S0143 Association of Search Engine Queries for COVID-19 and Diarrhea With COVID-19 Epidemiology

Abstract

INTRODUCTION: Coronavirus disease (COVID-19) is a far-reaching pandemic that has changed the landscape of human interaction. Gastrointestinal symptoms, such as diarrhea, are part of the spectrum of disease. Previous infectious disease studies have shown good correlation between online search engine queries and disease burden, thereby allowing tracking. We aimed to assess the relationship between Google searches for COVID-19 and diarrhea with COVID-19 epidemiology, including incidence and mortality.

METHODS: Google Trends, a publicly available and free service that tracks online search frequency, was utilized to identify online searches for combined diarrhea plus COVID-19 from March 3, 2020 to May 4, 2020. These results were stratified by state and then compared with publicly reported incidence data from the US Centers for Disease Control and Prevention (CDC) for the same timeframe. A control search of other COVID-19 associated symptoms listed by the CDC was performed. Additional control searches of the individual search terms diarrhea and COVID-19 were conducted. Correlations between geographic location and epidemiologic trends were analyzed using the Pearson correlation coefficient.

RESULTS: State-by-state searches for COVID-19 plus diarrhea were correlated with mortality reported as deaths per capita (R = 0.31; P = 0.03) [Figure 1], but not incidence reported as cases per capita (R = 0.19; P = 0.19). New Hampshire had the highest number of COVID-19 plus diarrhea searches while Alaska, Hawaii, Montana, and West Virginia had the lowest relative search frequency. Diarrhea was the second most searched symptom in conjunction with COVID-19 [Figure 2]. Fever and cough, which are thought of as more traditional upper respiratory viral symptoms, were also commonly searched. Diarrhea alone was more frequently searched than COVID-19 alone or COVID-19 plus diarrhea [Figure 3].

CONCLUSION: Our data show weak correlation between mortality and COVID-19 searches. This indicates that, for COVID-19, online search analysis is unlikely to be a good substitute for more traditional methods of patient testing, case tracking, and early detection. However, Google Trends of searches for COVID-19 plus diarrhea, in addition to being a marker for disease interest, may still be useful as part of a more complex model for tracking disease as they parallel search activity for COVID-19 in general.