Dissemination and Refutation of Rumors During the COVID-19 Outbreak in China: Infodemiology Study

During the outbreak of COVID-19, numerous rumors emerged on the internet in China and caused confusion among the public. However, the characteristics of these rumors in different phases of the epidemic have not been studied in depth, and the official responses to the rumors have not been systematically evaluated.

The aims of this study were to evaluate the rumor epidemic and official responses during the COVID-19 outbreak in China and to provide a scientific basis for effective information communication in future public health crises.

Data on internet rumors related to COVID-19 were collected via the Sina Weibo Official Account to Refute Rumors between January 20 and April 8, 2020, extracted, and analyzed. The data were divided into five periods according to the key events and disease epidemic. Different classifications of rumors were described and compared over the five periods. The trends of the epidemic and the focus of the public at different stages were plotted, and correlation analysis between the number of rumors and the number of COVID-19 cases was performed. The geographic distributions of the sources and refuters of the rumors were graphed, and analyses of the most frequently appearing words in the rumors were applied to reveal hotspots of the rumors.

A total of 1943 rumors were retrieved. The median of the response interval between publication and debunking of the rumors was 1 day (IQR 1-2). Rumors in text format accounted for the majority of the 1943 rumors (n=1241, 63.9%); chat tools, particularly WeChat (n=1386, 71.3%), were the most common platform for initial publishing of the rumors (n=1412, 72.7%). In addition to text rumors, Weibo and web pages were more likely to be platforms for rumors released in multimedia formats or in a combination of formats, respectively. Local agencies played a large role in dispelling rumors among social media platforms (1537/1943, 79.1%). There were significant differences in the formats and origins of rumors over the five periods (P<.001). Hubei Province accounted for most of the country's confirmed rumors. Beijing and Wuhan City were the main centers for debunking of disinformation. The words most frequently included in the core messages of the rumors varied by period, indicating shifting in the public's concern.

Chat tools, particularly WeChat, became the major sources of rumors during the COVID-19 outbreak in China, indicating a requirement to establish rumor monitoring and refuting mechanisms on these platforms. Moreover, targeted policy adjustments and timely release of official information are needed in different phases of the outbreak.