Assessment of the relationship between Google Trend search data on clinical symptoms and cases reported during the first wave of the COVID-19 outbreak in India.

Infodemiology and infoveillance approaches have been extensively used in recent years to support traditional epidemiology and disease surveillance. Hence, the present study aimed to explore the association between Google Trends (GTs) search of clinical symptoms and cases reported during the first wave of COVID-19. The GT data from January 30, 2020, to September 30, 2020, and daily COVID-19 cases in India and a few selected states were collected from the Ministry of Health and Family Welfare, Government of India. Correlation analysis was performed between the GT index values and the number of confirmed cases. Followed by, the COVID-19 cases were predicted using Bayesian regression and classical linear regression models. A strong association was observed between the search index of clinical symptoms and reported COVID-19 cases (cold: R=0.41, headache: R=0.46, fever: R=0.66, loss of taste: R=0.78, loss of smell R=0.86) across India. Similarly, lagged correlations were also observed (loss of smell, loss of taste, loss of taste and loss of smell, fever and headache show 3, 9, 1, 9, and 13 days lag periods respectively). Besides this, the Bayesian regression model was outperformed (MAE: 0.331164, RMSE: 0.411087) for predicting the COVID-19 cases in India and regionally than the frequentist linear regression (MAE: 0.33134, RMSE: 0.411316). The study helps health authorities better prepare and planning of health care facility timely to avoid adverse impacts.