Hypothesis: Improving Treatments Are Making COVID-19 Less Deadly


by James A. Bacon

The Wall Street Journal cited a revealing statistical measure of the COVID-19 virus that reflects upon the ability of different health systems to cope with the epidemic: the percentage of confirmed cases that result in fatalities. That number varies widely across countries and states.

The focus of the article was Singapore and Honk Kong, two densely packed cities that have kept the deaths-to-infections deaths ratio extremely low: 0.4% in Hong Kong and less than 0.1% in Singapore. Not only have those cities done a better job than the U.S. of containing the spread of COVID-19, the virus has proved to be less fatal when people get it.

By contrast, the case-fatality rate in the U.S. overall is 5.9%. That’s high, but not as high as the United Kingdom and Italy, where the figure stands at 14.%. In the U.S., New York City the rate is 8%, according to the Journal.

What about Virginia? Our case fatality rate peaked around 3.6% and has since declined to 3.2%, based on data published on public dashboards this morning. As can be seen above, the rate increased steadily from early April to early May, crested May 9 and 10, and has declined fairly steadily since.

What accounts for the differing case-fatality rates?

“When you overwhelm health systems a lot more people die,” the Journal quotes David Owens, founder of Hong Kong medical practice OT&P Healthcare as saying. “Hong Kong and Singapore “didn’t let the epidemic get out of control.”

Also, writes the Journal: “Keeping infections from spreading to communities most at risk allowed both cities to manage their medical resources. Doctors and researchers, free from being inundated by a crisis, have been able to focus on finding treatments.”

And that, ladies and gentlemen, brings us to a crucial but neglected point in the COVID-19 debate. As docs and researchers gain more experience with the virus and more scientific understanding of how it attacks the body, they are getting better at treating the disease.

That raises interesting questions here in Virginia. Have Virginia hospitals done a better job on average compared to their peers in other states of treating the virus? The fact that our case fatality rate is significantly lower suggests that they might have. On the other hand, Virginia hospitals, despite shortages of personal protective equipment, were never overwhelmed to the same degree they were in New York City. Also case fatality rates likely vary by the types of patients being treated. Elderly patients with more coexisting conditions are more likely to succumb. If Virginia hospitals were treating healthier patients on average, that might explain their better performance. Not having access to that data, I’d have to refrain from drawing any conclusions.

A stronger case can be made that Virginia physicians and hospitals, by  virtue of learning more about the virus, are getting better at treating patients — just as their peers are across the country. Are Virginia health practitioners figuring out which types of treatments work best, or which combination of treatments work best in specific cases? The declining case-fatality rates suggests that perhaps they are. There may be other explanations for the declining rate, but it is a reasonable hypothesis.

If that hypothesis holds up — if improving treatments are driving down the fatality rate — we can all take comfort.

There are currently no comments highlighted.

31 responses to “Hypothesis: Improving Treatments Are Making COVID-19 Less Deadly

  1. The denominator drives that number — total number of cases. Deaths are known, but total cases is unknown. Some “authority” in the RTD today reports infections in the US of A could be 10-15% of the population (uh, 30-50 million.) Working backwards from the CDC’s best estimate CFR of 0.4%, the case total would be 25 million. Other countries seem to have far better testing data that the US, and other states far better than VA. Too soon to know….

    But they figured out fast to keep people off the ventilators if they could. All these docs around the world have digital mechanisms for sharing what they are seeing and thinking, so I doubt VA gets brownie points for doing especially well. I suspect the CFR for somebody under 60 who has no metabolic or respiratory disease issues is quite, quite low. Still a nasty bug and you don’t want to get it.

  2. Well.. it was to be expected once it was realized that existing treatments did not work well for COVID19 and it may prove to be invaluable if the quest for a vaccine takes longer or even may not happen.

    And it will make it easier to open up as less deaths occur.

    so the idea no matter what we did – the same number of deaths would occur – may not be true at all.

    ergo – “herd immunity” may happen in a different way though it sounds like some/many of the antibody tests to verify it are bogus.

  3. If the data for cases refers to cases reported as positive, rather than to cases admitted to the hospital, the decrease appears to coincide roughly with the increase in testing in Virginia. Therefore, the decrease in the fatality rate may have more to do with the increase in testing and finding more cases, but which do not need hospitalization, than in improvements in treatments.

  4. I think this has a lot more to do with effectively quarantining the elderly than improved treatments.

    I also question how you can compare the case fatality rate between states or countries with very different levels of testing. More tests presumably means we are testing less symptomatic people. That means picking up more cases where the person has relatively mild symptoms and is likely to recover.

    Meanwhile, in what seems like an obvious event, cases in unlocked South Korea are rising. I’m not sure where singing falls on the breathing, talking, coughing, sneezing continuum but I think going to a karaoke bar is probably a “bridge too far”.

    As far as Northam and his masks … I’ve been staying in Maryland since March 18. Hogan ordered masks to be worn inside businesses on April 15. It was a non-event. Everybody wears masks and I’ve seen nothing about people refusing or protesting the wearing of masks. I imagine that the employees in a store would ask you to stay out or leave if not wearing a mask. Just like employees in most restaurants would ask you to leave if you are not wearing shoes. I’m not sure why this requires so much hand wringing in Virginia.

    https://apnews.com/72fae695d7df85accb1bf449568a8a1f

    • Meet me at the trench line at Cold Harbor and I’ll ‘splain it to you. Or the Byrd Statue on Capitol Square if you prefer! Me, I’m fine with the masks for now.

      This from our next governor….”I Will Not Consent! In lieu of the governors hypocritical Virginia Beach actions of not following his own face-mask mandates, and his 6’ rule; I’m encouraging my supporters, those who believe in our Rights as Virginia’s, supporters of Freedom, ‘We The People,’ and YOU … to stand with me and REFUSE to be mandated and dictated by this liberal agenda and the left any longer! I Will Not Be Mandated to Wear any Mask!.” (Chase, of course.)

      Oh, and this from “Virginia Constitutional Conservatives”: “…Then we need every single one of you to call the Governor’s Mansion and tell Radical Ralph you will never comply with his unconstitutional orders. Call right now at (804) 371-2642.”

      Sigh….

      • How did that stand at Cold Harbor work out for the Confederacy? Lol.

        Masks are a nit. Reopening the economy is not. I’d happily give in to King Ralph’s masks mandate in return for a clear description of how and when the economy will be reopened in full.

    • re: ” how you can compare the case fatality rate between states or countries with very different levels of testing.”

      I’m not following. Don’t folks get sick and go to the hospital THEN get tested such that overall testing over and above the hospital testing has no correlation?

      I guess I would have thought the correlation should be hospitalizations verses deaths in hospital… no? Then you could compare – even hospitals to each other if that level of data was provided.

      My bet is that not all hospitals are doing the same thing and that when one has success with a treatment – it may not be universally shared – at least right away.

      One thing they have backed off of is the ventilators…

      • I’ve been told by Virginia doctors that they have followed a strict protocol on when they can prescribe a COVID19 test. Before we ramped up testing a patient had to have unambiguous symptoms to get a test prescription. Now that we have more tests I assume that the rules for who can get tested are being relaxed. I hear there are some places where anybody who wants a test can get one.

        When you test only the obviously sick you get a lot of very sick people. When you test anybody who has been sneezing recently you get asymptomatic and semi-symptomatic people. Those who don’t show symptoms despite being infected don’t die. More positives, fewer deaths because you are picking up asymptomatic and lightly symptomatic people.

  5. Jeez, Larry, are you in the basement? Testing in drive through parking lots, testing at pharmacies, testing in public housing complexes, testing in places of work, teams testing everybody in a particular nursing home or prison – people showing up in primary care offices and ERs may now be a small minority of those tested. My granddaughter in Texas got an antibody test before getting her ear tubes replaced. Some of the states are getting 50,000 results a day.

    And people die in nursing homes or group homes without going to the hospital, or die at home after not going to see anybody about their symptoms. Not all deaths are inside the hospitals.

    • I’m sure I misunderstand but is there any correlation between the NUMBER of tests given and the death rate?

      One places could have few tests and a high death rate and another place a lot of tests and a low death rate, right?

  6. Well, I’ve got to say, if you track deaths as a percentage of hospitalizations, as opposed to cases, you get a very different curve. The fatality rate is getting worse, not better!

    I tracked deaths as a percentage of cases, because that was the Wall Street Journal’s basis for comparison.

    • well tracking deaths as a percentage of “cases’ where a case is a person who tested positive?

      well that’s far removed from hospitalizations, no?

      seems like if you want to determine if a given hospital is doing better at saving victims.. it would be to compare the number of victims hospitalized verses the number who survived… otherwise, several other factors are involved…

    • I think most people get lost in the correlation.

      Everything is just a subset to number of positive tests (ie. negative, hospitalized and deaths) . As any comparison must be made to that number, the largest issue is we don’t know how many cases there are.

    • That graph is interesting. It would be more interesting by age cohort. Are we successfully protecting the elderly or not?

    • Don’t let ’em take you to the hospital!

  7. Well, Larry, if you want the real fatality rate you need to know in full how many have been infected, and from the beginning the testing data has NEVER come close to reflecting that. You’d have to test everybody with a 100% accurate test, for active infection or antibodies. You can certainly calculate deaths per confirmed cases, but that won’t be the real CFR number. We’ll never know the real number, just better and better approximations as time goes on. Sadly, Sars-CoV2 is here to stay.

    • okay – so here’s the part I was focusing on:

      “the ability of different health systems to cope with the epidemic: the percentage of confirmed cases that result in fatalities. That number varies widely across countries and states.”
      ……… [deleted stuff]
      What about Virginia? Our case fatality rate peaked around 3.6% and has since declined to 3.2%, based on data published on public dashboards this morning. As can be seen above, the rate increased steadily from early April to early May, crested May 9 and 10, and has declined fairly steadily since.

      What accounts for the differing case-fatality rates?
      ……… [more stuff deleted]
      And that, ladies and gentlemen, brings us to a crucial but neglected point in the COVID-19 debate. As docs and researchers gain more experience with the virus and more scientific understanding of how it attacks the body, they are getting better at treating the disease.”

      so if you start with CASES – instead of hospitalizations – then don’t you muddy things up than if you looked at hospitalizations (regardless of the “cases” or numbers of tests) – just who actually got admitted to the hospital – then how many of those who actually were hospitalized that did not survive.

      Then compare that ratio among different hospitals to see if some hospitals did markedly better than others… perhaps because they actually did find better, more successful treatments?

      so how is that thinking wrong?

      I found Jim’s article a bit all over the map – apples, oranges, bananas, kumquats… 😉

      but yes how many went in and how many came out…on a per hospital basis. when you back off of that – you’re introducing all kinds of other factors.
      .

  8. So, the counter to my hypothesis goes like this: If we’re testing more, we’re reaching deeper into the population pool. Instead of focusing on the highest-risk patients (judging by their symptoms and severity of illness), we’re getting more people with sneezes and coughs who don’t have COVID-19. Therefore, as time goes by, it’s only natural that the ratio of positive tests is declining.

    If the prevalence of testing is the driving factor, explain this: Virginia, which has one of the worst testing records in the country (we all agree on that), has a significantly lower rate of case fatalities. It ought to be the other way around. If we’ve been testing only the sickest patients with the best match with COVID-19 symptoms, we should have a higher case fatality rate. True, testing in Virginia is spreading to a broader population and picking up more non-COVID cases. But so is testing in other parts of the country. I don’t think the counter-argument holds water.

    • Actually, the counter to your hypothesis is the plot by, oh, you. Deaths per hospitalized. If treatment were improving then the curve would be trending down. Hospitalized cases are way more likely to be true infections than the population screening tests, especially the serological tests.

      The problem with the first plot was just made public today — upto 50% of the serological tests may be false positives.

      More and more tests, more and more false cases, and the more it bends the curve down.

      You could fix it if VDH hadn’t mixed the results of the two test types.

      • As I’ve been saying for, like, eight weeks: I really want to know how FDA so totally screwed up the testing process. What really good test, maybe from a foreign source or private lab, got rejected, etc. Agreed the blood antibody tests are highly unreliable, but I think one of them had far better reliability.

        I would agree if the % of hospitalized who die seems to be stable, not a sign they are learning how to cure this.

        • This doesn’t bode well for The Age of Information as the economic follow up to The Industrial Age. This pandemic was the best opportunity to show the benefits of data mining and the new tools, and it was a failure of epic proportions.

          Given Twitter news as of late, we may have just skip the Information Age and gone directly to The Neo-Dark Ages where the role of The Church is played by social media.

          OTOH, it shows we’ve less to fear from Big Brother than Snowden implied.

  9. re: ” the blood antibody tests are highly unreliable”

    way worse than that – we don’t even know how many different ones are being used much less the reliability of each and apparently many were FDA-approved early on… and CDC and VDH were mixing the two different kinds of tests and calling them number of tests per day…
    woe is us

  10. Maybe it’s triage not treatment.

    https://nymag.com/intelligencer/2020/05/coronavirus-studies-updates-good-news-bad-news-herd-immunity.html

    Good roundup of preliminary studies. See #4.
    They found that three biological markers are so predictive of mortality, they can signal whether a COVID-19 patient will develop life-threatening illness with 90 percent accuracy more than ten days ahead of time. The three so-called biomarkers, all of which can be measured using a single drop of blood, were:
    1) Elevated levels of the enzyme lactic dehydrogenase (LDH).

    2) Low levels of lymphocytes (i.e., white blood cells).

    3) High-sensitivity C-reactive proteins, which are indicative of respiratory inflammation.

    By widely testing for these biomarkers, medical professionals could separate patients who will experience COVID-19 as a bad cold from those who will experience it as a mortal threat long before the distinction becomes apparent to the patients themselves. This would allow hospitals to better concentrate resources and thus, theoretically, reduce COVID-19’s overall fatality rate.

Leave a Reply