Federal officials still estimate the omicron variant is the dominant COVID-19 strain in the U.S., but not nearly to the same degree they did just a week ago.
The Centers for Disease Control and Prevention on Tuesday revised its estimate for the week ending on Dec. 18, surmising that the omicron variant accounted for about 23% of cases that week. Officials initially estimated that omicron accounted for 73% of cases then, and the agency announced on Tuesday it estimated the omicron variant accounted for 59% of cases in the week ending on Dec. 25.
Dr. Amesh Adalja, Senior Scholar at the Johns Hopkins Center for Health Security, told KCBS Radio’s Eric Thomas in an interview on Tuesday afternoon the revision could have had major real-world impacts. Many hospitals last week suspended the use of monoclonal antibodies because of a shortage of the lone treatment effective against the omicron variant, which wasn't yet the dominant strain.
"(There) could be patient harm that came from that because they assumed everything was gonna be omicron when there was still delta around and didn't treat people who might have been able to benefit," he explained.
The CDC has faced criticism this week after shortening the recommended quarantine and isolation period from 10 days to five on Monday, without requiring a negative COVID-19 test. Adalja called Tuesday’s revision "horrible" coming "from a public health communications standpoint" due to ample "mistrust of public health authorities" around the country.
What headlines last week about omicron’s estimated dominance failed to convey, Adalja said, were the projection’s upper and lower bounds.
The CDC’s website mentions up to a three-week lag time to publish reported sequencing data, and the agency relies on its "Nowcast" projections in the meantime, which Adalja said are based on a limited sample size. Those estimates also include a range of outcomes.
For instance, the agency’s omicron projections for the week ending Dec. 25 included a 95% prediction interval with an upper bound of 74% and a lower bound of 41.5%. In other words, 95% of the time, the omicron variant will have accounted for anywhere between 41.5% and 74% of cases for the week ending on Christmas.
"So we're always making statistical inferences based on a select number of samples, and there's wide confidence intervals around them, meaning that they might give you a number, but the confidence in that number is very low, and the upper and lower bounds of what it could actually be are wide," Adalja explained.
"And that's not often conveyed in a headline, and people make decisions, a lot of times, based on headlines rather than the small print," he added.
Given omicron’s transmissibility, Adalja said it "was going to be dominant at one time or another." But last week's projections, and the subsequent reporting, failed to paint an accurate picture of its spread.
Adalja said that’s a mistake that can’t be repeated as the pandemic enters its third year.
"I think there needs to be a lot more scrutiny of what happened, and a lot more criticism of what happened," he said.
Related
