Validation and Psychometric Properties of the Portuguese Version of the Coronavirus Anxiety Scale (CAS) and Fear of COVID-19 Scale (FCV-19S) and Associations with Travel, Tourism and Hospitality

The aim of this study is to determine the anxiety and fear related to coronavirus disease 2019 (COVID-19) and their associations with travel, tourism and hospitality, in the Portuguese population. The Coronavirus Anxiety Scale (CAS) and Fear of COVID-19 Scale (FCV-19S) were validated for the Portuguese population and correlations with issues related to travel, tourism and hospitality were established. CAS and FCV-19S presented a good adjustment model and solid reliability and validity. Correlations between CAS and FCV-19S and the perception of the impact of COVID-19 in travel, tourism and hospitality were found. Participants considered that COVID-19 mainly affected their holidays and leisure time. However, the strongest correlation established was between total FCV-19S and emotional fear FCV-19S and the fear of attending hotel facilities. The Portuguese versions of CAS and FCV-19S are reliable psychological tools to assess anxiety and fear in relation to COVID-19 for the general population. The use of hotel facilities is the most threatening issue related to travel, tourism and hospitality. The results suggest that hotels should invest in hygiene and safety measures that allow users to regain confidence in hotel equipment.


Introduction
The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the virus causing coronavirus disease 2019 (COVID- 19), the disease that was first discovered in 2019 in China [1]. As of December 28th, 2020, there were 80,879,693 people infected with this virus around the world, and 1,766,787 died from it. In Portugal, 394,573 people were infected, and 6619 died [2]. Thus, the COVID-19 pandemic is a massive global health crisis [3]. The most common symptoms at the onset of COVID-19 are fever, cough and fatigue [4], although other symptoms include headache, hemoptysis, diarrhea, dyspnea and lymphopenia [5]. Being male, elderly and having comorbidities have been significantly associated with the risk of death among COVID-19 patients [6].
As a result of the emergence of the COVID-19 outbreak, a socio-economic crisis and profound psychological distress occurred worldwide [7]. People who became infected Permission was received from the original authors [10,16] to validate the instruments in a Portuguese population. It was then translated from English to Portuguese using the back-translation technique [35] (Appendix A, Tables A1 and A2; Appendix B,  Tables A3 and A4). After the protocol was conceived, including the sociodemographic questionnaire, issues related to travel, tourism and hospitality, CAS and FCV-19S, it was submitted to the ethics committee of the University of Trás-os-Montes and Alto Douro (UTAD), having been approved in 1 September 2020. It was released to the general population through a social network page about the study, with data being collected between 1 October and 15 November 2020. Convenience and snowball samplings were used. Consequently, the sample is not representative of the Portuguese population. All procedures performed in this study were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki Declaration and its later amendments. Informed consent, in which the participants were familiarized with the aims of the study and in which the confidentiality and anonymity of the data were guaranteed, as well as the strategy to destroy data after being used, preceded the investigation protocol and the participants only accessed it after giving their consent.

Sociodemographic Questionnaire
This questionnaire included questions related to gender (man vs. woman), age (numerical), education (no university studies vs. university studies) and employment status (inactive-unemployed, sick, retired, on medical leave-vs. active-students, employees, housewives).

Questions Related to Travel, Tourism and Hospitality
Seven questions were designed to assess the participants' perceptions of the impact of COVID-19 on travel, tourism and hospitality in times of the COVID-19 pandemic. The instruction ("On a scale of 0 to 100, please indicate how much the pandemic situation caused by COVID-19 has . . . ") proceeded with the items: 1-. . . "changed your leisure activities"; 2-. . . "changed your vacations"; 3-. . . "prevented you from settling in a hotel"; 4-. . . "prevented you from traveling by plane"; 5-. . . "prevented you from traveling by train"; 6-. . . "prevented you from traveling by car"; and 7-. . . "made you feel fear of attending hotel facilities".

Coronavirus Anxiety Scale (CAS)
The CAS was developed with the purpose of filling a void in the mental health response to COVID-19 [16]. According to Lee [16], a brief mental health screener that could identify probable cases of dysfunctional anxiety and symptom severity associated with the coronavirus was needed. This is a five-item scale that assesses distinct physiological reactions of anxiety related to COVID-19, highly reliable as a cluster (α = 0.93) [16].

Analytical Approach
An internal replicability approach was employed by subjecting one half of the study's data to an exploratory factor analysis (EFA) and the other half to confirmatory factor analysis (CFA) to address the sampling error influences. The EFA was used to identify representative symptoms of coronavirus anxiety, while the CFA was used to test the replicability of the EFA results. The second half of the study's data was also used to perform a series of multiple group CFAs to determine if the construct of coronavirus anxiety presents differences across demographic groups. Pearson correlations between questions related to travel, tourism and hospitality and anxiety and fear towards COVID-19 were performed, as well as Spearman correlations between anxiety and fear and sociodemographic variables. Test-retest reliability was calculated using Pearson correlations to assess the CAS and FCV-19S constructs' stability and precision across time. According to the guidelines suggested by Vaz et al. [36], if the p-value is less than 0.05, and the Pearson correlation coefficient is above 0.7, then researchers have evidence of test-retest reliability. Statistical analyses were calculated using Statistical Program for Social Sciences SPSS version 27.0 (IBM Corp., Armonk, NY, USA), and CFA was run using AMOS version 27.0 (IBM Corp., Armonk, NY, USA).

Sample
Two independent samples with the same number of participants were used. As a whole, the sample consisted of 1122 participants, of whom 725 (64.6%) are women, with a mean age of 31.91 years of age (SD = 13.76), with 495 (44.1%) having university studies and the remaining (n = 627; 55.9%) without.
Concerning professional status, 932 (83.1%) are active and the remaining are inactive. The EFA sample (n = 561) was not different from the CFA sample (n = 561) in relation to sociodemographic issues (chi-square tests and Student's t tests), except for age [t(1111, 613) = 2.22; p = 0.027; d = 0.13], being that the EFA sample was slightly older (M = 32.82 years old; SD = 14.32) than the CFA sample (M = 31.00 years old; SD = 13.12).
The five items of the CAS were subjected to an EFA with varimax rotation. The maximum likelihood factor analysis with a cut-off point of 0.40 and the Kaiser's criterion of eigenvalues greater than 1 [39] yielded a one-factor solution as the best fit for the data, accounting for 67.64% of total variance explained. The five items meet the criteria for psychometrically sound items (Table 1). Structure coefficients ranged from 0.73 to 0.89, and communality coefficients ranged from 0.53 to 0.79. Correlations between items ranged from 0.45 to 0.73. These items were reliable as a single dimension (α = 0.85). If any item was deleted, alpha's value decreased.
The test-retest agreement was analysed item by item between the first and second (n = 31) evaluations. The correlations between the first and the second moments were all over r = 0.70 and the significance was always above p = 0.50. Notes: LD = structure coefficients; h 2 = extracted communality coefficients; M = mean; SD = standard deviation; S kw = skewness; K rt = kurtosis; Min = minimum; Max = maximum; CAS = Coronavirus Anxiety Scale.
The seven items of FCV-19S were subjected to an EFA with varimax rotation. The maximum likelihood factor analysis with a cut-off point of 0.40 and the Kaiser's criterion of eigenvalues greater than 1 [39] yielded a two-factor solution as the best fit for the data, accounting for 70.72% of total variance explained. The seven items meet the criteria for psychometrically sound items ( Table 2). Structure coefficients ranged from 0.71 to 0.86, and communality coefficients ranged from 0.57 to 0.79. Correlations between items ranged from 0.37 to 0.72. These items were reliable as a single dimension (α = 0.88) (if any item was deleted, alpha's value decreased) and as two-factor dimensions (first one α = 0.83; second one α = 0.82).
The test-retest agreement was analysed item by item between the first and second (N = 31) evaluations. The correlations between the first and the second moments were all over r = 0.70 and the significance was always above p = 0.50. Data screening results suggested that the seven items were suitable for EFA [37]: again, no issues relating to sample size, missing data, nonnormality, multicollinearity or singularity. The correlation matrices were also factorable [Bartlett's test of sphericity = p < 0.001; Kaiser Meyer-Olkin (KMO) test = 0.85)].
The seven items of FCV-19S were subjected to an EFA with varimax rotation. The maximum likelihood factor analysis with a cut-off point of 0.40 and the Kaiser's criterion of eigenvalues greater than 1 [39] yielded a two-factor solution as the best fit for the data, accounting for 70.72% of total variance explained. The seven items meet the criteria for psychometrically sound items (Table 2). Structure coefficients ranged from 0.71 to 0.86, and communality coefficients ranged from 0.57 to 0.79. Correlations between items ranged from 0.37 to 0.72. These items were reliable as a single dimension (α = 0.88) (if any item was deleted, alpha's value decreased) and as two-factor dimensions (first one α = 0.83; second one α = 0.82).
The test-retest agreement was analysed item by item between the first and second (N = 31) evaluations. The correlations between the first and the second moments were all over r = 0.70 and the significance was always above p = 0.50. To test the model found in EFA, a CFA was performed. The results supported the EFA findings (Figure 2). A two-factor model was found [χ 2 (11) = 39.56, p < 0.001] with a moderate fit for all of the indices [χ 2 /df ratio = 3.58; CFI = 0.99; TLI = 0.97; SRMR = 0.03; RMSEA = 0.07 (0.00, 0.05; 90% CI); PCLOSE = 0.09]. However, to achieve this model, a correlation between two items' errors (items 2 and 4) of the first factor and a correlation between two items' errors of the second factor (items 3 and 6) were established. As the items whose errors were correlated belonged to the same factor, in theory, the correlation is justified. To test if the fear of COVID-19 construct was measured the same way across genders (women vs. men), multigroup CFAs were performed. The results demonstrated gender differences, which were evidenced by the model fit [χ 2 (22)   To test the model found in EFA, a CFA was performed. The results supported the EFA findings (Figure 2). A two-factor model was found [ 2 (11) = 39.56, p < 0.001] with a moderate fit for all of the indices [χ 2 /df ratio = 3.58; CFI = 0.99; TLI = 0.97; SRMR = 0.03; RMSEA = 0.07 (0.00, 0.05; 90% CI); PCLOSE = 0.09]. However, to achieve this model, a correlation between two items' errors (items 2 and 4) of the first factor and a correlation between two items' errors of the second factor (items 3 and 6) were established. As the items whose errors were correlated belonged to the same factor, in theory, the correlation is justified. To test if the fear of COVID-19 construct was measured the same way across genders (women vs. men), multigroup CFAs were performed. The results demonstrated gender differences, which were evidenced by the model fit [χ 2 (22)    Correlations between the seven questions about travel, tourism and hospitality range from r = 0.19 to r = 0.63, being all the correlations significative at the p < 0.001 level. Cronbach's alpha suggests good reliability (Table 3). All questions concerning travel, tourism and hospitality correlate positively with the CAS and FCV-19S and their dimensions ( Table 3). The highest correlations were found between FCV-19S and emotional fear, on the one hand, and attending hotel facilities, on the other hand. The correlations between the same dimensions and avoiding settling in a hotel follow. However, if paying attention to the means of the items related to travel, tourism and hospitality (Table 3), it turns out that the highest mean relates to the impact of COVID-19 on vacations and leisure time; the lowest mean refers to traveling by car.

Discussion
This study aims to determine the anxiety and fear associated with COVID-19 and their associations with travel, tourism and hospitality, in the Portuguese population. The study is particularly important because there were no tools to assess these constructs for the Portuguese population. Accordingly, two instruments to evaluate anxiety and fear in relation to COVID-19 were assessed and validated. CAS [16] is a five-item scale assessing distinct physiological reactions of anxiety related to the coronavirus. FCV-19S [10] is a seven-item scale assessing fear of COVID-19 among the general population. In both instruments, a high score means more anxiety and more fear, respectively. Both CFAs of the original one-factor model for CAS and the two-factor model for FCV-19S, proposed by the authors of the original scale [10,16], showed a good fit with the most important indices, thus confirming the first stated hypothesis, H1. Additionally, both instruments revealed good internal consistency for the global score and the two subscales of the FCV-19S. The findings are consistent with those found for CAS in a significant number of studies [16][17][18][19][20][21][22]24,25,[27][28][29][31][32][33][34].
The validated tools were subsequently associated with questions related to travel, tourism and hospitality. All questions concerning travel, tourism and hospitality correlated positively with the CAS and FCV-19S and their dimensions, confirming the second hypothesis, H2, and corroborating previous studies [15,40,41]. Participants reported that the items most impacted by COVID-19 were vacations and leisure time, and the item lesser impacted was traveling by car, corroborating these results in the literature concerning leisure [42] and traveling by car [43]. In fact, leisure activities and holidays have undergone profound changes. However, although care shares have lowered in frequency, people continue to use their car alone to go to work. The strongest correlation was found between FCV-19S and emotional fear, on the one hand, and attending hotel facilities, on the other hand. Interestingly, this correlation with emotional fear suggests that, regardless of the measures that the hotel industry may be willing to carry out, emotional fear overlaps, which does not mean that the hotel industry should not continue to invest in hygiene and safety measures that guarantee users to regain confidence in such equipment, aiming to overcome resistance from clients.
These results are just a glimpse of how this pandemic has affected people's daily lives. All changes are generating anxiety and fear, especially when they are unwanted, as most of the changes that people have had to carry out in their life related to COVID-19 [43]. The change process alone causes anxiety. When the reason for the change causes fear, this overlap of anxiety and fear can be quite disturbing and can have lasting negative consequences. In fact, fear increases anxiety in healthy individuals and intensifies the symptoms of those with pre-existing psychiatric disorders.

Conclusions
In conclusion, this study provides the CAS and FCV-19S Portuguese versions as reliable and valid instruments, useful for measuring anxiety and fear related to COVID-19. The instruments showed good fit indices in the factor structure. The results also show good consistency indices for global scores and FCV-19S subscales. As far as it is known, these are the first instruments validated in a Portuguese population which evaluate anxiety and fear related to COVID-19. This study has some limitations. It is a cross-sectional design that hinders interpreting causality, and the questionnaire was self-applied. Future studies should test the fit of the instruments in a clinical sample.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by Ethics Committee of the University of Trás-os-Montes and Alto Douro (UTAD), having been approved in 1 September 2020 (no specific reference assigned, date acting as reference ID).

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to General Data Protection Regulation (GDPR).

Conflicts of Interest:
The authors declare no conflict of interest.

. English Version
Please indicate your level of agreement with the statements:  Por favor, assinale a sua concordância ou discordância em relação às seguintes afirmações:

Appendix B.1. English Version
How often have you experienced the following activities over the last 2 weeks?  4. I lost interest in eating when I thought about or was exposed to information about the coronavirus. 5. I felt nauseous or had stomach problems when I thought about or was exposed to information about the coronavirus.

Appendix B.2. Portuguese Version
Tendo passado mais de uma hora (no total) por semana a pensar no coronavirus 19 e a ver notícias nas redes sociais e na televisão sobre ele, por favor indique em quantos dias teve os seguintes sintomas: