If the state used the COVID-19 numbers Cortland County provides each day, Cortland County would be in a “red zone,” forcing businesses to close and banning mass gatherings, according to metrics the state released this week.
Cortland County and its 48,000 residents would be back where it was six months ago, with hundreds or even thousands more people out of work, a massive hit to sales tax revenue and virtually everyone stuck at home, staring at the walls.
So why isn’t it?
The state’s metrics for placing communities in yellow, orange or red zones concentrate on the number of new cases each day for the past week, and that number in relation to the number of people actually tested each day.
For counties with fewer than 50,000 residents, like Cortland, the state requires the most severe restrictions if:
• The county reports an average of 15 or more cases per 100,000 people over seven days — Cortland reports nearly three times that.
• And if the number of positive COVID-19 tests exceeds 6% of the total tests in a seven-day rolling average for 10 days. Cortland’s positivity for the past 10 days has ranged between 6% a day to more than 12%.
However, the state Department of Health records different numbers than the county. While the county has reported 382 cases of coronavirus between Oct. 5 and Tuesday, the state reports only 228. And while the county says it has tested 5,495 people, the state says it has tested 10,087.
The difference between those numbers is the difference between positive rates of 0.5% to 7% from the state, but 5% to 12% from the county.
And that’s the difference between a seven-day rolling average of 1.8% to 2.7% according to state numbers, but almost never drops below 6% for county numbers.
The spread might also be attributable to a cluster in a single institution. Gov. Andrew Cuomo in a news briefing Tuesday had mentioned that Cortland County was in a micro-cluster related to SUNY Cortland, but state officials have not said whether the college was sufficient cause to exempt it from the restrictions.
County officials, and public statistics, show the virus has spread offthe campus and into other parts of the community, either way.
Ask the state
Lisa Perfetti, the county interim public health director, averred when asked at a news conference about the discrepancy.
“That’s a good question for the state. We communicate twice a day with our state representative and even twice a day on the weekends, so they’re aware of our numbers,” Perfetti said.
A state Health Department spokeswoman, like Perfetti, said the state draws its data from an electronic reporting system.
“With the advent of rapid testing, those are not necessarily reported through the state system but we get them directly much quicker,” Perfetti said. “Pool testing results from the college, rapid testing sites those come directly to us.”
However, the state reports much more testing, nearly twice as many as the county reports.
County Legislator Beau Harbin, (D-Cortland) has been recording the county’s numbers since March, and sits on a multi-agency panel that includes county, city and SUNY Cortland and Tompkins Cortland Community College officials trying to share information and coordinate response.
He’s been asking the same questions. “I’ve asked about this, but I haven’t gotten a veryspecific answer,” Harbin said Wednesday.
He theorizes the difference is in what tests are counted — and there are different types of COVID tests. The standard test you may get at a testing site uses the real-time polymerase chain reaction method. It tests for the virus’ genetic material, and is nearly 100% accurate.
Pool tests, however, combine a number of samples and tests them together. If the pool tests negative, one presumes all its samples were negative. If it tests positive, further testing isolates who was actually sick. It’s much easier to test large populations, but depending on the method, it’s not as accurate. Sometimes, the viral material can be diluted beyond detection in a large sample.
The pool method is frequently used at colleges — that’s one of the reasons Tompkins Countyhas been able to report as many as 10,000 tests a day. And it’s used at SUNY Cortland.
“I believe their use of pool test is masking the actual growth of COVID in our community,” Harbin said.
While that’s possible, that’s not a given, said Christopher P. Morley, chairman of the department of public health and preventive medicine at SUNY Upstate Medical University in Syracuse. A pool test developed at the medical school yields accuracy on par with the PCR test. Besides, he added, the state wasn’t counting the pool tests in its numbers, only the follow-up tests to confirm who was sick.
State testing sites
The state has set up three testing facilities operating through the weekend. Expect their results to reduce the positivity rate.
With the existing facilities, patients are referred either because they meet diagnosis criteria, are about to undergo a surgical procedure, or work in specific professions.
The temporary facilities will test anybody, so people who are simply concerned they may have been exposed can go, even if they don’t meet other criteria. With just a few hours notice Thursday, a couple hundred had signed up for tests on Friday, and the tests continue through the weekend. More
testing — and more negative results — will move the positivity rate down.
But that also changes the consistency of the numbers, Morely said. “Once you alter the methodology, the numbers mean something different.”
The numbers reveal two discrepancies: the number of tests;and the number of tests that are positive.
On any given day, the county and state numbers may vary in the number of positives, Perfetti said, and it should even out over time.
But it’s not, the gap is increasing. In August, the numbers were virtually identical on a daily basis. On Friday, the state reported that Cortland County had 488 cases since the pandemic began. Cortland reports 733 — a 245-case difference. Analyze the numbers daily and you’ll find that on most days, the county has a few more positive cases in its tally than the state reports.
“We don’t know what’s accurate, really,” Morley said. “If you’re not testing every person every day, all you have are approximations.”
The state and county data report two extremes. The state shows the lower number of positive tests, but the higher number of tests administered — the best possible scenario. The county reports the higher positives, but fewer tests — the worst case.
Is one agency or the other necessarily incorrect? Is it possible that the county’s higher number of positives is more accurate, but the state’s report of more tests is a better reflection of testing in the community?
If the two sets of data are combined that way, that generates a seven-day rolling positivity rate somewhere between the 2% or less the state sees, and the 6% or more the county sees. The rolling positivity ranges from a bit less than 3%
to just over 4%.
If the 6% positivity is the level at which Cortland County becomes a “red zone” — forcing schools and businesses to close, scaling restaurants back to takeout only and churches to one-quarter capacity and banning all mass gatherings — 4% is the level for the “yellow zone,” the least severe of the state restrictions.
That designation would ban mass gatherings of more than 25 people and would limit churches to 50% capacity. But it would let non-essential businesses remain open and would let people eat in restaurants, albeit no more than four people at a table.
What to believe
Which numbers should one believe? At stake is how the community can respond, or would be required to respond, to the coronavirus spike. Harbin throws his trust behind the county numbers.
They’ve been recorded and publicized daily since March. As a legislator, he knows both the people who compile that information and how it’s compiled. He’s not as certain about the state.
And Morley suggests Harbin has reason, at least for the number of positive cases. Because of how students are tested — and insurance companies billed — at least some of that information is transmitted to their home communities, and the state doesn’t get that data. At least some of the 500 cases the college has reported in the past two months aren’t counted as coming in Cortland.
“A big skew is probably in the college numbers,” he said, and the county’s 733 cases as of Friday may be closer to the truth than the state’s 488.
But Harbin’s larger concern is the message the state is missing: “What these numbers are telling me is that we have a significant problem in Cortland County and we have to take this much more seriously,” he said. “We’re not getting the governor’s attention. They’re looking at the wrong data.”
However, Morley said, whether Cortland’s positivity rate is at 2% or 4% or 6% is just a line in the sand beyond which the state will force action. It’s not arbitrary, but the trend is more important than the line.
“Simply embrace the fact we’re on the precipice of something,” he said. “Whether you’re at 2% or 6%, you’re on the cusp of having a very bad winter.”