Systematic Review and Meta-Analysis of Fear of COVID-19

Background: Due to lack of preparedness of health systems, fast spread of the new virus, high mortality rates, and lack of a definite treatment, the outbreak of Coronavirus disease (COVID-19) led to high levels of fear and anxiety in different populations. In addition, isolation, mental disorders, and limitations in social interactions as a result of lockdown and travel ban increased the fear of the new coronavirus. Methods: International databases, including Scopus, PubMed, Web of Science, and Google scholar, were searched without any time limitation, and all observational studies published in English reporting the mean of fear of COVID-19 based on the Fear of COVID-19 scale (FCV-19S) were included in the analysis. Methodological quality was assessed using the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. Random effects model, subgroup analysis, and meta-regression analysis were used to analyze the data. Heterogeneity across studies was examined using Cochran's Q test and I2 statistic. All the statistical analyses were conducted using R software v4.0.3. Results: A total of 44 articles with a sample size of 52,462 were reviewed. A pooled mean of 18.57 was found for fear of COVID-19. The mean of fear of COVID-19 was higher in women than in men (20.67 vs. 18.21). The highest and lowest means of fear of COVID-19 had been found in Asia (18.36) and Australia (17.43) based on continent, and in hospital staff (19.51) and college students (17.95) based on target population, respectively. In addition, the highest and lowest means of fear of COVID-19 were related to items #1 and #3 of the scale, respectively. According to the results of meta-regression analysis, there was no significant association between the mean of fear of COVID-19 and sample size and participants' age. In addition, publication error was not significant (P = 0.721). Conclusion: The mean of fear of COVID-19 was high around the world; therefore, it seems necessary to pay more attention to the negative effects of the COVID-19 pandemic on mental health.


INTRODUCTION
The Coronavirus disease  pandemic was first reported in Hubei, China, in December 2019. So far, it has affected about six million people and has led to the death of more than 360,000 people around the world mostly due to severe acute respiratory illness (Ashamalla et al., 2020). Given the lack of an effective treatment for COVID-19, different countries around the world focused their efforts on reducing the risk of transmission through implementing public health measures, such as social distancing, self-isolation, and regular hand washing. In addition, unprecedented measures, such as controlling borders, contact tracing, and lockdown were taken (Ahorsu et al., 2020a;Alyami et al., 2020). These measures led to widespread fear so that in some countries people started to stockpile staple foods, toilet paper, and even guns (Bakioglu et al., 2020;Skoda et al., 2020). As the prevalence of COVID-19 increased, people started to isolate themselves, limit their social interactions, and avoid others for fear of getting the virus (Abuhammad et al., 2020). Fear is an adaptive response to one's environment and a defense mechanism to increase the chance of one's survival; however, it can be maladaptive when it is not proportionate to the actual threat (Steimer, 2002).
During the Severe Acute Respiratory Syndrome (SARS) and Ebola outbreaks, public fear worsened the negative effects of the actual illness; therefore, one of the most important challenges in the face of outbreaks is to control social reactions (García-Reyna et al., 2020). In order to reduce possible psychological problems, researchers recommend that the level of fear, worry, and helplessness associated with COVID-19 should be examined, because high levels of stress may prevent one from making logical decisions to protect themselves (Ahorsu et al., 2020a). For example, some patients who need medical care may refuse to go to the hospital due to experiencing illogical levels of fear (WongLaura et al., 2020). Some patients may postpone their surgical treatment for fear of contracting the virus (Vanni et al., 2020). In some cases, fear of COVID-19 can lead to hypochondriasis, so that some people may misinterpret their bodily sensations and attribute them to COVID-19 (Coelho et al., 2020). Some people may also excessively use medications recommended in COVID-19 treatment guidelines, such as hydroxychloroquine (Banerjee, 2020). On the other hand, fear can act as a motivator of behavioral change in the face of COVID-19 (Harper et al., 2020;Pakpour and Griffiths, 2020). The experience of fear can increase risk perception and reinforce protective behaviors, such as washing of hands and keeping physical distance (Broche-Pérez et al., 2020). When people take a threat seriously, they can perform preventive measures more efficiently, and perception of threat as a motivator facilitates the prevention of COVID-19. Harper et al. found that fear of COVID-19 strongly predicted improved social distancing and hand washing and had an important role in adherence to public health measures related to COVID-19 (Harper et al., 2020).
The COVID-19 pandemic has led to fear and negative emotions; however, it has also had positive consequences, such as encouraging people to engage in ethical behavior (Jian et al., 2020). According to what was explained above, measuring fear of COVID-19 has an important role in understating the implications of the pandemic for mental health and in designing interventions to reduce COVID-19 fear and anxiety. One of the most important instruments available to assess fear of COVID-19 is the Fear of COVID-19 Scale (FCV-19S) that has been translated to many languages. Studies form different parts of the world examining fear of COVID-19 have led to different results. Therefore, the goal of the present systematic review and metaanalysis is to estimate the pooled mean of fear of COVID-19 around the world.

METHODS
The present systematic review and meta-analysis aimed to estimate the pooled mean of fear of COVID-19 based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Moher et al., 2009).

Data Sources and Search Strategy
Search for articles was conducted in September 20, 2020, in databases of Web of science/ISI, PubMed, and Scopus using the following keywords: Wuhan Coronavirus, Sars-cov-2, 2019 Novel Coronavirus, COVID-19 Virus, Coronavirus Disease 2019 Virus, Wuhan Seafood Market Pneumonia Virus, Fear, and all possible combination to increase search sensitivity. In addition, references of the selected articles were reviewed to access more related articles.

Selection Criteria
All observational studies published in English examining the state of fear of COVID-19 using the FCV-19S were analyzed. This scale developed by Ahorsu et al. assesses fear of COVID-19 using seven items that are rated on a Likert-type scale ranging from 5 (totally agree) to 1 (totally disagree). Total score on this scale ranges from 7 to 35, and higher scores indicate stronger fear of COVID-19 (Ahorsu et al., 2020a). The inclusion criteria were as follows: participants aged at least 18 years and reporting the mean and standard deviation of fear of COVID-19 score. Articles with unavailable full texts, preprinted articles, and articles not reporting the fear of COVID-19 score were excluded from the analysis.

Data Collection
In the first step, two independent authors screened the articles and selected those having the aforementioned keywords in their titles or abstracts. Then, they extracted the article information and recorded it in a predesigned Excel sheet. This information included first author's name, publication year, mean age of patients, target population, mean and standard deviations of fear of COVID-19 (total score and score by gender). Because all the selected articles had been published in 2020, publication year was not included in the table presenting article information.

Quality Assessment
The two researchers independently evaluated the quality of the articles based on 10 items of Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist (title and abstract, study environment, objectives and hypotheses, sample size, inclusion criteria, statistical methods, descriptive data, interpretation of findings, limitations, and funding). Higher scores indicate better methodological quality. In terms of methodological quality, articles were divided into three categories: poor (4 or below), moderate (4 to 7), and good (over 7) (Vandenbroucke et al., 2007).

Analysis
Random effects model was used to estimate the pooled raw mean of fear of COVID-19. A forest plot was used to visually depict heterogeneity across studies in which the mean of fear of COVID-19 with a 95% confidence interval and also the pooled raw mean of the selected studies are reported. Heterogeneity across studies was assessed using I 2 statistic and Cochran's Q test. I 2 percentages of 25%, 50%, and 75% show low, average, and high heterogeneity, respectively, and in Cochran's Q test, P < 0.1 indicates significant heterogeneity (Higgins et al., 2003). Source of heterogeneity across studies was examined using subgroup analysis by gender, continent (Asia/America/Europe/Australia/multi-countries), and target population (general population, college students, pregnant women, and medical staff).
The association of the mean of fear of COVID-19 with sample size and mean age of participants was assessed using meta-regression analysis. To ensure that the meta-regression results were not affected by one or several articles, leave-one-out sensitivity analysis method was used that involved performing the analysis on the data by leaving out one study at a time. For studies reporting scores by drop out one study at a time and estimate the pooled raw mean of remained studies. In addition, the potential effect of small studies was assessed using a funnel plot based on Egger's regression test (Egger et al., 1997). It is worthwhile to note that all analyses were performed based on random effects model using R software, version 4.0.3, and that all statistical tests were two-tailed ones. In addition, the significance level for all tests except for examination of heterogeneity across studies was set at P < 0.05.

RESULTS
In the primary search, a total of 634 articles were found, of which 320 duplicate articles were excluded, and titles and abstracts of the remaining articles were reviewed. In addition, 270 articles focused on unrelated subjects were removed from the analysis. Figure 1 shows the flowchart of selecting and screening articles based on the PRISMA guidelines.
A total of 44 articles with a sample size of 52,462 were included in the final analysis. As shown in Table 1, among the 44 articles, 33 reported the total score of fear of COVID-19, 8   (Abad et al., 2020;Bakioglu et al., 2020;Broche-Pérez et al., 2020;Haktanir et al., 2020;Hossain et al., 2020;Mertens et al., 2020;Nguyen et al., 2020a;Sakib et al., 2020). These studies had been conducted with the general population, college students, pregnant women, and medical staff. Detailed information on selection of articles is provided in Table 1. In terms of methodological quality, eight studies had medium quality, and the remaining articles had excellent quality. According to the level of heterogeneity across studies, random effects model was used to combine the studies (P < 0.0001; Q = 8243.69, df = 32, P < 0.0001, τ 2 = 7.9730, and I 2 = 99.6%). In the present study, the pooled mean of fear of COVID-19 was found to be 18.57 (95% CI: 17.60-19.54). In addition, the prediction interval for the pooled mean of fear of COVID-19 was found to be 12.72-24.42 (Figure 2). Among the selected studies, 17 reported the mean of fear of COVID-19 by item; the highest and lowest mean scores were related to items #1 (3.32) and #3 (1.78), respectively. It is worthwhile to note that the mean of fear of COVD-19 for all items was higher in Asian studies compared to those conducted in other continents. The Frontiers in Psychology | www.frontiersin.org lowest scores on items #1, #3, #5, and #6 were reported by the American studies, the lowest scores on items #4 and #7 were reported by Australian studies, and the lowest scores on item #2 were reported by European studies. In addition, the mean of fear of COVD-19 on all items except for items #1 and #4 was higher in the general population than in college students. The pooled mean of fear of COVD-19 by item is presented in Table 2 ( Supplementary Figures 1-7).

Subgroup Analysis
The pooled raw mean of fear of COVID-19 was examined using random effects model. According to the results of subgroup analysis, the highest and lowest pooled means of fear of COVID-19 were reported in the studies conducted in Asia (18.36, and Australia (17.43,), respectively (Supplementary Figure 8). In addition, the pooled raw mean of fear of COVID-19 was higher in medical staff than in other groups (Supplementary Figure 9).
According to the results of subgroup difference test reported in Table 3, there was a significant difference in the mean of fear of COVID-19 in different continents (P = 0.0347), but there was no significant difference between different target populations (P = 0.0773). In addition, eight articles reported the mean of fear of COVID-19 by gender that was lower in men (18.21,) than in women (20.67, 95% CI: 18.62-22.73) (Supplementary Figures 10, 11). Moreover, the mean fear of COVID-19 was 17.68 and 19.70 in Asian men and women and 16.15 and 20.39 in American men and women, respectively (Supplementary Figures 12, 13).
The results of meta-regression analysis showed that mean score of fear of COVID-19 increased with mean age, but the relationship was not statistically significant (P = 0.797) (Figure 3 and Table 4).
Results of sensitivity analysis based on random effects model showed that none of the studies alone had a substantial impact on the pooled raw mean of fear of COVID-19 (Supplementary Figure 14). Results of Egger's regression test also indicated that publication bias was not statistically significant (P = 0.721) (Supplementary Figure 15).
Examination of the pooled mean by continent and target population showed that on all items, it was higher in the studies conducted in Asia compared to those conducted in other continents. In addition, the pooled mean of fear of COVID-19 on all items except for items #1 and #4 was higher in the general population than in college students (Table 5). Moreover, a significant difference was observed between studies conducted in different continents in terms of scores on item #2 of the scale (P < 0.0001), but there was no significant difference between different continents in terms of scores on other items. In addition, a significant difference was observed between different target populations in scores on items #6 and #7 (P < 0.05).

DISCUSSION
The present systematic review and meta-analysis aimed to estimate the pooled mean of fear of COVID-19. The results   showed that the participants in the reviewed studies had obtained 41% of the total score on the the FCV-19s. Fear of COVID-19 leads to stigmatization and social exclusion of patients and their families, and makes them vulnerable to adjustment problems, depression, irritability, anxiety, and anger (Abad et al., 2020;Satici et al., 2020b;Zhang et al., 2020). Therefore, it is important to pay attention to implications of COVID-19 for psychological health, because pandemics can lead to crisis in psychological, social, and economic domains (Xiang et al., 2020). Fear is not limited to the COVID-19 pandemic and has been observed in other outbreaks, including those of HIV and SARS (Ho et al., 2005). The pooled mean of fear of COVID-19 was higher in women than in men. This finding can be attributed to the fact that women are more delicate and vulnerable than men. In addition, Bakioglu et al. (2020) found that it was more acceptable for women to express their fears of illness. On the other hand, it is more acceptable for men to be strong and brave. In addition, because men are less likely than women to become ill, they tend to be less afraid of COVID-19. The results of the present study showed that the highest and lowest fears of COVID-19 scores were in studies conducted in Asia and Australia, respectively. This finding can be attributed to the fact that before spreading to other countries, COVID-19 was reported in China as an Asian country; therefore, people in China and other Asian countries experienced higher levels of fear of the new virus. Overall, different rates of fear of COVID-19 in different countries can be attributed to contextual and cultural factors and different levels of access to medical services. Isolation as a result of the COVID-19 pandemic led to increased rates of mental problems, such as anxiety, anger, PTSD, confusion, and even suicide (Giordani et al., 2020;Haktanir et al., 2020;Mamun and Griffiths, 2020a). For example, a Bangladeshi man killed himself because he thought that he had the new virus, but the autopsy showed that he actually did not (Mamun and Griffiths, 2020b). Therefore, misconceptions about COVID-19 can lead to increased xenophobia and suicide ideation. The same pattern had been observed during the SARS outbreak in Asia (Hong Kong) (Cheung et al., 2008). The pooled mean of fear of COVID-19 was higher in medical staff than in other target groups. This group became involved in fighting the new virus when health systems were not adequately prepared to respond to the pandemic (1). Longterm exposure to confirmed and also unrecognized COVID-19 patients, insufficient training on prevention and control of infectious diseases, and shortage of protective equipment were factors putting healthcare providers at higher risk of COVID-19 and, as a result, fear of the pandemic (Wang et al., 2020;Zhou et al., 2020). Nguyen et al. showed that the risk of testing positive for COVID-19 was three times higher in healthcare workers than in the general population (Nguyen et al., 2020b). Fear in healthcare providers is not limited to COVID-19 and has been reported during other outbreaks, including those of HIV (Montgomery and Lewis, 1995) and SARS (Ho et al., 2005). The highest and lowest pooled means were related to items #1 and #3 of the scale, respectively. This finding can be attributed to what the items assess. Item #1 directly assesses fear of COVID-19, while item #3 asks about symptoms of fear of COVID-19. One of the limitations of this study was the exclusion of preprint studies. Due to the large number of studies, it was not possible to review this type of articles in this meta-analysis and it is suggested that this type of articles be reviewed in future studies.

CONCLUSION
The excessive fear observed in previous outbreaks, including those of HIV and Ebola, has also been reported in the current COVID-19 pandemic and can be observed in the future. Excessive fear can negatively impact one's life in personal (e.g., mental health problems) and social (panic shopping and xenophobia) domains, while a normal (logical) level of fear can help one pay more attention to government measures aimed at reducing the spread of . The results of the present study showed that a moderate level of fear is required to reduce the risk of contracting COVID-19 and that fear of COVID-19 can be controlled and prevented from turning into excessive fear through providing effective training programs for different populations.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the Corresponding author.

AUTHOR CONTRIBUTIONS
FL and RGG: data collection and manuscript preparation. RGG: manuscript preparation and study conceptualization. QL: study design. SS and RGG: search strategy. QL and FL: final revision and grammar editing. SD: statistical analysis.