This afternoon, something quite unusual happened. Major news outlets like the New York Times, the Washington Post, CNN, the Wall Street Journal, USA Today, and Axios all reported that the CDC had stated that 73% of U.S. COVID-19 cases were now caused by the Omicron variant. Prominent public health figures such as Dr. Ashish Jha, Dr. Eric Topol, Florian Krammer, Dr. Eric Feigl-Ding, and Dr. Scott Gottlieb quickly shared these findings without much scrutiny.
However, upon closer inspection, that’s not exactly what occurred. There seems to have been a glitch in the CDC's automated statistical algorithm called the variant proportions Nowcast. Omicron isn't yet the dominant strain in most parts of the U.S., though all signs suggest it will be soon. Somehow, this error evolved from a quiet database update on a minor CDC webpage into a nationwide media frenzy within hours.
Let me clarify what actually transpired. No official announcement was made by CDC officials regarding the 73% figure, nor was any press release issued. Instead, this information surfaced on the CDC’s automated COVID data tracker site.
The Nowcast algorithm relies on the CDC’s variant proportions data. Each week, roughly 3.6% of U.S. COVID-19 cases undergo genome sequencing to identify the specific strain. This percentage varies by state. The CDC aggregates this data to produce weekly reports on the prevalence of each variant across ten regions in the continental U.S. However, the genomic data requires time to generate, process, and analyze before it's incorporated into CDC reports—typically around two weeks behind (as of today, the data is 16 days old; for the week ending December 4).
To provide an estimate of current prevalence, the CDC uses the Nowcast algorithm to project forward from two weeks ago. Historically, this tool has been incredibly accurate and useful. As Alpha, Gamma, Delta, Mu, and other variants spread across the U.S., I closely followed the Nowcast, which proved invaluable in tracking their progression. The CDC deserves praise for developing such a sophisticated system.
Unfortunately, something appears to have gone awry with the Nowcast this week. During Monday's update, the algorithm produced results that don't align with reality.
Looking at the national numbers tells part of the story. The data from the previous two weeks indicates that Omicron accounted for 0.0% of cases sampled (regionally adjusted) in the week ending November 20, 0.1% in the week ending November 27, and 0.7% in the week ending December 4. According to the Nowcast, these numbers surged to 12.6% in the week ending December 11 and 73.2% in the week ending December 18.
In essence, the Nowcast isn’t suggesting that Omicron accounts for 73% of current U.S. COVID-19 cases. Instead, it claims that Omicron made up 73.4% of cases during the average day of a week starting on December 12. This implies that Omicron’s prevalence increased twentyfold in one week, with a doubling time of about 36 hours sustained over two weeks.
Such growth is inconsistent with everything we know about Omicron. Responsible estimates from countries with better data, like the UK and Denmark, suggest doubling times ranging from two to four days. This has led experts to describe Omicron’s spread as "explosive" and predict its dominance in the U.S. within weeks. Yet a doubling time of 36 hours sustained for over two weeks defies all available data—it simply doesn’t add up.
A glance at regional numbers reveals even more anomalies. Region 10, comprising Oregon, Washington, and Idaho, shows 0.0% Omicron prevalence in the last measured week (ending December 4), but the Nowcast claims it jumped to 43.4% in the week ending December 11 and 96.3% in the week ending December 18. With such low actual data in Region 10, how could the algorithm accurately predict such high figures?
Moreover, these numbers don’t match national case counts or our understanding of how variants typically interact. While significant case spikes have occurred in Region 2, where Omicron accounted for 2.4% of cases in the latest measurements, similar surges haven’t been observed in most other parts of the country. This pattern resembles what happened when Delta replaced Alpha earlier this year, but it isn’t happening now.
National case numbers have remained relatively stable until recently, increasing only slightly due to sharp rises in a few states. If Omicron had truly dominated last week, this would mean Delta inexplicably collapsed nationwide at the same pace and timing as Omicron’s explosion. Furthermore, Omicron’s growth would have abruptly slowed once Delta cases reached zero, yet we haven’t seen a corresponding spike in national cases. Region 10, supposedly experiencing the fastest Omicron growth, has even reported declining numbers.
In reality, the CDC’s Nowcast algorithm has malfunctioned, and Omicron wasn’t truly at 73% prevalence nationwide in the week ending December 18. Trust me on this.
So, what does this mean? What happens next? Here’s what you need to know:
1. This will be corrected. The CDC will issue an explanation. Experts will retract their initial statements. Media outlets will publish clarifications. It may happen tomorrow, later this week, or next week, but it will happen.
2. None of this reflects poorly on the CDC. They didn’t announce this. It was an automated algorithm, and one with a strong track record prior to this incident. The emergence of Omicron has accelerated events, and the genetic surveillance infrastructure, including the Nowcast, hasn’t been tested under conditions of extreme exponential growth. Bugs were bound to emerge, and this one happened to gain traction rapidly.
3. The press missed this story because they aren’t monitoring the details closely. Outlets like Science magazine, Endpoints, and the MIT Technology Review (and myself) haven’t covered this. Disappointingly, Stat News and the Financial Times have, but they likely encountered the Nowcast for the first time, don’t fully grasp it, and didn’t investigate thoroughly. Social media pundits, including some experts, propagated the initial simplistic coverage. This is just another example of the press’s broader struggles with scientific reporting during the pandemic.
4. Omicron is indeed spreading explosively and will likely dominate the U.S. soon. It’s probably already dominant in places like New York, New Jersey, Florida, Hawaii, and Puerto Rico, where case numbers are surging. This algorithm error and media frenzy have jumped ahead of reality, but they are temporary glitches. The bigger picture remains that Omicron is spreading explosively and will soon take over.
What should you do?
1. Continue preparing for Omicron’s dominance in your area within weeks—or sooner. This is real and ongoing.
2. Watch how this story evolves.
3. Stay cautious about what you read.
Stay safe, everyone.
Dining Table And Chairs,West Elm Dining Chairs,Kitchen Chairs,Black Dining Chairs
CAMOSA HOME , https://www1.camosafurniture.com