This was a cross-sectional study conducted using an online survey questionnaire from April 30 to June 9, 2020. For this purpose, a specific questionnaire was created at the www. Google forms.com, and a quick response (QR) code was designed specifically for the link to the online survey. The Hospital d Clínicas, Universidade Federal do Paraná (HC-UFPR), Brazil, review board approved the study. The online survey link was distributed randomly in personal and professional networks (communities and web groups) and email. The response to the forms was received online. Participation in the survey was voluntary and completely anonymous, without the possibility of anonymity break. The electronic form could be accessed only once by each google email address.
Assessment of fear
Development of the scale Relevant and possible items, which could expose the person to risk of infection by SARS-CoV-2 on daily life necessities and activities were identified by two authors (SMA and CPV). After removing items with similar content or not specifically related to the construct “fear”, 15 items were selected; on a second review by the same two authors four items were excluded. This resulted in the 11 items scale, which was circulated on-line. The semantic evaluation of the scale was performed by the researcher. Initial assessment of scaleParticipants from the same institution of the researchers and self-identified as responders, answered the online form on the initial day of circulation [06 women and 01 man; median (IQR): age 51.00 (42.50; 55.00) years; years of education 22.00 (18.00; 23.50)]. They were asked to obtain initial assessment of the scale, as if there were any problems in understanding the items themselves; their thoughts about the scale items, accessibility to the form through the link or QR code, and time to answer the form, which was reported one to two minutes. These participants indicated no changes. The authors propose an English version of the SCoV-2-FS. For the English version of the SCoV-2-FS, forward-backward translation method was applied (Table 1). The items were translated into English by a Portuguese native speaker fluent in English, reviewed and corrected by an English native speaker, further translated back into Portuguese by another bilingual Portuguese native speaker fluent in English professional who had not seen the Portuguese version of the SCoV-2-FS (back translation). SARS-CoV-2 Fear Scale (SCoV-2-FS)The fear scale, which was titled SARS-CoV-2 Fear Scale (SCoV-2-FS), comprised 11 closed-ended questions, in Portuguese, focused on the fear of the risk of infection with SARS-CoV-2, on daily life necessities and activities, such as, going out on the street, going to the market, pharmacy, hospital, using an elevator, contact with other people, and public transport. Fear about being infected with SARS-CoV-2, in person or someone close. Fear to die of complications of SARS-CoV-2, in person or someone close (Table 1).
The participants indicated their level of fear using a five-item Likert-type scale, which was used to measure the level of fear in the study. The degree of fear was reflected using 1–5 point, with higher scores indicating greater fear. Details of the scale are as follows: 1, no fear; 2, mild; 3, moderate; 4, a lot; and 5, severe. A total score was calculated by adding up each item score (range, from 11 to 55).
Demographic data and risk factors
Participants were stimulated to report birth date, gender, years of education, religion, city, state, country, and ZIP code. Identification of risk factors for poor prognosis of COVID-19 or SARS-CoV-2 infection (yes/no): necessity to go to work, health workers, and existence of comorbidities (chronic diseases).
Participants
The target population was the general Brazilian population. Participants were randomly recruited online in personal, professional networks, and emails. The participant was asked to disclose the electronic questionnaire among their contacts and social networks (snowball sampling strategy). Inclusion criteria: participants aged 18 years or above, literate, fluent in Portuguese, and familiar with the use of social networking. Exclusion criteria: responses from participants under 18; incomplete responses, or participants from other counties were excluded.
Statistical analyses
Categorical variables were described by frequency of distribution (n, %), while continuous variables were described by median (IQR). Analysis of the psychometric properties of the SCoV-2-FS was described by Skewness and Kurtosis to check for asymmetry and kurtosis were in the range from -1 to +1 in the case of normal univariate data distribution. To check if the normal distribution model fits the observations, the Shapiro–Wilk normality test was used.
Reliability (internal consistency)
Reliability (i.e. internal consistency) of the SCoV-2-FS was examined by Composite reliability (cutoff >0.7; Netemeyer et al., 2003); Cronbach’s α (cutoff >0.9; Barret and Kline, 1981), Guttmann ɣ 6 (G6, smc), McDonald´s ω (0 .70; Graham, 2006).
To confirm the factor structure (model fit), the SCoV-2-FS was examined with the Confirmatory factor analysis (CFA; Kline, 2015). Several indices were used to assess model fit, including standardized root mean square residual (SRMR, cutoff <0.08), comparative fit index (CFI, cutoff >0.9), goodness-of-fit index (GFI, cutoff >0.8; Doll et al.,1994); normed of fit index (NFI, cutoff >0.9), and standardized factor loading. To evaluate how each individual item affects the reliability of the scale, the measurements were calculated on the overall score and removing each of the items of the scale. The correlation coefficients between each item of the SCoV-2-FS were calculated with a standardized variance/covariance matrix (correlation matrix, using Pearson's correlation coefficient). Statistical significance was considered as p < 0.05.
The cutoff of SCoV-2-FS was calculated by the confidence interval using the Z statistic, determined by confidence level (Z= 1.96), and sample mean to generate an interval estimate of the population mean.
The analysis was carried out using the statistical package R software (R Core Team, 2019, Vienna, Austria) version 3.6.1.