COVID-19 Update: Time to Re-Evaluate Usefulness of “Confirmed Cases” Statistic

by James A. Bacon

The latest Virginia Department of Health (VDH) data continues to look alarming: 562 new reported cases yesterday, 75 new hospitalizations, and 27 new deaths. However, I’m sticking by my argument that the epidemic in Virginia is close to peaking. Here’s the nub of my argument, the number of confirmed COVID-19 cases is bifurcating from other data sets measuring the spread of the disease. I am questioning whether it is even a meaningful figure anymore.

Steve Haner has pointed me to an ABC News article which makes the point that COVID-19 infection is far more prevalent than indicated by officially recorded tests. Recent community testing in Santa Clara County suggest that the actual number of infections is 50 to 80 times higher than the original numbers. A Wall Street Journal op-ed today makes the same point.

If those same ratios apply to Virginia, the number of actual COVID-19 infections, as opposed to officially confirmed infections, could stand around 400,000 to 500,000. Of those, fewer than 1,300 have been hospitalized and only 258 have died. Those are tiny percentages, not much different from a typical influenza season. The bifurcation in trend lines can be seen in this graph showing new confirmed cases, hospitalizations and deaths reported by the VDH:

The number of reported cases is increasing far more rapidly than the number of hospitalizations and deaths.

Rather than providing an accurate measure of the spreading of the disease, I would suggest, the number of new cases is an indicator of Virginia’s healthcare providers ability to target their tests. In the early phases of the epidemic in Virginia, testing was scattershot. The number of negative results was extremely high. Doctors and nurses were testing a lot of people who did not have the virus. But as healthcare providers got better at recognizing systems, and as they prioritized patients they deemed most likely to have the disease — typically those who were in the hospital already — the percentage of “hits” has increased over time. You can see that trend in this graph, which expresses the new COVID-19 confirmed cases as a percentage of the number of tests on a given day.

If the official number of COVID-19 cases reflects one out of 50 cases in the real world, it is almost worthless as an indicator of how rapidly the disease is spreading through the population. What better measures are there? The number of hospitalizations and deaths. Those data series have their own issues, but we can feel pretty confident that they are a more accurate reflection of trends than the confirmed- cases number, especially trends among those most severely affected by the disease.

We can tentatively draw important conclusions to guide us as we seek the appropriate balance between saving lives and salvaging the economy:

  1. The disease has spread far more rapidly than anyone heretofore suspected;
  2. It is not not nearly as fatal as feared;
  3. Social-distancing measures have indeed “flattened the curve” and Virginia hospitals are in no danger of being overwhelmed, and
  4. There is far more herd immunity in the population than previously believed.

I conclude by presenting John Butcher’s updated logarithmic graph that shows a bending of the curve for the three key metrics.

There are currently no comments highlighted.

8 responses to “COVID-19 Update: Time to Re-Evaluate Usefulness of “Confirmed Cases” Statistic

  1. The daily report on DOC offenders testing positive shows a definite increase. 65 total positive cases in facilities, with 6 of those in a hospital. (Yesterday, it was 47 total, with 6 in hospital.) That increase in active cases has caused the cumulative total to jump from 52 in yesterday’s report to 80 in the latest report. The big increase seems to be in Unit 13 in Chesterfield County, which now has 20 cases. That is at least 10 percent of the facility’s population. It is a minimum security, dormitory facility.

    Collateral damage for DOC is 50 staff members testing positive.

  2. Cranky’s flattening curve is good news, and fingers crossed. But it is not flat and not descending. And I’d stop tracking “cases.” As testing swells, so will cases. We wait for that to descend and we will release the lockdown in 2022.

  3. Well are we actually , accurately tracking deaths attributable to COVID19?

    Are the medical examiners testing deaths that are not from hospitals?

    is there still a “normal” amount of deaths that occur outside of hospitals?

    IF we are, then I AGREE that the lethality is not what was projected earlier.

    Would also point out that the idea that people who recover , recover fully. There is evidence that some people suffer lifelong organ damage.

    Finally – all the emphasis on the govt “shutting down the economy”.

    If the public fears contagion – you don’t have to worry about the govt shutting down the economy, it will happen organically.

    Only those not in a right mind are going to go to a sit-down restaurant or attend a sport event if we are seeing thousands and thousands of cases even if the deaths are just a percent.

    Who wants to take a chance of getting it and dying?

    Yep – there are quite a few who will hide under their bed and are more than willing to let Darwin do it’s will!

    As an aside… how many can imagine telling their father or mother that they are “toast” because we have to get back to work? Tough concept.

    Life is not that simple… except to some.

    • Kidney failure and dialysis in ICU Covid patients https://www.nytimes.com/2020/04/18/health/kidney-dialysis-coronavirus.html

      To respond to your other point, I don’t believe that restaurants, entertainment, or personal services are going back to normal until people feel safe. That’s going to ultimately be a much bigger constraint than any government edict.

      If remdesivir is as effective as early rumors suggest, that’s very helpful. If the pill form antivirals are effective, that’s even better. If the vaccine efforts are successful, that changes it completely.

      People don’t want to be hospitalized for weeks, or end up with organ damage or failure. They don’t want to accidentally kill family or friends. Dying yourself isn’t the only risk.

      • I agree that the gov’t decrees won’t matter. This will be a class-based reopening.

        Middle and upper middle and upper class workers will not be going “back to work” or “back out” when the gov’t decrees expire. They will continue to work from home and have groceries and food delivered. They will do this for 2-4 weeks to make sure the “coast is clear” and there isn’t some spike in cases or hospitalizations.

        The virus’ class and racial effects have been disproportionate so far. The “reopening” has the potential to make those disproportional numbers explode.

  4. Jim says:
    “We can tentatively draw important conclusions to guide us as we seek the appropriate balance between saving lives and salvaging the economy:

    The disease has spread far more rapidly than anyone heretofore suspected;
    It is not not nearly as fatal as feared;
    Social-distancing measures have indeed “flattened the curve” and Virginia hospitals are in no danger of being overwhelmed, and
    There is far more herd immunity in the population than previously believed.”

    Yet, still we had the disaster, and potential of debacle in New York City. How do we avoid that again near term with required certainly while returning to near normalcy needed too survive economically and socially in New York? And how does this relate to various areas of Virginia, in all their variety, socially, economically, health care wise, especially in short term.

    I see short term still fraught with dangers. Yet we have to be unafraid to proceed, take the risks, accept the losses, in order to achieve the gains we need to get across this near term dangerous ground. As you point out, Jim, this is what we do already on many fronts of risks and risky business we take and engage in willingly to gain the benefit of managing that risk to acceptable levels. We need to adjust to this new kind of risk mentally as well as physically. That’s where new federal policy is trying to push us, nudging us at first, then pushing. How will the various states respond? And Federal government react? A key question.

  5. The Covid checks from Trump will be the reason the government reopens to many, ready or not. By the way, good arguments from all. The truth is, we just don’t know!

Leave a Reply