BACKGROUND
The sudden outbreak of COVID-19 has placed an unprecedented pressure on China's public health system. It is imperative to strengthen the capacity of early surveillance and early warning to build a sound public health system. Therefore, it is necessary to improve the multi-channel monitoring and early warning mechanism to improve the ability of real-time analysis and judgment.
OBJECTIVE
To explore the correlation of COVID-19 spread with Baidu search data in Beijing, so as to evaluate the possibility of monitoring the epidemic situation of COVID-19 with Baidu search data.
METHODS
This study compared the daily case counts of COVID-19 outbreak from January 20 to March 1, 2020 with Baidu search data for the same period in Beijing. After keyword selection, filtering and composition, the most correlated lag of the COVID-19 Baidu Search Index (CBSI) was used for comparison and linear regression model development.
RESULTS
Our findings showed a positive relationship of CBSI and the confirmed cases of COVID-19 (ρ=0.711, P < .001). The strongest correlation between COVID-19 confirmed cases and indices, CBSI, was at a lag of -11 days. The regression coefficient β1 of the established regression model was equal to 1.042 (P<.001), R2 was equal to 0.7, which indicated that Baidu search data could reflect 70% of the variation in COVID-19 cases.
CONCLUSIONS
COVID-19 Baidu Search index may be a good monitoring indicator for early detection of COVID-19 outbreaks.