Introduction

As the COVID-19 pandemic progresses, several hospitals have gone through changes in their usual work dynamics, among which are the elevated demand for intensive care unit (ICU) beds, the increased death rate, and preventive measures to avoid infection during healthcare. All of those elements are present in the daily routine of the ICU staff, interfering with their mental and physical health. Those factors may lead to increased workload and stress, therefore enhancing the risk of developing acute or chronic psychological disturbances (e.g., burnout, substance abuse [SA], acute stress disorder [ASD], and post-traumatic stress disorder [PTSD]) [1–4].

Even in non-pandemic periods, healthcare professionals working in ICUs suffer high stress loads possibly due to the environment and the severity of patients [5]. In this context, recurrent stressful stimuli may lead the professionals developing SA [6, 7] and ASD [8], more often than in other areas of healthcare [9–11]. Additionally, recent studies suggest those workers have been experiencing a high impact in their mental health during the pandemic [12, 13]. However, limited research has been conducted on the occurrence of those phenomena in healthcare professionals working in Brazilian ICUs during the COVID-19 pandemic. Furthermore, there is no data providing an independent analysis of the eventual associations between each single stressor element that may be present in the ICU environment and the occurrence of SA, PTSD, and ASD in that population [14].

Therefore, the aim of this study was to evaluate the occurrence of SA, PTSD, and ASD in professionals working in Brazilian ICUs, either having direct contact with patients or not, using validated questionnaires as tools to assess the prevalence of those conditions. The secondary objective was to identify associations between epidemiological and labor characteristics of those workers and the development of those psychological disturbances in that population.

Methods

Results

Three hundred seventy-six ICU professionals agreed to participate in this survey, 340 with direct exposure to COVID-19 infected patients (see Table 1). Professionals directly exposed to infected patients included 54 physicians (14%), 53 physiotherapists (14%), 226 nurses/technicians (60%), 38 pharmacists (10%), 5 nutritionists (1%), 7 psychologists (1.5%) and 3 social service professionals (0.7%). 71.5% of the participants were female. The age variable was categorized, and participants who were less than 38 years of age were the most frequent (56.4%). More than 80% of the workers who were directly exposed participated in this study. It was not possible to determine the answer ratio among workers without direct exposure as most of them (e.g., security, cleaning services) were employed by third-party companies. Less than ten individuals refused to participate.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Socio demographic, personal, and occupational characteristics of participants.

https://doi.org/10.1371/journal.pone.0263892.t001

Most participants (66.5%) had a weekly exposure greater than 24 hours. 50% of the participants had been infected by COVID-19, and 38% of them had had a relative infected. Although most of them shared their homes with individuals from vulnerable populations (e.g., elderly relatives), only 21% of them sought alternative accommodation away from home during the pandemic. Forty-three participants (11.5%) declared to have a history of psychiatric disease. Table 1 describes the characteristics of the studied individuals.

IES-R score results

Most participants (76.3%) had scores ≥12 (see Table 2). One-hundred twenty-seven (33.8%) of them had scores higher than 33, which strongly suggests a diagnosis of PTSD and the need of psychiatric evaluation as soon as possible. Eighty-nine (23.7%) participants were free from PTSD signs or only had a few of them. Results are summarized on Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Frequencies of IES-R score categories.

https://doi.org/10.1371/journal.pone.0263892.t002

Psychiatric medical history.

The relationship between IES-R scores and history of psychiatric disease was assessed, but those factors did not present a statistically significant association (p = 0.341, OR = 2.68 [95% CI = 0.35–20.51]) (see Table 3). Then, the relationship between IES-R scores and history of anxiety disorders (which included ASD and PTSD) was assessed, but those factors also did not present a statistically significant association (p = 0.44, OR = 2.23 [95% CI = 0.29–17.13]). Table 3 summarizes those findings.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Factors without an association with IES-R scores.

https://doi.org/10.1371/journal.pone.0263892.t003

Personal and occupational factors.

Among personal factors, being a female and having a relative infected by COVID-19 were associated with higher IES-R scores (Table 4). When compared to men, women had a higher median score (27 vs 19 in males [p = 0.0001]) even after adjustment with Wilcoxon rank sum test. Professionals with infected relatives (OR = 3.91 [p = 0.031, 95% CI = 1.13–13.50]) were also significantly more likely to score higher in the IES-R, therefore more likely to suffer from PTSD or ASD (Table 4). The relationship between those who had sought alternative accommodation away from home during the pandemic in order to protect their relatives and IES-R scores was also assessed, but it was not statistically significant (OR = 5.72 [p = 0.091, 95% CI = 0.75–43.32])—see Table 3 above.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Results of multivariate logistic regression of factors associated with increased IES-R scores.

https://doi.org/10.1371/journal.pone.0263892.t004

Among occupational factors, direct exposure to infected patients was associated with higher IES-R scores. When compared to professionals who were not directly exposed, those who had a direct exposure were more likely to achieve higher scores, with an OR = 5.62 (p = 0.001, 95% CI = 2.10–15.03). Those findings are presented in Table 4.

Associations with substance abuse

Results indicate that, according to the ASSIST 2.0 form, 24.7% of ICU workers had scores suggestive of alcohol abuse. Scores compatible with tobacco abuse and dependence were seen in 11.2% and 2.3%, respectively. Eighteen workers (4.8%) had scores compatible with hypnotics abuse.

There was no statistically significant association between personal or occupational factors and substance use. However, two association trends were found. For tobacco and alcohol, higher scores had an association trend towards history of psychiatric disease, direct exposure to infected patients and those who sought alternative accommodation away from home during the pandemic in order to protect their relatives.

For cannabis, stimulants, and cocaine, higher scores had an association trend towards history of psychiatric disease and direct exposure to infected patients. The prevalence of each substance use is detailed on Table 5 below.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Prevalence of substance use according to the ASSIST 2.0 categories.

https://doi.org/10.1371/journal.pone.0263892.t005

Discussion

This research found associations between participants’ personal and occupational factors and signs of ASD/PTSD in a COVID-19 ICU in this Brazilian hospital. Sex (female), direct exposure to infected patients and having a relative infected by COVID-19 were significantly associated with higher IES-R scores. Other factors such as history of psychiatric disease, history of anxiety disorders, and seeking alternative accommodation away from home during the pandemic were assessed but did not present a statistically significant association with higher IES-R scores. The prevalence of alcohol and tobacco abuse was high. The evaluation of substance use among ICU workers with the ASSIST 2.0 form was not able to determine any statistically significant associations with personal or occupational factors and higher scores. Nonetheless, there was an association trend towards tobacco, alcohol, cannabis, stimulants, and cocaine use in those with a history of psychiatric disease and with direct exposure to infected patients. In a recent study evaluating the mental health of ICU workers in England [19], authors found similar results, with high prevalence of traumatic stress disorders and alcohol abuse, especially among nurses (which account for the most numerous subpopulation of the presenting study).

We believe that those factors associated with higher IES-R scores were directly related to an increased stress load in those individuals. Whilst the cause is unclear, it could be suggested that, when compared to most of the occupational factors, personal life characteristics, as well as direct exposure to infected patients, were more strongly related to the likelihood of having PTSD or ASD symptoms. Researchers investigating the impact of personal and occupational factors on the mental health of Canadian ICU workers exposed to COVID-19 infected patients found that, following a multivariate analysis, those elements related to the anxiety about being infected during work were the best predictors of scores indicative of traumatic stress disorders [20]. We further hypothesize that worrying about the safety of their relatives is another possible cause of this overwhelming stress.

In terms of substance use, the most frequently consumed were tobacco and alcohol. Those were also the substances with higher frequencies of abuse and dependence. It is possible that those results reflected the availability of those substances. Also, the abusive consumption of those substances could be a coping mechanism during the pandemic. Even considering the socio-cultural differences among countries, the current results are in accordance with those presented by other authors, in other populations of ICU workers (e.g., Netherlands [4], Canada [20], United States [21], and England [19]). Thus, it is likely that the findings of the presenting study could be generalizable to ICU workers from different populations and should be taken into consideration by professionals working globally.

This study has several limitations. Firstly, as reported and studied in the field of substance abuse, this study may be subject to social desirability bias, which is the tendency of a participant providing answers based on what is considered a desired behavior within a specific socio-cultural environment [22, 23]. The fear of being treated with prejudice by their peers may also be a contributing factor for this bias in ICU workers [24]. Secondly, some ICU staff during the application of this study either refused to participate or were on medical leave. Given that some of those workers may have had psychological or psychiatric reasons for work leave, our findings can be underestimated. Thirdly, the lack of statistical significance could be due to a small sample size in some sub-groups and thus statistical testing lacked power. However, with regards to the non-statistically significant results, we believe they should be taken into account because it is reasonable to hypothesize that some of those factors may also contribute to the psychological burden in the ICU workers. Fourthly, this study had no follow-up sessions. Therefore, it was not possible to determine causal relationships or risk factors. Finally, it was a single-centered research and, therefore, influenced by local environmental biases. Nevertheless, our sample was large enough to find some statistically significant associations between participants characteristics and the development of PTSD/ASD symptoms during the COVID-19 pandemic. It was also possible to highlight the alarming prevalence of substance use among those individuals, which may carry devastating consequences for their personal and professional lives.

Considering the duration and the consequences of the COVID-19 pandemic, specifically for the population of this study, our results should bring to light the need for institutional support for those individuals. From the experiences with the severe acute respiratory syndrome virus (SARS) outbreak in the early 2000’s, it has been demonstrated that ICU professionals may need long term support for mental health issues developed during their work time with infected patients [25]. As recently proposed by other authors [26, 27], intervention plans should be traced not only during the pandemic but also after it, due to the prolonged course of diseases such as PTSD, ASD, and SA. For that reason, we believe that this study could be used for guiding larger, multicenter studies and institutional interventions to prevent and reduce the suffering of professionals dealing with similar situations in Brazil and in other countries.

Supporting information