The Effect of the COVID-19 Pandemic on Health Care Workers’ Anxiety Levels: A Meta-Analysis

Background. The COVID-19 pandemic has been declared a public health emergency of international concern; this has caused excessive anxiety among health care workers. In addition, publication bias and low-quality publications have become widespread, which can result in the dissemination of unreliable findings.This paper performed a meta-analysis with the following two aims: (1) to examine the prevalence of anxiety among health care workers and determine whether it has increased due to the COVID-19 pandemic, and (2) to investigate whether there has been an increase in publication bias.Methods. All related studies published/released from 2015 to 2020 were searched in electronic databases (Web of Science, PubMed, PsyArXiv, and medRxiv). The heterogeneity of the studies was assessed using the I2 statistic. The effect size (prevalence rates of anxiety) and a 95% CI for each paper were also calculated. We used a moderator analysis to test for the effect of COVID-19 on health care workers’ anxiety levels and detect publication bias in COVID-19 studies. We also assessed publication bias using the funnel plot and Egger regression.Results.The pooled meta-analysis prevalence was 35.3% (95% CI: 32.2%−38.4%). The moderator analysis revealed no significant difference between articles related to COVID-19 and those unrelated to COVID-19 (p = 0.831). Moreover, no significant difference was found between articles related to COVID-19 and preprints (p = 0.981). Significant heterogeneity was found in each subgroup. Egger’s tests revealed publication bias in articles related to COVID-19 and preprints (p < 0.001).Conclusions. Concluding whether the anxiety state of health care workers is altered by the COVID-19 pandemic currently is difficult. However, we have strong evidence that their anxiety levels are always high, although this could also be a false positive caused by a large publication bias. Moreover, without comparing the results with those of non-health care workers, we cannot be sure that their anxiety is particularly high. Furthermore, we found a large publication bias in studies; however, the quality of the studies is relatively stable and reliable.