Geographical Social Inequalities in Information Seeking Response to the COVID-19 Pandemic in China: A Longitudinal Analysis of the Baidu Index


The median of the national daily search index for terms related to Covid-19 was 4,533 (IQR (interquartile range) = 1,301) before the outbreak of COVID-19 (from January 1, 2017 to December 30 2019), and 314.718 (IQR = 445,074) after the outbreak (December 31, 2019 to March 15, 2021). The median provincial-level search index ranged from 63 (IQR=7) in Tibet to 1138 (IQR=302) in Guangdong before COVOD-19, and ranged from 1386 (IQR=983) in Tibet to 38,061 (IQR=45,784) in Guangdong after the COVID-19 outbreak. The raw relative change in the median of the search index ranged from 2099% in Tibet and 2034% in Hainan to 3872% in Beijing and 4284% in Liaoning (Table 1). 89,936 cases of SARS-COV-2 occurred nationwide (ranging from 1 case in Tibet to 68,021 cases in Hubei) from December 31, 2020 to March 15, 2021. The number of confirmed cases outside Tibet and Hubei ranged from 18 (0.1%) in Qinghai to 2,245 (10.6%) in Guangdong province. Along with these research patterns, 13%, 76% and 11% of confirmed Covid-19 cases were reported in January 2020, February 2020 and March 2020 to March 2021 respectively.

Table 1 Comparison of search index in the COVID-19 and pre-COVID period.

Model of estimated change in research index by Human Development Index (HDI) categories


As shown in Table 2, there were 10% (risk ratio (RR)=1.10, 95% CI 1.07–1.13, p

Table 2 Model of estimated change in search index by HDI categories.
Figure 1

Baidu search index by province and number of new confirmed cases over time. (A) Observed daily search index (log transformed) by province and HDI category over time. The aggregated search index by HDI category over time is shown in Fig. S1. (B) New COVID-19 confirmed daily in China (cases in Hubei provinces are excluded).

Initial wave of COVID-19

During the initial wave, the search index increased by 41%, 62% and 58% as of December 31, 2019 among low regions (RR = 1.41, 95% CI 1.34-1.49 , p

Similarly, the search index was multiplied by 107, 125 and 125 between January 18 and 25, 2020, the period shortly after the official announcement of human-to-human transmission (THH), among regions with low (RR = 106.8, 95% CI 100.1–114.0, p

Beijing outbreak

The Beijing outbreak was associated with a 91%, 34%, and 112% increase in seeking index among low regions (RR=1.91, 95% CI 1.79-2.03, p

Qingdao outbreak

The Qingdao outbreak was associated with a comparable 31%, 34%, and 41% immediate increase in seeking index among low regions (RR = 1.31, 95% CI 1.23-1, 40, p

Shijiazhuang outbreak

The Shijiazhuang epidemic in January 2021 was associated with an immediate increase of 100%, 167% and 145% in the search index among low regions (RR = 2.00, 95% CI 1.85– 2.16, p

Figure 2
Figure 2

Immediate relative change in search index at different exposure periods (A) December 31, 2019, estimated start of the first Covid-19 wave. (B) from January 18, 2020 (official announcement of human-to-human transmission) to January 25, 2020 (shortly after the lockdown and the estimated peak of the daily search index in the initial wave of Covid-19). (VS) Outbreak in Beijing from June 11, 2020. (D) Outbreak in Shijiazhuang from January 3, 2021. Specific point estimate of relative change and corresponding 95% CIs are provided in supplementary material.

Association between HDI, GNP per capita, education, life expectancy and magnitude of change in search index

The results of the models where the HDI or its component was coded as a continuous variable were consistent with the results of our main analysis. As shown in Table S1, pre-pandemic trends in two provinces differing by HDI, GNPPP (gross national product per person), year of education or life expectancy by one standard deviation were similar (p > 0.1). The immediate relative increase in the seek index in a province with a higher standard HDI was statistically higher (initial wave: ratio of RR = 1.09, p


About Author

Comments are closed.