Why it’s so hard to make quick, accurate estimates on new COVID variants

On the finish of December, the CDC sharply revised its estimates of the prevalence of Omicron relative to different COVID variants throughout the US. Earlier than Tuesday, it reported that Omicron made up 73 % of recent COVID circumstances from December 12 to 18. Now, it places the prevalence for that week at 23 %, and estimates that Omicron made up 59 % of recent circumstances between December 19 and 25.

As Popular Science reported recently, that decrease estimate makes it even more durable to grasp the US’s outbreak—Delta is clearly nonetheless a dominant pressure, and is probably going accountable for most deaths.

Why such a dramatic swing? The CDC can’t really see each COVID case, so it’s counting on a modeling idea known as a nowcast. (Now + forecast.) That mannequin predicts unfold primarily based on the traits of COVID variants and the overall inhabitants, and tethers that prediction to real-world information.

Proper now, there’s nonetheless appreciable uncertainty over how shortly Omicron spreads relative to Delta, and the way simply it’s in a position to infect individuals with prior immunity. “The mathematical fashions that describe COVID transmission dynamics have gotten increasingly advanced,” says Anass Bouchnita, an infectious illness modeler with the College of Texas at Austin’s COVID-19 Modeling Consortium. “We’ve got a number of new results”—like booster efficacy and prior immunity—“that should be integrated, and now we have to do that very quickly, given how shortly the panorama of COVID is altering.”

Bouchnita and colleagues produced a model of Omicron’s spread in mid-December, primarily based on the info that was then obtainable concerning the variant’s properties. “In essentially the most pessimistic situation we had been anticipating that the height would attain about 500,000 circumstances per day,” he says. Since then, it’s develop into extra possible that the Omicron surge will come on quicker and sharper. “After we revise the projections, we anticipate that circumstances will attain 700,000 per day,” someday in January.

However quirks within the underlying information additionally led to the change. As a result of Omicron spreads so shortly, small modifications in our understanding of its preliminary prevalence can result in massively totally different mannequin outcomes. To a sure extent, these information quirks are most likely one thing fairly primary: The US was on the lookout for Omicron circumstances, and so it discovered a number of them. “One of many the explanation why the CDC adjusted their prevalence numbers is as a result of so many labs particularly appeared for [Omicron] samples,” says Krista Queen, who oversees COVID sequencing at Louisiana State College Well being Shreveport. “If you go on the lookout for that [variant], it’s going to inflate the prevalence.”

Omicron is comparatively simple to identify with a PCR check, even with out sequencing. PCR exams search for a handful of particular sequences from the virus’s genome. And it simply so occurs that a few of them are designed to search for a sequence that, on Omicron, has mutated into an unrecognizable type. When such exams are used on a case of Omicron, they return a partial constructive: three yeses for the intact sequences, and one no, known as an “S gene goal failure.” That signature isn’t a slam dunk, because it seems in a number of different strains. However it provides researchers a clue.

[Related: Omicron isn’t overtaking Delta as quickly as the CDC thought—and that’s bad news]

These apparent PCR outcomes are a big a part of how the US’s genomic surveillance system confirmed the presence of Omicron shortly. Well being departments throughout the nation appeared for uncommon circumstances with S gene goal failure, and sequenced them. “Early in December,” says Queen, “the Louisiana Division of Well being”—like well being departments throughout the nation—”was prioritizing samples that had S-gene goal failure.”

Queen’s group at present sequences each pattern that is available in from LSU’s testing lab, which works with healthcare services and neighborhood testing websites, which supplies them a clearer image of Omicron’s prevalence. Nonetheless, that nationwide sequencing bias possible led to an overestimate of the true prevalence of Omicron.

Then, modelers must account for testing bias. Vaccinated individuals are much less prone to get COVID in any respect, however usually tend to catch Omicron, with its immune evasion properties, than Delta. And a few information means that vaccinated individuals are additionally extra prone to search testing. “It’s solely individuals who go to get examined that we’re sequencing,” says Queen. “So are we seeing extra Omicron as a result of it’s really extra prevalent? Or due to who’s working out to be examined?”

However focusing an excessive amount of on Omicron’s prevalence nationally could be deceptive: The pandemic seems to be very totally different in other places, with lethal Delta outbreaks raging within the higher Midwest, as Omicron surges in extremely vaccinated cities like Seattle.

Extra essential is knowing the prevalence of Omicron at an area stage. That’s partly as a result of the illness behaves considerably in another way, spreading quicker and extra simply among the many vaccinated, which calls for various security precautions. However, Queen factors out, it is usually essential for allocating COVID medicine. “Demand proper now for any remedy is thru the roof,” Queen factors out. Some antibody remedies for COVID sufferers don’t work as properly for Omicron, so areas with the brand new variant want totally different medicines. “Are states in a position to meet the demand? That’s what issues far more than a gross total estimate for the US.”


Leave a Reply

Your email address will not be published.