Acute Stress Reaction (ASR) is a psychological reaction in the presence of a challenging or threatening event (Adler et al., 2020). It is defined by polymorphic and unstable anxiety symptoms, which usually appear immediately, within minutes of a catastrophic event and are accompanied by noticeable autonomic arousal, resulting in cognitive and somatic disruptions (World Health Organization, 1992).

These symptoms may include panic attacks, depersonalization or derealization, psychophysiological symptoms such as tension headaches, various gastrointestinal problems, tremors, sleep problems, cardiovascular variations such as increases in blood pressure, excessive worry about getting sick or dying (or what will happen to a loved one), a feeling of hyperarousal and hypervigilance, among others. This reaction usually resolves within hours or days (Svetlitzky et al., 2020).

On the other hand, Acute Stress Disorder (ASD) is a slightly different diagnostic category, included in the Diagnostic and Statistical Manual for Mental Disorders, 5th Edition (American Psychiatric Association, 2013). Unlike ASR, which can be diagnosed immediately at the scene based on observation of behavior, the ASD diagnosis requires clinical assessment at least three days following the event. Symptoms of ASD include a persistent clinical picture characterized by intrusion, avoidance, arousal, negative mood, and dissociation. Unlike ASR too, ASD is considered a mental health disorder and includes criteria of significant impairment in everyday functioning.

The American Psychiatric Association (2013) also proposes what is deemed as a related disorder: Post-Traumatic Stress Disorder (PTSD). According to the DSM-5 (APA, 2013), PTSD is a psychiatric condition associated with four clusters of symptoms (i.e., intrusion, avoidance, negative alterations in cognitions and mood, and hyper-arousal) that last at least 30 days. Taken together, there appears to be a time continuum such that ASR can occur in the immediate context of the event, ASD can occur in the first few days or weeks after the event, and PTSD can occur after the first month (Svetlitzky et al., 2020).

The challenging or threatening events that may lead to acute stress can be of diverse nature, such as wars, economic collapses, and health emergencies (Boehm-Tabib, 2016; Hauksdóttir et al., 2013; Naumann et al., 2017). The coronavirus pandemic goes beyond a merely biomedical event. It has had a health dimension, evidenced in fatality rates and disabilities (Smith, 2022), among other health outcomes, but it also has a socioeconomic dimension, evidenced in unemployment and job loss (Demena et al., 2022), among others. These health and socioeconomic outcomes may, subsequently, have psychological effects as well. Nevertheless, individuals are not condemned to passively suffer from the pandemic: they may cope with it resiliently. Resilient coping has been understood as a way of confronting adversity to rebound from and positively adapt to significant stressors (Sinclair & Wallston, 2004). Thus, a negative association can be expected between acute stress and resilient coping.

Fear states have been considered adaptive, since they promote problem solving and adaptation to danger (Rosen & Schulkin, 2004). However, these states may become overloaded and exceed individual abilities to adapt, leading to psychopathology (Rosen & Schulkin, 2004). Fear has been related to specific disorders, such as PTSD and other stress disorders, anxiety and mood disorders, among others (Carleton et al., 2011; Gonzalez & Martinez, 2014;Shin & Liberzon, 2010).

As previously mentioned, in the case of the coronavirus pandemic, there are several challenges or threats: on one hand, there are health risks, for you and significant others; on another hand, there are economic risks and outcomes of the pandemic. Therefore, when assessing fear that may result from the pandemic, both kinds of threatening events that may lead to fear should be considered: treats of negative health consequences of the pandemic and treats of negative economic consequences of the pandemic.

An observational cross-sectional study was carried out with the aim of developing measures on pandemic-related acute stress and fears. It was also aimed to test validity and reliability of these measures in a sample of Costa Rica’s inhabitants. As likely evidence of construct validity, we hypothesized a negative association between resilient coping and these new measures on pandemic-related acute stress and fear, as well as a positive association between acute stress and fear. This study was part of wider research on the impact of the pandemic on mental in Costa Rica, whose results are expected to be reported somewhere else.

Participants were 699 adults (18 years or older), inhabitants of Costa Rica, who voluntarily decided to fill out an online questionnaire on mental health consequences of the pandemic.

This questionnaire was advertised in a social network (Facebook). Each participant was presented first an informed consent form and expressed agreement before having access to the questionnaire. This study was approved by the Ethics Committee of the University of Costa Rica (reference number: CEC-470-2020).

Concerning the new scales tested in this study, they were developed as follows. A group of mental health professionals (psychologists, psychiatrists, and other mental health researchers) initially elaborated a set of items on 1) pandemic-related acute stress and a set of items on 2) pandemic-related fears. The items were elaborated considering what was deemed as contextually relevant for adult inhabitants of Costa Rica by these mental health professionals. Further details on these scales, and its development, is provided in the next paragraphs.

The study questionnaire included several scales. Two of these scales were new, to address specificities of the psychological reaction to the pandemic, and one of these scales was taken from previous literature and research. This last scale, on resilient coping, was included in order to allow testing of construct validity, by its correlations to the new scales.

Pandemic-related Acute Stress Scale (PASS). As mentioned, a set of items were elaborated by mental health professionals. They were focused on acute stress symptoms and was assumed to be one-dimensional. These items were introduced by a stem (“Since the COVID-19 Pandemic emergency began, have you presented any of the following symptoms?”), which was followed by 8 items, on different symptoms (e. g.: “Feeling restless, such as a feeling of not being able to sit still or wanting to run away”). Response options used a 5-point Likert format, from (1) “almost never” to (5) “almost always”. The complete set of 8 items, in its original Spanish language, and a free translation into English, is presented in Annex A, Table 3.

Alternatively, mental health professionals also developed and selected a different set of items, conceptualizing the acute stress as bidimensional (cognitive stress and physiological stress). We also tested them. Results on this are reported in Annex B.

Pandemic-related Fear Scale (PFS). A set of items was elaborated by mental health professionals, including what they deemed as two clusters of concerns related to the COVID-19 Pandemic: (1) concerns on health and (2) economic consequences of the pandemic. Items were introduced by a question (“To what extent do you experience the following fears in the current health emergency?”). There were 4 items focused on health-related fear (e. g.: “become infected and sick from COVID-19”). There were 2 items focused on economy-related fear (e. g.: “Facing material or economic losses”). Response options used a 5-point Likert format, from (1) “totally disagree” to (5) “totally agree”. The complete set of 8 items, in its original Spanish language, and a free translation into English, is presented in Appendix A, Table 4.

Brief Resilient Coping Scale (BRCS). This scale was originally developed and evaluated by Sinclair & Wallston (2004), where evidence of internal consistency (. = .69) and test-retest reliability (. = .71) was reported, as well as evidence of validity. In diverse cultural contexts and languages, including Spanish, this scale has been reported to be one-dimensional (Cosco et al., 2016; Kocalevent et al., 2017; Limonero et al., 2014; Tomás et al., 2012). It consists of 4 items, such as “I look for creative ways to alter difficult situations”, with a Likert 5-point response format, from (1) “does not describe me at all” to (5) “describes me very well”. In this study reliability was . = .79 in the exploratory subsample and . = .89 in the confirmatory subsample.

Sociodemographic variables. Information on several sociodemographic variables was also collected, such as age, gender, citizenship, family income, education, and province of residence.

About half of the sample (N = 347) was randomly selected to conduct, with this first subsample, exploratory factor analysis (estimation method: Maximum Likelihood). Data from the second half of the sample (N = 352) was then analyzed by confirmatory factor analyses (estimation method: Maximum Likelihood). In other to assess fit acceptability, a set of indices and their corresponding cutoff values were chosen, based on previous recommendations (Cangur & Ercan, 2015; Hu & Bentler, 1999). These indices and values were the following: the ratio χ2/df, where values close to 3 are considered acceptable (Cangur & Ercan, 2015); the Comparative Fit Index (CFI), a measure of incremental fit, where values of .90 have been traditionally used as a cutoff, although more recently, values close to .95 are preferred (Cangur & Ercan, 2015; Hooper et al., 2008; Hu & Bentler, 1999); and the Root Mean Square Error of Approximation (RMSEA), a measure of absolute fit, where values of .06 or lower are deemed as ideal (Hu & Bentler, 1999), although lower than .08 are still acceptable (Cangur & Ercan, 2015).

Reliability analyses were also conducted, for each scale and subscale. A Cronbach’s Alpha (.) ≥ .70 was considered acceptable (Cortina, 1993).

Finally, to test for evidence of construct validity, correlations of these new scales with resilient coping were also computed. As mentioned in the introductory section, there is a theoretical rationale to expect negative associations between resilient coping and both, acute stress and pandemic-related fear. A positive association between acute stress symptoms and fears was also expected.

Overall, the sample (N = 699) consisted of 74.1 % women (25.6 % were men, 0.1 % non-binary), with a mean age of 36.05 y. o. (S. D. = 13.31, Q. = 25, Q. = 34; Q. = 44). Most were Costa Ricans (96.6 %). Most participants (55.7 %) reported a monthly family income ≤ ₡ 1000000 (approx. $ 1577). Most reported to have an educational level of either uncomplete or complete college degree (81.2 %). The province of residence that contributed the most participants to the study was San José (40.8 %), where the capital, San José city, is located. Since both subsamples (N = 347 and N = 352) were generated randomly, there were no differences between them in terms of sociodemographic composition (. > 0.5).

PASS. The results from the first subsample, based on an EFA for the scale on acute stress during the pandemic, suggested it to be one-dimensional: only one factor had an eigenvalue < 1. This factor explained 49.09 % of variance. Reliability of the PASS in this subsample was . = .85, and can, then, be considered acceptable. All item-total corrected correlations were between . = .47 and . = .67, what is considered excellent (Qin, 2006), and provides further evidence of internal consistency.

Results based on a CFA from the second subsample also supported a one-dimensional model, where an acceptable fit was found: χ2/df = 2.72, CFI = .96, RMSEA = .070, 90 % CI [.048; .093]. Reliability in this subsample was . = .86, what is deemed as acceptable.

Table 1 summarizes information on item means and factor loadings from both subsamples and factor analyses. No loading was found to be . ≤ .50.

Table 1
Descriptive statistics and factor loadings of the PASS items

In the EFA: KMO = .872, Bartlett test = 944.52 (p < .001).

Concerning construct validity of the PASS, a correlation of . = -.36 (. < .001) was found between this new acute stress measure and the BRCS in the first sample, and a correlation of . = -.39 (. < .001) was found in the second sample. These correlations support what was hypothesized based on theoretical grounds.

PFS. For the new scale on pandemic-related fears, results from the first subsample, based on an EFA suggested a bidimensional solution: two factors had eigenvalues > 1. The first eigenvalue explained 55.87 % of variance, while the second one explained 17.41 % of variance. Cumulatively, both factors explained 73.29 % of variance. The correlation between these factors was . = .47 (. < .001). Reliability of the health-related fear subscale was . = .85. Reliability of the economy-related fear subscale was . = .75 (._SB = .75). Reliability of these subscales was then considered acceptable. Item-total correlations in the health subscale were between . = .67 and . = .72. In the economic subscale these correlations were . = .59. This is considered excellent (Qin, 2006), and provides further evidence of internal consistency.

Results based on a CFA from the second subsample did not completely support the bidimensional model suggested previously by the EFA. Its fit indices were as follows: χ2/df = 4.43, CFI = .96, RMSEA = .099, 90 % CI [.067; .133]. As seen, although relative fit (CFI) was acceptable, both, the χ2/df ratio and absolute fit (RMSEA) did not reach cut-off levels recommended in the literature (Cangur & Ercan, 2015; Hu & Bentler, 1999).

Modification indices suggested that, within the subscale on health-related fear there were items with correlated residuals, and two likely subdimensions: fear concerning health of significant others (items 1 & 2, MI = 10.760) and fear concerning health of your-self (items 3 & 4, MI = 12.573). Therefore, this model was specified so that the economic subscale remained the same (two items with one first-order factor), but the health subscale was modified so that there were two first other factors (fear related to self-heath and fear related to other’s-health), and one second-order factor (overall health-related fear).

The fit of this modified model was found acceptable: χ2/df = 1.77, CFI = .99, RMSEA = .047, 90 % CI [.000; .089]. The correlation between the economic fear factor and the overall health fear factor was . = .43 (. < .001). The correlation between both health-related fear subscales was . = .61 (. = < .001). These correlations were not large enough as to suggest important overlap between variables (Tabachnick & Fidell, 2014). Reliability in this subsample was . = .81 for overall health fear (other’s health fear, . = .77, ._SB = .77; self-health fear, . = .71, ._SB = .71). For economic fear, reliability was . = .79, ._SB =.80. Thus, internal consistency of these subscales can be deemed acceptable.

Table 2 summarizes information on item means and factor loadings from both subsamples and factor analyses. No loading was . ≤ .65.

Table 2
Descriptive statistics and factor loadings of the PFS items

In the EFA: KMO = .810, Bartlett test = 860.98 (p < .001), items presented a clear one-dimensional pattern (there was a difference in loadings of, at least, .20 between factors). In the CFA, the loadings of the second-order factor (overall health fear) on first-order factors were β = .90 (Other´s health) and β = .91 (Self-health).

Concerning construct validity of the PFS, in the first subsample a correlation of . = -.21 (. < .001) was found between economy-related fear and the BRCS, as well as a correlation of . = -.24 (. < .001) between health-related fear and the BRCS. In the second subsample a correlation of . = -.10 (. < .05) between economic fear and the BRCS, as well as a correlation of . = -.21 (. < .001) between health-related fear and the BRCS.

Additionally, in the first subsample the correlation between economic fear and the PASS was . = .43 (. < .001), and the correlation between health-related fear and the PASS was .56 (. < .001). In the second subsample, the correlation between economic fear and the PASS was . = .38 (. < .001), and the correlation between health-related fear and the PASS was . = .62 (. < .001). These correlations provide additional evidence of construct validity for both, the PASS and the PFS.