46 | 1.01 | 0.016 | |
| Specialty/Oncology | -3.49 | 1.57 | 0.026 | |
| Gender/Male | -1.07 | 0.50 | 0.033 | |
| Redeployed/Yes | 1.10 | 0.53 | 0.040 | |
| Specialty/Radiology | -3.41 | 1.88 | 0.069 | |
| Country/China | -2.43 | 1.34 | 0.070 | |
| TSQ | Role/Senior Resident | 1.22 | 0.35 | 0.001 |
| Role/Junior Resident | 1.68 | 0.52 | 0.001 | |
| Age/31–40 | 0.84 | 0.35 | 0.016 | |
| Age/41–50 | 0.97 | 0.41 | 0.018 | |
| Age/51–60 | 1.12 | 0.52 | 0.033 | |
| Specialty/Medicine | -1.15 | 0.61 | 0.061 | |
| Gender/Male | -0.57 | 0.30 | 0.061 | |
| Age/61–70 | 2.00 | 1.11 | 0.071 | |
| Country/France | -0.86 | 0.48 | 0.072 | |
| Specialty/Radiology | -2.02 | 1.13 | 0.074 | |
| Country/China | -1.39 | 0.81 | 0.085 |
Note: Positive estimates signify a positive association with the score, while a negative estimate corresponds to a negative association. Only the features with a P value of less than 0.1 are shown AIS-8=Athens Insomnia Scale-8, PHQ-2=Patient Health Questionnaire-2, TSQ=Trauma Screening Questionnaire
Discussion
The authors used three validated scales to screen 515 ICU workers for indications of depression, insomnia and PTSD. Across the various countries explored, 16% to 44% of respondents exceeded the threshold for signs of depression, 60% to 86% for insomnia and 17% to 35% for PTSD. There was a negative correlation between the time elapsed since the peak infection and the scores achieved by the various countries. Staff redeployed to ICU and female gender were associated adversely with self-reported outcomes. Doctors had overall lower scores than nurses and healthcare support workers, however, after adjustment in linear regression analysis, junior and senior residents were consistently highlighted at increased risk. The authors identified an increase in all scores with longer time spent in PPE, which reinforces causality.
There is currently a paucity of European data reporting mental health effects, and specifically describing the experience of ICU staff and redeployed workers, calling for focused studies in this regard (Holmes et al, 2020). As hypothesised, in the study reported in this article, the cohort of ICU staff caring for COVID-19 patients self-reported a significant mental burden across depression, insomnia and PTSD scales. Besides the clinical demands required to treat such patients, staff are exposed to mental burden following sustained intense stress from their day-to-day activities. Onder et al (2020) reported 13-20% mortality rates in confirmed COVID-19 patients in Italy aged 70 or more. In a cohort of 5700 COVID-19 patients in the USA, 20% required mechanical ventilation and a quarter died (Richardson et al, 2020). Such frequent deaths may result in what is known as death anxiety, a negative feeling or anxiety and dread at the thought of death or its prospect (Peters et al, 2013; Karanikola et al, 2015).
Even at baseline, the public health sector suffers from high levels of stress, leading to absenteeism. In the UK, the public health sector has the highest rates of sickness absence of all industries in the UK, and the 2020 NHS Staff Survey reported 44.0% of staff feeling unwell due to work-related stress (Survey Coordination Centre NHS England, 2021). This will only have been worsened by the COVID-19 crisis.
Similar to the severe acute respiratory syndrome (SARS) epidemic of 2003 and Ebola in 2014-2016, the sudden emergence of a large-scale contagious and deadly disease resulted in damaging pressure on healthcare workers (Liu et al, 2012; Ji et al, 2017). In China, some authors have reported anxiety in up to 25% of frontline healthcare workers and PTSD in up to 45% (Huang et al, 2020). Up to 65% of Chinese frontline healthcare workers have reported subsequent insomnia (Wu and Wei, 2020). A recent systematic review of 13 studies during COVID-19 included 33 000 healthcare staff, mostly from China, reported an overall prevalence of 23% of anxiety, 23% of depression and 39% of insomnia. The data in the study undertaken by the authors is congruous with this review, which demonstrated that gender and occupation were associated with the outcome, with females, nurses and allied health professionals being at higher risk of anxiety and depression than males and doctors (Pappa et al, 2020). This study's ICU cohort self-reported an even higher mental burden: up to 44% depressive symptoms, 35% of PTSD and 86% of sleeping disorders. Whereas Pappa et al's review included general hospital staff, this duty cohort consisted exclusively of ICU staff who have been exposed to higher levels of patient severity and higher rates of death and may therefore report greater rates of affective disorders (Schäfer et al, 2018). Adverse mental burden may be exacerbated via pressures on personal resilience during social lockdown (Willan et al, 2020). Healthcare staff may endure individual isolation and loss of social support, combined with a change in work shift pattern (Lung et al, 2009; Bohlken et al, 2020; Xiao et al, 2020).
The authors identified a relationship between the time elapsed since the local peak of COVID-19 infection and mental health burden across the outcome measures. It is likely that some of the stigma experienced by staff improves with time, as new COVID-19 admissions reduce, the number of deaths drop, patients are discharged from ICU and national lockdowns are eased (Nobles et al, 2020). Antonovsky's theory of ‘salutogenesis’ provides a useful framework for analysing the relationship between health, stress, and coping (Antonovsky, 1979). The key concept is that individuals might regain health and wellbeing after a stressful event if their ‘sense of coherence’ is preserved, that is their belief that the event they encountered is comprehensible and manageable, and that reaction to the event was meaningful. The sense of meaning is generally strong in the healthcare domain (Dixon-Woods et al, 2014; Schäfer et al, 2018). Additionally, public support may have empowered staff and helped regain morale. Wellbeing measures for ICU staff such as psychological counselling or mindfulness may help in the coping process, and indeed, 85% of ICU staff in this study were in favour of receiving individual counselling sessions.
Limitations
The study has a number of limitations. The authors could not determine the response rate due to the method of study advertisement using non-specific healthcare emails and social media platforms. The presence of a reporting bias inflating the scores cannot be excluded, since subjects suffering from mental burden might have been more likely to respond. Robust conclusions cannot be drawn from countries where the sample size was particularly limited (eg Egypt, Belgium). Conclusions drawn from examined outcome scores are not exclusively representative, but rather reflect an association of confounding factors (culture, professional experience, pre-existing social support) (Sull et al, 2015). Despite this, the data provides a useful indicator of mental wellbeing across ICU staff in order to recognise such impact, and address them via targeted local and national policies (Qiu et al, 2020).
Conclusion
The results of this study suggests that the COVID-19 crisis has negatively impacted the mental health of healthcare workers globally. All ICU staff should have access to early and effective mental health resources as part of a wider staff health and wellbeing strategy. This is especially pressing as hospitals and governments prepare their workforce for potential further surges of COVID-19 patients.
KEY POINTS
- The study included 515 respondents across seven countries (France, the UK, Italy, Mainland China, Taiwan, Egypt and Belgium, who responded to validated mental health surveys during the first COVID-19 outbreak
- A significant proportion of respondents screened positive for depression (16–49%), insomnia (60–86%) and post-traumatic stress disorder (17–35%)
- Adverse outcomes were associated with longer time spent in PPE, female gender, advancing age and redeployed status
- Putting in place mental wellbeing support for healthcare workers, to help sustain subsequent outbreaks and help to alleviate sickness/stress related absenteeism should be a consideration for policymakers
CPD reflective questions
- Consider whether the redeployment of staff to unfamiliar surroundings improved the culture of teamwork. Did it also change traditional hierarchies, and was the latter considered a barrier by some staff members?
- Do you feel the provision of mental wellbeing support for healthcare workers has improved since the COVID-19 outbreak? How can we be more open about mental health at work?
- Reflect on how your experience as a healthcare worker during COVID-19 has affected your family and those close to you. Have you experienced stigma as a key worker during the outbreak peaks?