This AI tool from ASU could predict when people should stay inside to avoid Valley fever

A new Valley fever risk prediction system is under development by Arizona State University.

The system works by using soil attributes, wind speeds and annual rain and temperature data to create an algorithm that predicts when chances of inhaling the disease-causing fungus is highest.

Fangwu Wei is an associate research scientist at ASU’s Decision Theater, a research facility. He said the model uses AI to measure peaks in potential exposure.

“We need to use these AI models algorithms to analyze, combine with these environmental factors to conduct some time series forecasting and see when we can notice there's a peak based on the results generated,” Wei said.

Theater director Manfred Laubichler says knowing those patterns is essential for public health in Arizona, where the fungus is most common.

Laubichler says the theater's mission is all about translating data and scientific knowledge into research to create solutions that allow health care professionals and others to make the right kind of decision.

Though unclear when it will be available, he says it will first be implemented in emergency rooms to get people diagnosed and treated faster.

“So different types of treatments, and the earlier you get the Valley fever treatment, the easier it is to treat it,” said Laubichler, “that's the context for what we are doing here.”

When asked if we will ever be able to fully contain Valley fever, he said no. He says ongoing research is trying to figure out how areas where Valley fever is prominent may expand due to climate change.

As global temperatures continue to increase, so will the outbreaks. The disease is caused by a fungus called Coccidiodies that lives in the soil.

Though it can’t be fully contained, it can be managed.

“So to know those patterns is therefore essential for public health in Arizona, the Southwest, and you know, an ever-expanding potential region where the fungus actually will take hold,” Laubichler said.

Knowing there is a high incident rate will allow people to adjust their outdoor activity accordingly to prevent inhaling the spores.

Laubichler says the system is good for prediction, but has to be refined to localize the incident rate. They are also working on mapping Valley fever to particular soil conditions and creating a risk map of where people are more likely to encounter Valley fever.

EDITOR'S NOTE: This story has been updated to correct the spelling of Fangwu Wei's name.