COLUMBIA, S.C. (WIS) - Dr. Alex McLain says modeling a virus' spread is like tracking a hurricane. But, the COVID-19 predictions were so off, he says it's like forecasters thought a storm would hit North Carolina but hit Maine instead.
In the middle of May, the IHME model from the University of Washington, which is the model used on DHEC’s website, predicted new daily deaths from COVID-19 would stop around May 29th.
Instead, there have been about 300 deaths since that date.
“The number of new deaths is outside of their uncertainty interval 70 percent of the time. It’s supposed to be inside those uncertainty intervals 95 percent of the time. That’s terrible...they are doing a terrible job.,” University of South Carolina epidemiology biostatistician Alex McLain said.
McClain says one of the main issues is models assumed people wouldn't get restless.
“All the models took into account that people would follow these CDC recommendations indefinitely, and as we know that didn’t happen. People started to go back to work [and] started to go back to restaurants. So that’s what went wrong in the models,” he said.
In addition, McLain says the models weren’t taking into account a spike over Memorial Day weekend, the protests that erupted in the wake of George Floyd’s death, or states reopening before reaching the White House Task Force’s prerequisite milestones. But, he says the spike in our state and others has taught modelers more about the coronavirus.
“I think we have a much better idea of what is the impact of a stay-at-home order...the impact of face mask rules on the number of daily cases we are going to see,” he said.
McLain says while it's not a perfect study, researchers are able to learn more about modeling the virus by examining the new daily case trends from different regions of the county.
“In the U.S. we have a lot of different task mask policies. They are different in New York, Georgia, and South Carolina as we know, as we see on the news every day...and we’ve been able to use the number of new cases that we’ve seen in those different areas to get some idea of how much of an impact those types of rules have on the number of new cases,” McLain explained.
However, just because modelers are seeing some things that the curve to go up and down, they're still not able to predict the future of this virus' spread.
“If you tell me how much people are going to social distance and if you tell me what the policies are going to be all summer. Yes, you can give some predictions that will be somewhat accurate. But we don’t know those things,” he said.