| Abstract [eng] |
Background and Objectives: The probability of future COVID-19 waves remains high, thus COVID-19 surveillance and forecasting remains important. Online search engines harvest vast amounts of data from general population in real time and make this data publicly accessible via such tools as Google Trends (GT). Therefore, the aim of this study was to systematically review literature about possible use of GT for COVID-19 surveillance and prediction of its outbreaks. The objectives were to determine the main findings about GT, also most frequently used keywords, languages, statistical methods and time periods. Materials and Methods: In November 2021, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, we collected articles about possible use of GT for COVID-19 surveillance. We resulted in 44 publications that were used in this review. Results: The majority of the studies (84.1%) included in this review showed positive results of the possible use of GT for forecasting of COVID-19 outbreaks. Most of the studies were performed in English-speaking countries (68.2%). The most frequently used keyword was “coronavirus” (47.7%), followed by “COVID-19” (27.3%) and “COVID” (20.4%). Many authors have made analyses in multiple countries (50%) and got the same results for the majority of them, thus showing the robustness of the chosen methods. Various methods including long short-term memory (4.5%), random forest regression (4.5%), Adaboost algorithm (2.3%), autoregressive integrated moving average, neural network autoregression (2.3%), and vector error correction modelling (2.3%) were used for the analysis. It was seen that most of the publications with positive results were using data from first wave. Later the search volumes reduced even though the incidence peaked. Conclusions: In most countries, the use of GT data would be beneficial for forecasting and surveillance of COVID-19 spread. However, effectiveness of the use of it might be influenced by many factors, thus longer studies are needed in order to find the most effective solutions for prediction. |
