55.7 F

Davis, California

Friday, April 12, 2024

Social media can predict case count of COVID-19 ahead of official reports

Online posts give researchers insight on people with COVID-19 symptoms  

Defined as social media posts clearly mentioning symptoms or diagnoses of COVID-19, “sick posts” on the Chinese social media site Weibo are able to predict the daily number of COVID-19 cases up to a week ahead of official statistics, according to a study conducted by UC Davis faculty in the department of communication. 

Wang Liao, an assistant communications professor, explained that this study utilizes people’s tendencies to seek help in uncertain situations in order to make sense of the public’s use of social media in pandemic conditions. 

“We need to communicate,” Liao said. “We shout out and scream when there is uncertainty or risk in the environment. That kind of instinct or human nature is probably the very reason behind this kind of study or why it would work.”

Since this is a novel virus, there is no current system in place efficient enough to collect true signals from the population regarding who is infected with the virus, according to Jingwen Zhang, an assistant professor in the Department of Communication. The only way to determine who is infected is if the patient decides to report symptoms to a local clinic or hospital — which already causes a delay, as patients have likely been experiencing symptoms prior to their visits. 

Early on, China did not allow doctors to inform patients infected with COVID-19 about their conditions, which led to widespread uncertainty about why people were getting sick, Zhang explained. This controversial move spurred much activity on Chinese social media, with users expressing concerns and seeking help. 

Although this is not the first study to use social media for disease surveillance and prediction, what differentiates this research is isolating true signals of disease by sifting through noise on social media, Liao said. Rather than relying on keyword searches, which would include posts about COVID-19 in general, the team focused on posts that included specific diagnoses and symptoms which then served as better indicators for predicting daily cases.

According to both Zhang and Liao, because of the noise prevalent on social media and the multitude of people contributing to the platform, people should be mindful of the information they consume. Zhang recommends relying on the accounts of familiar organizations for accurate information.

“Social media is where people find information and misinformation, where people seek and receive support, where people connect with their friends and family, where government and health authorities can effectively communicate with their constituents about preventive measures and public policy, and where scientists and public health agencies can find data to inform their response efforts,” said Cuihan Shen, an assistant professor in the Department of Communication, via email.

Although its study focused on a social media platform in China, the team believes that this epidemic surveillance modeling would prove to be similar with American social media platforms, such as Twitter, but with a few differences. Zhang expects there to be some cultural variations, as Eastern Asian cultures may be more reluctant to talk about diseases and death on a public platform compared to Western cultures. 

Liao also said the differences in the trajectory of the pandemic in China and the U.S. may contribute to variations in the predictive power of social media. The team will be exploring how topics discussed online evolve with the timeline of the pandemic and different forms of misinformation in the next steps of its research.

 “Fundamentally, it’s about human psychology, disclosing your problems to others,” Liao said. “We’re social animals and we look for help when we are in trouble and we communicate those helping messages to others. Because of that kind of fundamental human motivation, I would say that the pattern we found in Chinese social media would apply and also be found in social media in the United States.”

Written by: Michelle Wong — science@theaggie.org


Please enter your comment!
Please enter your name here