Internet Search Data Can Help Identify a ‘Twindemic’

Internet Search Data Can Help Identify a 'Twindemic'

In current times, the Google search box might be the most popular place to look for medical advice.

People search for things like “loss of taste” and “how long contagious” on the site often enough for researchers at Georgia Tech to be able to use the information to correctly predict waves of flu-like illnesses and COVID-19 infections.

Their predicting models work for the whole country and for each state. This gives healthcare systems a new way to learn about possible “twindemics” that could cause problems.

The model was made by Shihao Yang and his team at the H. Milton Stewart School of Industrial and Systems Engineering. It was released in the Nature Journal of Communications Medicine.

“We find that there might be a way to predict the short term using search data,” said Yang, an assistant professor who co-wrote the paper with Simin Ma, a Georgia Tech Ph.D. student, and Shaoyang Ning, an assistant professor of statistics at Williams College.

Internet Search Data Can Help Predict a Looming 'Twindemic'

The team’s “hack” uses 23 key search queries, like “loss of taste,” to see how bad flu-like illnesses and COVID-19 cases will be in four weeks. The model uses government data on hospitalizations, deaths, and outpatient visits for flu-like illnesses from COVID-19.

This is the official benchmark because it is hard to tell the difference between flu cases and other viral infections with similar symptoms.

“We provide a unique view. “This data, which you might think of as alternative data, isn’t really used by a lot of other people,” Yang said. “People who work in health care and epidemiologists use information from monitoring, surveys, and hospital records.

That’s the normal range. As a mathematician who works in engineering, my view of our healthcare system is a little different.

Yang said that standard epidemiological models are much better than his way of making predictions about the next six months or more. His team’s work is best in the short term, when it helps public health officials and hospitals prepare for a sudden increase.

The researchers gave the Centers for Disease Control and Prevention the results of their study.

Yang has been using search data to predict flu outbreaks for years. In early 2020, he started to use what he knew about the flu to predict the coronavirus pandemic. Over time, he got the idea that the new tools he and others had made for COVID forecasts could help with flu modeling and vice versa.

As worries about a “twin demic” of both diseases grew last fall, Yang said, “We realized this seems to be the time to start treating flu and COVID kind of the same.” “So, the whole idea is, why don’t I just put the two diseases together and make a model for both?”

Read More:

How to Solve “Chat GPT is at Capacity Right Now”?

What Are the Best Alternatives to CloudConvert?

Is Wordtune the Best AI writing Assistant?

Leave a Comment