New Haven Register (New Haven, CT)
COVID tracking more confusing than ever
Experts: Changes in virus test reporting have distorted metrics
“We’re still grappling with recognizing the limitations of the data that’s available.”
Dr. David Banach, hospital epidemiologist at UConn Health
How nervous you are right now about Connecticut’s COVID-19 uptick might depend on which metric you’re tracking.
By test positivity rate, Connecticut is in crisis. Over the past week, more than 11 percent of the state’s tests have come back positive, a level not previously recorded since January, amid the state’s winter surge.
By case counts, on the other hand, the situation appears more manageable. As of Friday, Connecticut had averaged 900 cases a day over the past week, up from about 300 a day during parts of March, but nowhere near the 10,000 a day the state saw in midJanuary. So which of the two metrics is truly accurate in representing Connecticut’s current level of COVID-19? Well, maybe neither one.
Due to changes in how tests are reported, as well as the emergence and popularity of at-home testing, experts say both positivity rate and case counts have become skewed, making comparisons with previ
ous COVID-19 waves difficult. This comes as hospitalization totals, another stalwart metric throughout the pandemic, are distorted by the presence of “incidental” cases — those in which someone tests positive for COVID-19 after being hospitalized for an unrelated reason.
The result: A lot of confusion about Connecticut's true level of transmission, among those who follow the pandemic professionally and the general public.
“We're still grappling with recognizing the limitations of the data that's available,” said Dr. David Banach, hospital epidemiologist at UConn Health. “I'm still struggling, and I think a lot of people are, with how to interpret it.”
Imperfect metrics
The COVID-19 metrics reported each day by the state Department of Public Health have always been flawed and incomplete in certain ways, but during previous waves they at least offered a common language with which to discuss the pandemic: A 5 percent positivity rate was high, while a 10 percent positivity rate was extremely high. If your town was in red on the state's case map, maybe it was time to mask up.
But the emergence of at-home testing has changed that in a hurry. Unlike during previous COVID-19 waves, when nearly all test samples were collected by care providers and reported to DPH, patients are now able to test themselves, with no obligation — or even ability — to report their results to health authorities.
This, of course, has led to a dramatic under-counting of cases. Dr. Scott Gottlieb, a Connecticut resident and former head of the Food and Drug Administration who has closely followed the pandemic, estimated last month that the Northeast was detecting “one in seven or one in eight [COVID-19] infections.”
But while home tests have led to an under-counting of cases, they may be contributing to an inflation of Connecticut's positivity rate, as asymptomatic people who need a negative test to see family or attend an event are now more likely to stick with an athome test, leaving PCR tests for those with significant symptoms.
“Why is our positivity 10 percent? In part, because I think the people getting PCRs already know they're positive and want that confirmation by PCR, or had been testing negative on self-test kits, but are symptomatic and are going in and getting a PCR,” Dr. Manisha Juthani, Connecticut's public health commissioner, said last week. “So I think our positivity rate does not necessarily reflect what it did before.”
The state has also recently tweaked how it reports COVID-19 tests, no longer sharing negative results from rapid tests, which previously accounted for about 10 percent of the state's total. That means thousands of negative samples are missing from the denominator of Connecticut's high positivity rate.
Amid these changes, Banach says it has become difficult to rely on positivity rate or to compare the current numbers against what the state experienced earlier in the pandemic.
“The positivity rate, I'm still trying to understand how to interpret that based on what the denominator is, because I think that's a little unclear right now,” he said.
But the confusion doesn't stop there. Whereas hospitalizations were once a reliable way to gauge how many people were seriously ill, albeit with a lag of a week or two, those numbers are now inflated by people who arrive at the hospital for one reason, then test positive for COVID-19 upon arrival. These incidental cases can make up as many as 60 or 70 percent of total hospitalizations, one Hartford HealthCare official said Friday.
And while deaths are, in the end, perhaps the most important metric of all, they lag weeks behind cases, and therefore, have never been especially useful in revealing the state of a COVID-19 spike at a particular moment.
It was likely with all this confusion over metrics in mind that the Centers for Disease Control and Prevention created its COVID-19 threat maps, offering county-by-county risk assessments. But even there, the data has become cloudy. The CDC drew criticism earlier this year for drastically changing the thresholds used for its signature map, turning high-risk counties into low-risk counties overnight.
Based on the CDC's old map, all of Connecticut easily meets the criteria for “high transmission” and has for a while. Based on the new map, three counties narrowly qualify as for the “high” category, while four others are still “medium.”
All told, Connecticut residents have access to plenty of data — little of it easy or straightforward to interpret.
The tools available
Still, it's clear that COVID-19 is present in Connecticut in large numbers, and with enough context, it's possible to evaluate how much.
Wastewater surveillance in New Haven, for example, shows a sharp increase in COVID-19 over the past month, but suggests that transmission has plateaued at a level far lower than the state experienced over the winter.
Banach said he looks at the CDC's community transmission map, the more cautious one, as a tool in assessing a community's risk level. He also considers hospitalizations a valuable
measure, less for determining how many people are seriously ill and more for tracking how many people are showing up at hospitals with the disease.
Dr. Ulysses Wu, chief epidemiologist at Hartford HealthCare, said he's concerned less about absolute case counts and positivity rates and more about trends in those numbers. Because while those two metrics might now be telling different stories, they're both pointing in the same direction.
“We are still following the same metrics,” Wu said. “We are looking at positivity rate, we're looking at hospitalizations, we're looking at vaccination rates, we're looking at deaths, we're looking at
ICU data.”
The balance of evidence suggests Connecticut has
seen a significant rise in COVID-19 transmission — maybe more significant than was initially predicted when this uptick began in March — but nothing resembling the state's winter surge, when hospitals struggled to find staff and space amid a crush of patients.
What may ultimately be most important to know, experts say, is simply that COVID-19 is present in Connecticut in large quantities, more than enough for residents to consider taking precautions to protect themselves and their families.
“The risk of COVID in the Northeast at this point is still substantial,” Banach said. “All the metrics would point toward that.”