Governor and health commissioner say they look at data models to guide their decisions, but 'don't have something to share'

Graphs from the rt.live website show the estimated transmission rate of the coronavirus in Kentucky and some other states. A rate of 1 estimates that one infected person will infect one other person.

By Dalton Stokes and Al Cross
The (Frankfort) State Journal and Kentucky Health News

Gov. Andy Beshear said this Wednesday about his slow, gradual steps to reopen the state: “If it proves that we can't do any of them safely, it is always subject to pause.”

How will he know when to pull the plug?

In some other states, governors give considerable attention to two statistical models used to track the spread of the coronavirus, but Beshear and Health Commissioner Steven Stack are keeping to themselves whatever models they are using.

Epidemiologists measure reproduction of a virus with a figure called R0, or “R naught.” It represents the number of infections, on average, that stem from a single infection.

A related number, called Rt, for transmission, uses statistical adjustments to create a more current estimate of how fast a disease is spreading, or what epidemiologists call the effective reproduction number of a virus.

Asked at Beshear’s briefing Wednesday if the state Department of Public Health calculates those figures, and if so, what they are, Stack said: “Yes, we do try to calculate those numbers, but I don’t have any great confidence in them,” because they depend on so many varying factors.

“We do look at those things,” Stack said, “but I don’t have something to share.”

A website created by the co-founders of Instagram estimates the virus’ Rt value in each state daily. Natalie E. Dean, an assistant professor of biostatistics at the University of Florida, told Vox that the model, as well as the data behind it, seem sound.

The site estimates that Kentucky’s Rt on April 25 was 0.88, just about in the middle among the states.

When Beshear issued his “healthy at home” order March 26, Kentucky’s Rt was 1.3, the website estimates. It dropped to 1 on April 10, and settled at 0.88 on April 21. Preliminary data indicate it has remained there since, the site says.

Among the states bordering Kentucky, most have lower Rt estimates; Illinois has about the same, and only Indiana has seen a resurgence. Its estimate fell to 0.94 in early April, then rose to 1.04 on April 15. Its latest estimates have been just below 1.

“You want it to be less than 1, which means it’s decreasing; if it’s over 1, it’s increasing, and if it gets too high, it increases very rapidly,” Stack said.

New York Gov. Andrew Cuomo said Tuesday that his plan uses Rt as a “circuit breaker,” set to 1.1. If the state jumps to 1.1, or higher, his reopening efforts will be reversed.

Dean told Vox that the accuracy of the R0 and Rt estimates hinges on the amount of testing in a state. If states don’t increase testing proportionate to the growth of the virus, they will report the same relative number of new cases each day, even if there are actually more.

While Stack wouldn’t say what numbers he and Beshear are using, he said the difference in the Rt estimate and the basic R0 number for the coronavirus shows the effectiveness of the measures they have taken.

“The R0 for this disease is ... closer to 3 when it’s unchecked,” he said. The number means that if no measures had been taken to contain the virus, each infected person would have infected three other people, and those three would have infected nine more, and so on.

“We are trying to do the best we can,” Stack said, “to balance the societal need to resume activities that are so important to society, but to keep people safe.”

Information for this story was also gathered by Melissa Patrick of Kentucky Health News.