Maker Of COVID-19 Forecast Models Is Urging Caution

A microscopic image of the first U.S. case of COVID-19
A microscopic image of the first U.S. case of COVID-19. Courtesy of the Centers for Disease Control, Prevention
A microscopic image of the first U.S. case of COVID-19
A microscopic image of the first U.S. case of COVID-19. Courtesy of the Centers for Disease Control, Prevention

Maker Of COVID-19 Forecast Models Is Urging Caution

An Illinois State University professor who has been developing models of the coronavirus pandemic since January is urging caution.

“Every model we look at agrees on one thing … it is too soon to remove the lockdown measures,” said Olcay Akman, who also serves as editor in chief of “Letters in Biomathematics.” The peer-reviewed journal publishes mathematics and statistics research related to biological, ecological, medical, and environmental settings.

The rates may differ, he said. For example, his model forecasts 170,000 coronavirus-related deaths nationwide by August; the Centers for Disease Control forecasts 147,000. Both agree on the value of restrictive measures, said Akman.

“Illinois is one of the better states,” said Akman, noting that it took action, such as closing schools, more quickly than some states. “I think we’ll reap the benefit.”

Akman thinks the country needs “extensive testing, vaccinations and robust contact tracing policies before we start letting people mingle together again.”

He said the course of the pandemic has been following his model closely — “unfortunately.” The one surprise for him was the big jump in cases in mid-February.

“We knew it wasn’t going to go to zero, but we were surprised that is exploded so exponentially,” said Akman.

Several factors go into developing models, he explained, including how readily the disease spreads; probability of exposure, such as density of the population; and the probability of infection, which involves the activity required to be infected.

One reason why COVID-19 spreads so easily is that people can be contagious before the show symptoms, said Akman. “You may not even know you’re infected.”

Carla Pohl, a nurse practitioner and associate professor of nursing who teaches public health classes at ISU, said computer modeling would have been helpful when she was working as a midwife during the H1N1 pandemic a decade ago “to prepare for what we were seeing in the hospital.”

Having information on how many cases to expect and how quickly “makes all the difference in the world,” she said.

“You can try to implement procedures to cope with the onslaught of patients. … It’s a planning document,” said Pohl, speaking as someone who has studied and taught public health. 

To get such information out more quickly, “Letters in Biomathematics” has put in place a rapid review process.

“We didn’t think we could wait for a year to get information to help our decision makers with these policies,” said Akman.

The reviewers, who are epidemiologists and mathematicians, agree to review each manuscript in four to five days. They are published online as “provisionally approved,” while a more thorough peer review continues, Akman explained.

The manuscripts cover such things the impact of lifting lockdowns, mortality rates for different age groups, and frailty models. They eventually will be combined in a special issue on COVID-19 about a year from now, he said.

Although COVID-19 is a new virus, “it comes from a family,” said Pohl. “We know how the others in the family act. … It hangs around in the air, so it’s going to be contagious.”

That is among the reasons for recommendations such as limiting social contacts and wearing masks.

“You can only give people the information that they need to know and, hopefully, make the right choices, whether it’s exercise, nutrition or wearing a mask,” said Pohl.

The difference is that, while the choice not to exercise regularly or eat healthy diet affects only you, the choice not to wear a mask “impacts everybody else, it puts everybody else at risk,” said Pohl.

Originally published in The (Bloomington) Pantagraph.