This is the second of a two-part editorial. The first one ran in the April 24 edition of The Franklin News-Post.
Our inquiry into the data began with a single question: On April 16, the Highland Medical Center in Highland County announced it had diagnosed a coronavirus patient. As of April 21, though, the map on the VDH website still showed no coronavirus cases in Highland County. Why not?
Meanwhile, former New Castle Town Council member Jordan Labiosa tweeted that it had taken eight days for cases diagnosed in Craig County to show up on the website. Is the data really running that far behind? No, but those two situations highlight the dangers of looking at this data on the granular level. The numbers listed on the map on the VDH website shouldn’t be taken as Gospel. Virus reports — from labs or, in some cases, from doctors who have diagnosed a patient without a lab test — include zip codes. But some zip codes cross city and county lines, which means cases can get wrongly attributed on the map. That’s the case with the Highland patient, who was initially listed under Augusta County, said Dr. Molly O’Dell, director of Communicable Disease Control for the Alleghany and Roanoke health districts. Many Craig County residents share zip codes with Roanoke County and Montgomery County. She pointed out the map showed one case in Covington. There are no cases in Covington, she says. Instead, that’s an Alleghany County case with a Covington zip code — the map hadn’t been updated yet to reflect that.
If you’re obsessed with this map every day to decide how risky things are, that’s a problem. However, if you’re an understaffed health department fighting a pandemic, fixing such “noisy data” is hardly the top priority. Whether that case is shown in Covington or Alleghany County doesn’t really change the public’s understanding of how many cases are in the Alleghany Highlands.
O’Dell said the numbers on the map often run behind the cases her office knows about. In about half the localities she’s responsible for, they were off last week by one or two or sometimes three cases. The website will catch up eventually but people shouldn’t expect it to be updated in real-time.
The 14 days of declining cases standard isn’t how epidemiology normally works, she says. In theory, 14 days in which each day is one case lower than the one before constitutes a decline — but, statistically speaking, that’s more of a plateau than an indication the disease is dying out.
“Traditionally for an epidemic the epidemic is gone when you go two incubation periods with no new cases,” O’Dell says. Since the incubation period of COVID-19 is about 14 days, that means we’d need 28 days with no new cases reported. Even that wouldn’t guarantee the virus is gone, because people could still be carrying it and not showing symptoms. That 28-days-of-no-new-cases is a medical standard politicians may not have the spine for. Regardless, basing a political argument on the number of new cases, is not the best measure for either side.