Medical Moment: Using AI to predict the spread of COVID-19

Published: Sep. 21, 2022 at 5:38 PM EDT
Email This Link
Share on Pinterest
Share on LinkedIn

(WNDU) - The newest subvariant strain of COVID BA.5 is spreading through the U.S. and is now the cause of most COVID-19 cases.

But what if you could predict the spread of the virus down to your hometown?

New research may make localized early COVID warnings a reality. For the past two years, scientists have been using data and computers to predict where the virus will spread next, and how many people may get sick. Scientists are testing a method of forecasting COVID that could help cities prepare information that is localized, much like a daily weather forecast.

“If you are looking at the weather, you would not like to know the weather for the United States,” said Ben D. Sawyer, PhD. “You’d like to know the weather for your surrounding area.”

The researchers used artificial intelligence to forecast the spread of the disease. Compared with current methods of forecasting COVID, the AI model predicted COVID cases that were closest to the actual numbers.

Scientists say with the machine learning model, local experts anywhere in the world would be able to more accurately predict the number of people who would get sick, the number of hospitalizations, and the number of deaths.

“One of the most useful things we can do is start working on giving people useful tools, to understand how the disease will impact their life,” Dr. Sawyer explained.

And prepare to fight the virus for the long term.

As of mid-July, the CDC reported the country’s seven-day average on new cases had ballooned to over 100,000 new infections a day, more than three times higher than this time last year.