Google searches for specific, common gastrointestinal (GI) symptoms correlated with incidence of COVID-19 in the first weeks of the pandemic in 5 states with high disease burden, researchers report in the November issue of Clinical Gastroenterology and Hepatology. Specifically, searches for loss of appetite or taste correlated with increases in COVID-19 cases in high-incidence states.
Google Trends allows measurement of search term popularity in specific locations over time. It is not an epidemiologic tool for determining incidence, but it can be used to search volumes for specific diseases or disorders over time. Google search volumes correlate with infectious disease incidence and can be used in forecasting.
Imama Ahmad et al used Google Trends to investigate whether searches for GI symptoms correlated with COVID-19 incidence data, obtained from Harvard Dataverse. The authors analyzed incidence data from 15 states with top, median, and lowest COVID-19 burden for a 13-week period from January 20 through April 20, 2020. GI symptoms attributed to COVID-19 included ageusia (loss of taste), abdominal pain, loss of appetite, anorexia, diarrhea, and vomiting.
Ahmad et al found that Google searches for ageusia, loss of appetite, and diarrhea increased 4 weeks before the increase in COVID-19 cases in most states, with maximum correlation estimates of 0.998, 0.871, and 0.748, respectively.
Time-lag coefficients became stronger with increasing lag in weeks. A lag time of 4 weeks yielded the strongest correlation between symptom search volume and COVID-19 case volume, specifically for loss of taste (5 states), diarrhea (3 states), and loss of appetite (1 state). Plots of symptoms vs cases over time demonstrated an increase in search volume followed by an increase in COVID-19 incidence after 3–4 weeks (see figure).
These findings support the concept that GI symptoms are an important indicator of COVID-19 and that Google Trends can be used to predict pandemics with GI manifestations. Further studies are needed to determine peak incidence, seasonality, and the optimal query time frame for generating predictive models for COVID-19.