Burnout Syndrome During COVID-19 Second Wave on ICU Caregivers

This study has obtained the agreement of the ethics committee of Nimes (ref 20.0026). No written informed consent was required. The authors guarantee the anonymization of all data collected.

The COVID Impact study was conducted during the second wave of COVID-19 in France. First, we developed a survey, composed of two parts: caregiver characteristics and scales. We used the Hospital Anxiety and Depression Scale (HADS) [11, 12, 13]. The surveys were distributed between October 20th, 2020 and November 20th, 2020. Six intensive care units participated: one which had no COVID patients and five admitting only COVID-19 patients. All staff working in the ICU had the survey delivred. A total of 208 professionals completed the survey.

The HADS is used to evaluate anxiety and depressive symptomatology and severity of symptoms. The scale is composed of 14 items, 7 for anxiety and 7 for depression. Every question has 4 responses, all from 0 to 3. The minimum possible score is 0, the higher score is 42. The scores for the anxiety questions and the depression questions are added together to obtain 2 scores that are then added together. [12]

The Impact of Event Revisited Scale (IES-R) was used to assess Post Traumatic Stress Disorder. The scale is composed of 22 items, each statement having a score between 0 and 4. then the scores of all the items must be added together. The minimum possible score is 0 and the higher is 88.

Maslach Burnout Inventory (MBI) is a 22-item scale. The respondent answers questions related to his feelings and experiences at work over the last 7 days. This scale evaluates 3 components of burnout: the emotional exhaustion (EE), the depersonalization (DP) and the personal accomplishment (AP).

For the EE dimension, a score between 0 and 16 corresponds to a low level of EE. A score between 17 and 26 corresponds to a moderate level of EE and a score greater than or equal to 27 corresponds to a high level of EE. For the DP dimension, a score between 0 and 6 corresponds to a low level of DP. A score between 7 and 12 corresponds to a moderate level of PD and a score greater than or equal to 13 corresponds to a high DP. For the AP dimension, a score greater than or equal to 37 corresponds to a low level of AP. A score between 31 and 36 corresponds to a moderate level of AP and a score between 0 and 30 corresponds to high AP.

A high EE or DP score or a low AP score is sufficient to speak of burnout. Depending on the number of dimensions affected, there are stages of severity. The degree of burnout is as follow:

Descriptive statistics were performed using median [interquartile range, IQR], and frequency (percentage) for qualitative variables. Differences between groups were estimated using the Wilcoxon Rank Sum Test for quantitative variables, and by the Chi-squared test for qualitative variables or Fisher test when applicable. In order to select the predictor of severe burn-out, a bidirectional stepwise regression analysis based on the Akaike information criterion was performed. For the modelling strategy, we removed the answer from the physician, because of convergence problems (None of the 25 Medical staff presented a severe burnout). From the final model, we estimated the odds ratio associated with their 95% confidence interval and respective p value, and to assess the calibration of the model, the c statistic from the model was estimated. Missing variables were imputed for the regression using multivariate imputation by chained equations. Statistical significance for the p value was set at 0.05. Statistical analysis was performed by using R software V 3.6.3 and the package survival.

The cohort (Table 1) includes 208 professionals, 152 women (73.4%), the most represented age group is under 30 years old with 99 professionals (47.8%). One hundred and twenty-six people were married (61.5%). Among the cohort, 52.4% (109 people) were caregivers and 12% (25 people) were physicians. Forty-nine percent of respondents did not work in the ICU before the COVID-19 pandemia. During the first wave, 69.7% (145 people) were already in ICU, 30.3% (63 people) weren’t volunteer to work in ICU, and 58.7% had less than 12 months of critical care experience. During the first wave in April, 65.2% (162 people) of the cohort was in COVID+ ICU compared to 86.9% (179 people) in November. A total of 29.9% (56 professionals) did not feel trained enough. Approximately 10% (23 people) of the cohort had started psychological follow-up since the beginning of COVID-19. One hundred and two professionals (55.7%) felt that institutional psychological support was insufficient. Nineteen-point seven percent (41 people) of the team had severe BOS.

Table 1 presents univariate analysis results. Women were more affected by BOS 24% VS 9% (p=0.033), engaged professionals were more affected by BOS 26% VS 10% (p=0.009). Only 4% of doctors have been affected by BOS VS 44% of the reinforcement team and 23% of nurses (p=0.012). When professionals didn’t come to the ICU voluntarily, they were more affected by BOS 37% VS 5% (p<0.001). The fear of being contaminated increases the risk of developing burnout 27% VS 14% (p=0.033). When the professional thinks he or she might need psychological help, burnout prevalence increases (p<0.001). It’s the same if the professional describes his experience in ICU as a trauma for himself (p<0.001).

Figure 1 shows the distribution of severe BOS according to the professional job.

Fig. 1

Concerning anxiety: 48% of the professionals who suffer from an anxiety disorder also suffer from severe burnout. When the professionals were not anxious, only 6% suffered from severe BOS (p<0.001).

Concerning the anxio-depressive syndrome, 56% of the professionals who suffer from an anxio-depressive syndrome suffer from severe burnout. When the professionals did not suffer from this syndrome, only 7% suffered from a severe BOS (p<0.001).

Concerning PTSD, 53% of the professionals who suffer from PTSD suffer from severe burnout. When professionals do not suffer from PTSD, only 11% suffer from severe BOS (p<0.001).

Important to notice: there was no BOS for doctors, they’ve been excluded from multivariate analysis.

We observed a high correlation between EE and PTSD, anxiety, and Depressive symptomatology, and a moderate correlation between DP and AP and the PTSD, anxiety, Depressive symptomatology (Table 3) The independent risk factors for BOS are (Table 2).

Being engaged (OR = 3.61 (95% CI, 1.44;9.09)

Don’t working in ICU when it’s not COVID-19 pandemic (reinforcement) (OR = 37.71 (95% CI, 3.13;454.35)

Being a volunteer (OR = 0.10 (95% CI, 0.02;0.46). This variable corresponds to 3 statuses: I am ICU personnel (a), I am a back-up and I do not come to help voluntarily (b), I am a back-up and I come to help voluntarily (c). Therefore, taking as a reference the ICU staff (a), I have no excess risk of being a non-volunteer (b) compared to the ICU staff, but it is a real protective factor to be a volunteer (c).

And finally, it is important to point out that paramedics have a real excess risk that is not highlighted here due to lack of power.