![]() The 3-year (October 1, 2017–October 25, 2020) time series GT trend data of ‘all categories’ for keywords of symptoms that may be related to COVID-19 was queried using R package gtrendsR. COVID-19 data and Google Trends (GT) data were separately analyzed in nine different regions: Japan (JP) and eight English-speaking countries, namely, Australia (AU), Canada (CA), Great Britain (GB), Ireland (IE), India (IN), Singapore (SG), United States (US), and South Africa (ZA). A statistical level of less than 0.05 is considered significant if not stated otherwise. ![]() Third, because COVID-19 and its symptoms have attracted intensive attention worldwide, the influence of media coverage on GT symptom keywords is inevitable, which has hardly been adjusted in a statistically favorable manner.īased on the above analytical concerns for earlier studies, by using the vector autoregression (VAR) model, which is designed to deal with time-series data and is robust against weakness as observed in case of using correlation, we aim to identify statistically more reliable symptom keywords for which GT trends may be used as a predictive measure for future COVID-19 positivity trends, and to validate the earlier study results.Įxtracting Google Trends and COVID-19 dataĪll the following data handling and analyses were performed using R 3.5.2 (R Foundation for Statistical Computing, Vienna, Austria). ) without adequate adjustment for multiple comparisons, which would also increase the risk of false-positive results. Second, the Pearson/Spearman correlation tests were repeated for each of the included symptom keywords (e.g., fever, cough, pneumonia, anosmia, sore throat, headache, etc. This is sometimes critically inappropriate in the context of time-series analyses because time-series data often contains unit-root and the correlation between such series often results in high coefficient value and t-statistics, and thus it can increase the likelihood of obtaining spurious correlations. First, Pearson (or Spearman’s rank) correlation is often applied to assess the correlation between the time-series trends of COVID-19 cases/deaths and GT trends in symptom keywords without confirming the stationarity of these time series. In many earlier studies analyzing GT trend data as an epidemiological tool, with a few exceptions, analytical fallacies were of concern. within the initial months following the outbreak. As for coronavirus disease (COVID-19) that became a worldwide pandemic in early 2020, the potential use of GT to predict COVID-19 cases or deaths has been reported with regard to GT trends and keyword searches of “COVID-19” or any of its symptoms, including chest pain, anosmia, dysgeusia, headache, shortness of breath, etc. ![]() It is used as one of the “infodemiology” tools to study epidemiological trends of certain disease outbreaks such as the Middle East Respiratory Syndrome epidemic and the Ebola outbreak. Google Trends (GT) is a publicly available source of online Google search trafficking data ( ), which allows users to visualize changes in time series related to the general public’s online interest in certain keywords. Our results suggest that some of the search keywords reported as candidate predictive measures in earlier GT-based COVID-19 studies may potentially be unreliable therefore, caution is necessary when interpreting published GT-based study results. “Sense of smell” and “loss of smell” were the most reliable GT keywords across all the evaluated countries however, when adjusted with their media coverage, these keyword trends did not Granger-cause the COVID-19 positivity trends (in Japan). Our Granger causality-based approach largely decreased (by up to approximately one-third) the number of keywords identified as having a significant temporal relationship with the COVID-19 trend when compared to those identified by Pearson or Spearman’s rank correlation-based approach. In addition, the impact of media coverage on keywords with significant Granger-causality was further evaluated using Japanese regional data. We extracted the relative GT search volume for keywords associated with COVID-19 symptoms, and evaluated their Granger-causality to weekly COVID-19 positivity in eight English-speaking countries and Japan. In this study, we aimed to apply statistically more favorable methods to validate the earlier GT-based COVID-19 study results. However, many of the earlier GT-based studies include potential statistical fallacies by measuring the correlation between non-stationary time sequences without adjusting for multiple comparisons or the confounding of media coverage, leading to concerns about the increased risk of obtaining false-positive results. Google Trends (GT) is being used as an epidemiological tool to study coronavirus disease (COVID-19) by identifying keywords in search trends that are predictive for the COVID-19 epidemiological burden.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |