The Acute Stress Scale in Healthcare Professionals Caring for Patients with COVID-19. Validation Study

Backgroud: This study designed and validated a scale to measure acute stress experienced during the care of patients with COVID-19 (EASE scale). Methods: The COSMIN recommendations were followed in the design and validation of the EASE scale. Internal consistency was calculated using Cronbach's Alpha and McDonald's Omega. Qualitative research was conducted to assess content validity. Exploratory and Conrmatory Factorial Analysis were used to assess construct validity. Mann-Whitney-Wilcoxon test and Lineal Regression were applied to assess criterion validity. Responsiveness and interpretability were also assessed. Results: A total of 228 health professionals from the Spanish public health system responded to the ten items, of the EASE scale. Internal consistency was 0.87 (McDonald’s Omega). Two factors explaining 55% of the variance were isolated. Goodness-of-t indices conrmed this factor structure. The highest level of stress was found among professionals working in health services who accumulated a higher number of deaths from COVID-19 (P<.05). By means of translation-counter-translation, it was adapted to Latin American Spanish, Brazilian Portuguese, and English. Conclusions: EASE gathers adequate metric properties. Its systematic use would allow the identication of professionals with a higher risk of developing affective/anxiety disorders and Moral injury that could limited that patients received safety and integrated care.


Background
The rate of reproduction, lethality and social alarm that accompanies SARS-CoV-2 (1, 2) poses a challenge to health systems in all countries. The Spanish health system has been overwhelmed by the number of patients with COVID-19, as of May 26th, 236.259 cases and 27.117 deaths. In addition, 20.4% of COVID-19 patients were health professionals (3).
The physical and mental effort involved in caring for patients with COVID-19 has caused acute stress, compassionate fatigue, and other affective pathologies which, together with psychosomatic reactions, have affected work morale (4)(5)(6). Integrated care has been jeopardized. Healthcare professionals have been reassigned to areas where they have no expertise or preparation, protocols made overnight and constantly changing and continuity of care interrupted at all levels.
Under these conditions, healthcare has been compromised (7). Secondary and tertiary prevention (effects in the 6 months at the end of the current care pressure) is di cult because most of these professionals are reluctant to seek help (8).
The aim of this study was to design and validate a self-applied Acute Stress Scale for those people who were in direct care of patients with COVID-19 (EASE). The utility of this instrument focuses on facilitating awareness of the level of stress endured in order to assess the impact of organizational changes in the healthcare systems to cope with the professionals' Moral Injury (9) as consequence of limited resources available to treat patients during the rst times of this pandemic and the interrupted care for Non-COVID19 patients, and contribute to secondary prevention of emotional and anxiety responses once the critical phase of the pandemic is over.

Methods
The study was conducted between March 20 and April 19, 2020 in Spain. Healthcare providers could reply using a website page and a mobile app (Fig. 1).
This study design and validation procedure followed COSMIN (10) recommendations considering reliability, validity, responsiveness, and interpretability. This study was approved by the Research Commission of the Sant Joan D'Alacant University Hospital.
Participants. A convenience sample of physicians, nurses, and other health personnel. The sample size was adjusted according to the number of items (maximum of 10), applying the criterion of a minimum of 20 subjects with valid responses per item. To preserve complete anonymity, no information on age or experience in the workplace was requested.
Reliability. Internal consistency was calculated using Cronbach's Alpha and McDonald's Omega. A value greater than .70 was considered acceptable for this analysis.
Content validity. A pragmatic literature review of items assessing acute stress in healthcare professionals was conducted for possible inclusion. Eight physicians, four nurses, ve psychiatrists, six psychologists and 10 safety responsible from 10 hospitals and four health centres participated collecting and grouping by similarity the categories of problem situations (content validity). These professionals presented the resulting scale to colleagues to assess their representativeness and understanding (Face Validity). They also assessed the understanding of the response options (four-point-scale).
Construct validity. Exploratory factor analysis (EFA) was used to determine the factorial structure of this scale (Construct Validity). The Principal Components technique followed by Varimax rotation was applied. Eigen values greater than .40 and factor loading greater than .5 were considered an acceptable level of missing data. The suitability for this EFA of the inter-item correlation matrix was calculated using the Bartlett's Sphericity test and the Kaiser-Meyer-Olkin test. Additionally, a con rmatory factor analysis (CFA) was used to con rm the underlying structure, estimating several t indices to test which CFA model best represented the dataset.
The translation-counter-translation method was used to ensure equivalence between the Latin American Spanish, Brazilian Portuguese and English versions (cross-cultural validity).
Criterion validity. Using the Mann-Whitney-Wilcoxon test and linear regression analysis, Enter Method, the capacity to classify the responses of the professionals in the EASE was assessed, according to the number of deaths registered in the geographical area where the health centre was located during the period from March 10 to April 19.
Responsiveness. We compared the EASE scores in two different periods of the evolution of the pandemic (March 25 to April 1, and between April 14 and 19, 2020) marked by a different daily number of cases and deaths by COVID-19 (between 800-900 versus < 500/deaths per day). The data published by the Ministry of Health on 22 April 2020 were taken as a reference.

Results
A total of 228 health professionals responded. Forty-two per cent were physicians, 28 per cent were nurses and 30 per cent were health support staff (nursing assistants and care attendants). They mostly worked in Madrid (28.9%), Andalusia (21.9%), Valencia (17.5%) and Catalonia (6.6%).
Content and Face Validity. The most relevant sources of acute stress, pointed out by the professionals experiences were: constant changes in instructions; shortage of material to avoid contagion; reduction in the number of staff due to risk exposure or contagion; bitter feelings when seeing patients die lonely; fear of infecting their families; making decisions reserved for situations of major catastrophes with a high component of ethical con ict; and the passing away of colleagues. These issues were represented into 17 possible reactive items. This number was nally reduced to 10 items (version 0 of the scale) once participants considered their representativeness and comprehension (Supplementary material).
Reliability. Internal consistency values were adequate, Cronbach's Alpha score was .85 and McDonald's Omega was .87.
Construct Validity. Two factors explaining 55% of the variance were isolated by EFA. The rst factor measured affective responses and the second fear and anxiety responses. The CFA, in the second stage of the study, indicated an acceptable t to the data ( Table 1, Fig. 2).

Discussion
This instrument has adequate metric properties and seems useful for professionals to become aware of their level of stress supported caring COVID-19 patients. The average level of self-reported stress was between 9 and 11 points up to 30. This level of distress was measured during the phase of an unprecedent care pressure. In this period there was relevant organizational changes, uncertainty about the evolution of patients, a lack of personal protective equipment and the number of professionals infected increased. It is expected that by decreasing this pressure the emotional response of the professionals will increase as they become fully aware of their experience (9,11).
In the forthcoming months we can expect professionals be affected by the so-called Moral Injury as consequence of having seen their professional codes violated (9) due to insu cient resources to care for COVID-19 patients, contradictory instructions or the interruption in continuity of care of other non COVID19 patients. Also, we can expect an increase of affective and anxiety reactions and symptoms, including in some cases post-traumatic stress among the professionals who saw their health and that of their loved ones threatened (12,13). This scale can help to identify and prevent such progression in health professionals, the second victims of SARS-CoV-2. Its mayor utility is expected in the next 3 to 6 months after the worst of the current health crisis has been overcome.
These same professionals should continue to care for patients with COVID-19, patients with sequelae from COVID-19 and all patients whose care processes have been interrupted during the acute phase of the pandemic. In this scenario, if rapid action is not taken, the capacity of professionals and therefore the quality and safety of patients may be compromised.
This scenario may be a valuable opportunity to consolidate integrated care, now it's the time to consider appropriate strategies to introduce measures to increase well-being at work, strengthen clinical leadership, and commit to a model such as the Quadruple Aim that invites consideration of patient outcomes as dependent on how caregivers are cared for.
The Quadruple Aim recognizes this focus within the context of the broader transformation required in our healthcare system towards high value care. While the rst three aims provide a rationale for the existence of a health system (14), the fourth aim becomes a foundational element for the other goals to be realized (15). The key is the fourth aim: creating the conditions for the healthcare workforce to nd joy and meaning in their work and in doing so, improving the experience of providing care (16). For this reason, it seems that "caring for the caregiver" (17) is necessary in the transformation of the health system and having instruments to be able to monitor the effects of measures implemented can be very useful: "caring for the caregiver in times of pandemic".
A limitation of this study is that it does not discriminate between professional categories, nor does it consider separately critical services during this crisis such as Critical Care and Resuscitation, Internal Medicine, Pneumology and Infectious Diseases.

Conclusion
The EASE scale meets adequate metric properties to be considered a reliable and valid scale. Its usefulness can be twofold. To help professionals to become aware of the emotional overload supported and, secondly, to measure the degree of affectation in order to avoid the progression towards more severe psychopathological conditions. Availability of data and materials: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Figure 1
Page 16/16 Flow diagram of the EASE Scale validation process.

Figure 2
Factorial structure based on CFA.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. S1EASEScale.docx