In a now-familiar pattern, what goes down must go up. Nominally, the seven-day moving average of the U.S. rate of new COVID-19 cases rose to 36 per 100K per day (up from about 31, last Friday).
But in reality, that’s mostly or entirely the final effect of Memorial Day on the reporting of new positive tests. In all likelihood, the actual new case count, absent the data reporting glitches, was more-or-less flat.
Data source for this and other graphs of new case counts: Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 6/7/2022, from https://github.com/nytimes/covid-19-data.” The NY Times U.S. tracking page may be found at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html
In percentage terms, the increase we saw today was slightly less than the drop we saw a week ago. The product of the two one-day changes is less than 1.0, suggesting that we haven’t started on any summertime increase in the official COVID-19 case count yet.
But what’s the real infection rate?
Source: Percent with self-reported COVID-like symptoms from Carnegie-Mellon COVIDcast. Percent with positive test results from NY Time data cited above.
The chart above shows the official count of positive tests (orange line), and the fraction of respondents to a Carnegie-Mellon university on-line survey who report having COVID-19-like symptoms.
Fully acknowledging how loosey-goosey on-line surveys are, for sure, of late, a whole lot of people think they might have COVID-19, compared to the official count of positive tests.
But if we look at the fraction of all physician visits that appear to involve COVID-19, these people aren’t checking in with their doctor about their symptoms. By and large, the physician-visit line mirrors things like hospitalizations. Those haven’t risen much during the Omicron-II wave of COVID.
Source: Percent of physician visits for COVID from Carnegie-Mellon COVIDcast. Percent with positive test results from NY Time data cited above.
Taken together, these two pieces of data match my subjective assessment of what’s going on.
- Yes, there’s still a lot of COVID-19 in circulation in the community.
- Yes, a whole lot of people are still being infected, daily, with COVID-19.
- No, that’s not generating a lot of serious illness.
- No, people aren’t having contact with the health care system or getting any sort of official test.
That’s about as far as I can take it. Based on the number of people reporting symptoms, COVID-19 is still pretty much rampant in the population. But because there’s so much cumulative immunity built up (via vaccination or prior infection), an increasing fraction of new cases requires no medical intervention.
And so, as Omicron-II has taken over (BA.2.12.1), new infections have increased, in line with the overall greater infectiousness of BA.2.12.1 relative to the original strains of Omicron. But that’s not translating into a proportionate number of cases requiring medical intervention. As a result, the official count of positive tests, and hospitalizations, and deaths, all show little new activity.
I guess this is how COVID-19 finally makes the transition into being a flu-like illness. Most people don’t need or seek help dealing with a case of seasonal flu. Best guess, that’s what’s now happening to COVID-19 under Omicron-II.
And that’s probably why you keep hearing about so many people who have been infected recently, despite no profound uptick in any of the official measures. Testing at home and dealing with it yourself has become the new norm.
Going forward
I’m not quite sure what this means for tracking COVID-19 versus flu going forward. Really, in anticipation of a winter wave of both.
Right now, flu season is on the decline, at least based on lab testing. So there’s little doubt that what we’re looking at, in the self-reported symptoms chart above, is COVID-19. This is from the most recent week of CDC flu tracking, as of this writing:
Source: US CDC, Weekly influenza surveillance report, accessed 6/7/2022.
But think about the difficulty of tracking a disease that is common, but for which people rarely seek testing or treatment. How would you do that?
For flu, at the end of the day, the CDC does its overall impact-of-flu estimates by starting from the hard (i.e., reliable) data, hospital admissions for flu. It then uses the historical relationship between total flu cases, and flu hospitalizations, to inflate the hospitalization number up to an estimate of all symptomatic flu cases in the country.
Meanwhile, other tracking systems rely on symptoms. One way or the other, they look at people who show up (in a survey, in a physician’s office) with “influenza-like illness”. And if the symptoms are flu-like, that’s good enough for the running day-to-day count.
But with COVID-19, there is no stable historical relationship between new cases and new hospitalizations. That number has been all over the map during this pandemic, and now appears to be falling.
And we now have two disease that will be common — flu and COVID-19 — that, at least on the surface, seem to share a lot of symptoms for mild cases. I have to wonder about the extent to which we’ll ever be able to keep them separate in any symptoms-based tracking, and I wonder about the extent to which the CDC and other flu surveillance systems are now responding to both flu and COVID-19 in the community.
All told, I’m not expecting to see a whole lot of clarity on this issue, come this winter. At the end of the day, this probably only matters greatly for the unvaccinated-and-uninfected population, for those whose health is fragile, and for institutions such as nursing homes and assisted living facilities.
For those institutions serving the elderly, the vastly greater infectiousness of COVID-19 means that they probably would like to know when COVID-19 is prevalent in the community, compared to flu. And at this point, I’m not seeing any way for them to know that.
