Preliminary German Study Shows a COVID-19 Infection Fatality Rate of About 0.4 Percent

Good news from a population screening study


Preliminary results are out from a COVID-19 case cluster study in one of the regions worst hit by Germany's coronavirus epidemic. They are somewhat reassuring.

One often-heard statistic is the "case fatality rate"—that is, the percentage of people diagnosed with a disease who will die of it. This afternoon that figure stands at 3.5 percent for COVID-19 in the U.S., but this rate is significantly inflated because it does not count asymptomatic cases or undiagnosed people who recover at home. What we really need to know is the infection fatality rate: the percentage of all the people infected who eventually die of the disease. That's what the German study attempts to do.

Over the last two weeks, German virologists tested nearly 80 percent of the population of Gangelt for antibodies that indicate whether they'd been infected by the coronavirus. Around 15 percent had been infected, allowing them to calculate a COVID-19 infection fatality rate of about 0.37 percent. The researchers also concluded that people who recover from the infection are immune to reinfection, at least for a while.

For comparison, the U.S. infection fatality rates for the 1957–58 flu epidemic was around 0.27 percent; for the 1918 Spanish flu epidemic, it was about 2.6 percent. For seasonal flu, the rate typically averages around 0.1 percent. Basically, the German researchers found that the coronavirus kills about four times as many infected people than seasonal flu viruses do.

The German researchers caution that it would be wrong to extrapolate these regional results to the whole country. But they also believe these findings show that lockdowns can begin to be lifted, as long as people maintain high levels of hygiene to keep COVID-19 under control.

NEXT: 4 Reasons Small Businesses Are Struggling With the Paycheck Protection Program

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. Have they started testing on twins yet?

    1. You know who else tested on twins?

      1. Doublemint gum?

      2. Bob Sagat?

        1. The Germans formerly known as the Germans?

          1. Racist, they call themselves the “Deutsch”.

      3. The Director of the Montreal Laboratory?

      4. Posturepedic?

      5. Dr. Hefner?

  2. But they also believe these findings show that lockdowns can begin to be lifted…

    Writing this sentence in a public forum is literally a crime now.

    1. If you lift the lockdowns, the terrorists win.

  3. The German researchers caution that it would be wrong to extrapolate these regional results to the whole country.

    I’m not an epidemiologist, but I would think the CFR would extrapolate pretty well to larger populations with similar healthcare standards, assuming that the pool tested included a diverse sampling across ages and pre-existing health conditions.

    I could understand that infection rates might depend on population density (among other things), but CFR shouldn’t.

    1. It actually might. There is evidence that people who continue to take in new doses of a virus after having already been infected by it are likely to have more severe symptoms than someone who gets infected and then is able to prevent himself from taking in any new doses. The more densely populated a place is, the more likely you are to keep taking in those new doses of the viruses, even if you never step outside of your abode and especially if you live in a densely packed building, like many of the buildings in New York.

      1. This isn’t true.


          First of all, we need to be grateful to the greatest President ever. Donald J. Trump, who led us in the fight against the second worst invader on our shores. To the talented and timely team of scientists helping out, Jerome Adams, Deborah Birx, Anthony Fauci, and all others. To the magnificently courageous health workers across the country who risk all every day. To the workers in essential industries who kept on keepin’ on in the face of desperation. To the entire American populace, who did the social distancing without complaint, and thereby saved perhaps 2 million lives mainly among the 40 million USA people who have two or more comorbidities.

          Thank you.

          Ok, Germany took a sample of the entire population, and confirmed that 15% of their entire population had or used to have C-19. The CFR was 0.4%. Why? The denominator was much, much higher than the number who had been identified as cases in Germany.

          The US is likely to have the same CFR result, a CFR of 0.40%. If we end up with 60,000 deaths, the number ever infected will be 60000 / 0.04 = 15 million.

          Germany did identify a good denominator for the C19 flu CFR. However, we continue to believe the numerator is overstated.

          From probability, let P(CFR|C19) be the probability that a person dies given they have C19 – and gets counted in the CFR as a C19 death.

          Let CM be the 40 million American who have two or more comorbidities. 2.6 million Americans die from these comorbidities each year. 7000 per day.

          The people counting the C19 deaths for the numerator of the CFR presume that P(CM|(C19 and CM) = 0. That is, just because you have the dry lungs, etc, you therefore died because of C19. Dead wrong. Our guess is that a more likely probability is that P(CM|C19 and CM) equals 0.8 – P(Dying|C19 and CM), and that P(C19|C19 and CM) = 0.20. C19 was a contributing cause, but we can’t discount the probability that the comorbidities were a huge contributing factor. Many of these people would have died, sooner or shortly later, even if C19 was not around – certainly having an added comorbidity pushed these wonderful parents, grandparents, all, over the edge.

          If we eventually prove that 15 million Americans had the disease or at least have antibodies against the disease, we are at least partway towards herd immunity.

          Overall, we estimate a practical multiplier to convert the numerator for the CFR is to take .2 x 1 + .8 x .2 = 0.36. 20% of the dead had no comorbidities, and 80% of the victims did have two or more comorbitities. If we lose 60,000, we perhaps should only count 60,000 x .36 = 21,600 as caused BY the C19, as oppose3d to people dying WITH the C19. This would shrink the 15 million exposed population to 5.4 million.

          And reduces the actual CFR to 0.4% x .36 = 0.14%, much like the annual flu.

          America did the right thing, to close us down and practice social distancing. When C19 cycles back again, we should repeat the closing down, but hopefully leave a lot more people working and instead plan to isolate many of the 40 million Americans who are at greatest risk.

          1. Ooops.

            60000 / 0.004 = 15 million.

          2. This post showed immediate signs of closed head injury in claiming Trump to be the greatest president. He’s just another turd in a long line of shit going back decades.

            But rest assured, the next one will be just as bad.

    2. I can understand the caution about that. That area was their epicenter – and everyone there knew it. Which also means resident’s behaviors were affected by being at the epicenter. They may have been able to sample broadly representative demographics, but you can’t correctly sample survivor bias in either direction. Plus from the Guardian story, this is just their first involvement in what is going to be a very comprehensive set of follow-ups and stuff. That always gets a different group of people than those who simply agree to a one-time thing. The entire project is exactly the sort of good public health and preventive practice that almost can’t happen here.

      But I’m sure we don’t need any of that caution. A bit of this from Germany, a bit of that from Singapore, a bit of the other from South Korea, a smidgen from Iceland and Taiwan. Just apply the shiniest objects from elsewhere to ourselves – a bit of we’re America – and hey presto in a couple days Sullum will write an article invoking the 0.1% too and we’ll be right back to this is just the flu.

      1. Interesting article in German re that area. They’ve been under close surveillance by public health since the first case was confirmed after a Feb15 carnival in Gangelt. As of a couple weeks ago, the town next door had 0.5% of its peeps as confirmed cases – compared to 0.007% average in Germany – so 71x higher case rate. Very very low death rate too overall in that district – 0.2% of total German fatalities, 6% of total German cases. May be a lot of reasons for that – but it too could not simply be extrapolated to elsewhere. They got the whole contact tracing/testing stuff very early (mostly late Feb and early March) – but they also got the tires slashed and refusal to deliver supplies from neighboring areas.

        Just not a similar universe at all – but the ideal place for a focus study

  4. //But they also believe these findings show that lockdowns can begin to be lifted, as long as people maintain high levels of hygiene to keep COVID-19 under control.//

    So, not that deadly and, wash your fucking hands. If only early calls for rational responses weren’t routinely drowned out by apocalyptic predictions and the dissemination of insane models where everyone dies.

    1. GG: Let’s do a conservative calculation using these heartening rates and see what it yields. Let’s conservatively say that only 25 percent of the US population is eventually infected. 330,000,000 x 0.25 = 82,500,000. Now let’s multiply that with the preliminary German infection (not case) fatality rate. 82,500,000 x .0037 = 305,250 deaths. Would that be too many excess deaths for 1 year or not? Remember the last year’s flu deaths totaled about 34,000.

      1. Ron,

        If 82 million people is your “conservative” estimation of the number of those likely to be infected, I’d really hate to see what your panicked estimation looks like. Not sure where you are getting that number from. You are also assuming, and I don’t know why, that the CFR of all infected cases is going to be consistent against all demographics – that is, that after the most vulnerable die, the numbers of those dying will continue to stay the same.

        But, your estimate is now 302,000 people. Tell you what Ron, if we can anywhere even remotely near that number, the next time Reason does it’s annual begathon, I will donate $1,000.00.

        But, if we don’t, you have to change your Reason account name to “Ron Burgundy, Panic Merchant.”


        1. Lol. I’ll throw in ten bucks for this.

          1. Never going to let Ron forget how bad this whole debacle is making him look

            1. $: I won’t let you forget it either.

              1. How bad it made you look? Ok.

                1. That’s funny.

                  1. Does this mean you’re taking Geraje’s bet?

                    1. ¯\_(ツ)_/¯

                  2. Ron.. you should take up the bet but limit the name change to 1 (work) week with at least one use a day.
                    I may not agree with your positions on some things but have great respect for your continued engagement of this group… its a tough crowd. Plus, like anything else – and regardless of the motivations of others – a public acknowledgement of error is not humiliation, it is humility. I think others could respect that as well… and hopefully it could help shape your approaches to issues [IF you end up losing, of course]


          Are you still confident? We are not at peak yet, and Corona is already the leading cause of death in the US.

          1. From the article:

            “ Although the virus has taken the leading spot for fatalities in the nation, data reports show localized case numbers are diminishing in some of the heaviest hit areas. New York Gov. Andrew Cuomo (D) reported Thursday that hospitalization rates are lowering in the city for now.

            Anthony Fauci, a leading member of President Trump’s coronavirus task force, also said Thursday that early projections estimating between 100,000 to 240,000 fatalities from the virus are now less likely to occur.

            Fauci and other members of the task force are following newer and more data-driven models such as one from the University of Washington, which now projects 60,000 fatalities by August, a stark decrease from early estimations.”

            What was your point again ?

            1. That he is dumb as fuck.

            2. Fauci and other members of the task force are following newer and more data-driven models such as one from the University of Washington, which now projects 60,000 fatalities by August, a stark decrease from early estimations.”

              Just a reminder, that “data-driven model” at UW is from IHME, which was short on its hospital resource requirements for the entire nation by 50% as of 4 April, and some states by as much as 84%, “updated” its projections, and memory-holed its previous predictions so their current model curves are a lot lower.

              They’ve been using these from the beginning, which is why the “HOLY SHIT MILLIONS ARE GOING TO DIE AND WE DON’T HAVE ENOUGH HOSPITAL RESOURCES TO TAKE CARE OF EVERYONE THAT WILL GET SICK” doomer-tardation compelled our nation’s leaders and state governors to slit the economy’s throat, with the gleeful assent of the nation’s population of Karens and soy vacuums.

              Just as a reminder, that 60,000 is the equivalent of a bad flu season, and when compared to previous flu epidemics like the one in 1968-69, is not nearly as bad given the far higher population numbers the country now has.

              1. Yeah now this is somehow supposed to be about saving everybody and getting down to 0 cases. What the fuck? Weren’t we just flattening the curve to not overload the system, preventing the EXCESS deaths we would see in such a scenario?

                1. That would have to admit, in the long run, there is no way to prevent this from infecting a large majority of the population.

          2. Holy fucking shit you’re dumb. You’re comparing a peak rate to an average rate over a year.

          3. Per the CDC, COVID-19 is _not_ the leading cause of US deaths:

        3. Ron, now that it’s been shown that 0.37% is LESS than the 2.5% of Spanish flu, are you going to walk back your article “As More Death Data Becomes Available, COVID-19 Looks Less and Less Like the Flu” or at least put an amendment that yes, it does look a lot like the flu after all?

      2. Depends how you record the deaths.
        Currently, anyone who dies who has (or might have) the covid is being attributed to “covid deaths” regardless of the actual cause.
        It’s stat padding.
        But let’s cover every flu season like this, and let our technocrat overlords dictate every facet of our life while we cower in fear

        1. Further, what percentage of infected people are confirmed? How many are unrecorded?
          The technocrats have completely fucked the people of the world based on a CFR that 1) attributes every death of an infected person to cover, even if they were hit by a bus, and 2) has no clue how many people are actually infected.
          And we get… 0.4%

          1. Yeah, I could (maybe) see a whole different category for “suspected” cases. But by lumping all the stats together, all you’ve done is made ALL your numbers unreliable.

        2. Even with the stat padding, which is obvious, do you think we are going to hit 300,000 dead, or anywhere even near that number?

          1. Lol
            It’s fucking idiotic

          2. Geraje….The IHME model says no, we won’t hit 300K deaths. Merely 60K-ish. With extreme mitigation measures in place.

            The Germany data is an indicator. Not sure I trust their data 100%, without knowing a lot more detail. The good news is that it isn’t 3.5% or even 1%. I have not been shy about expressing my extreme concern with the Wuhan coronavirus. I still believe The KungFlu is far more dangerous than influenza. We’ll see. If 2-3 more large metro areas else go logarithmic, we’re talking a different ball game here.

            You’re being a little rough on Mr. Bailey, who at least is bringing real data to the table, and not mindless pablum. Ok, maybe just a tiny smidgen of pablum. 🙂

            Yo, I know you’re in NYC. Stay safe, Ok? My friends tell me Crown Heights is bad.

            1. I guess if you define “real data” as “incomplete, misleading, and chock full of unwarranted, non-scientific assumptions,” then I suppose we can agree that Bailey is bringing real data to the table.

              I’m being rough on him because he’s supposed to be a journalist. He’s supposed to be a science guy, but I never see any bona fide scientific discussions in any of his articles, let alone libertarian ones.

              And now, he’s blasting out absurd numbers that even the official panic merchants in government aren’t blasting out. Why? To prove a point on the internet? To demonstrate that he is willing to stand by his absurdities, no matter what? Or, to stoke panic and generate clicks?

              Anyway, I think Bailey is a hack. I’ll stay safe. It ain’t so bad here. There’s 10 million people living in this city. The hospitals are fine. There’s enough room for people. Lots of hysteria in the media that does not reflect the reality on the ground.

            2. Ronnie just offered fake data in the comment thread to push a panic.

        3. You forgot your source. Sources I have read have noted the huge spike in people dying at home in NYC, and their deaths are not being recorded as corona. So much so that they will start to count some of those deaths as corona going forward, but haven’t done so yet.

          1. If people are dying in their homes of coronavirus, so much fucking good the lockdown did.

            1. Same thing apparently happened in Italy, too.
              “We need to shelter in place so we flatten the curve! Oh, we did that and people died anyway? Shows we should have bolted everyone in their house like that smart Chinese government did!”

          2. Did you consider that people with health problems are afraid to go see their doctors and arent being treated for their non covid ailments dummy?

            1. Did you consider that many of those people are dying of covid and not being counted as covid deaths, though, and the whole point that I’m refuting is that covid deaths are being over counted?

      3. “My conservative calculation is that 300,000 Americans will be killed by the covid even though it’s completely illogical”
        -tranny Ron

        1. lmao

        2. I’m afraid to ask but, is Ron a tranny?

          1. “Transhumanist”
            I’m not a fan of that particular cult

          2. I know great platform Transen in München where Ron spent a lot of his time… So…

      4. Shorter Ron “I am still desperate to shit my pants over this “

      5. According to your own article, the German infection rate was only 15%. On what basis do you almost-double that to 25% for the calculation of US population? Using both the German observed rates, the US prediction equals 183,150.

        And while you are correct that last year’s flu deaths were about 34,000, you should also note that the 2018-2019 season was lower than average. Flu during the year before (the 2017-2018 season) killed over 56,000 and even that is not the highest recorded. The year-to-year variation is large.

        1. R: The answer is that epidemic is not over. Those numbers are useful for calculating the infection fatality rate and have nothing to do with what the ultimate percent of the population will become infected is.

          1. So you just randomly picked it, and ignored the actual data, got it.

          2. Don’t weasel out now, Ron. It’s going to be 25% of the U.S. population and that is your “conservative” estimate. It could very well be 50%, if we’re going to be pulling numbers out of our ass.

            By the way, do you have any timeline for when the entire country is going to get infected, and more than a quarter million will die? Because I don’t want you coming back in 10 years with an “I told you so.”

            1. GG: Citing data from past pandemics is using “actual data,” not randomly picking anything.

              1. Well, the Spanish flu had a CFR 10% – 15%. But, you are ignoring that. If you’re going to be randomly picking and choosing which stats to carry over in a comparison, and which to ignore, you’re not engaging in scientific analysis. You are pulling numbers out of your ass.

                So, would you care to propose a timeline? And, while you are at it, explain why your estimate of deaths is five times larger than the CDC’s?

                1. Why do you think the Spanish flu CFR has any bearing on the Coronavirus CFR?

                  1. It doesn’t … and neither does the fact that it apparently infected 25% of the population mean a damn thing when it comes to COVID.

                    Bailey is saying he is using “actual data” from a flu epidemic a hundred years ago to inform his outlook on the CFR and infection rate of a completely unrelated virus he insists is nothing like the flu.

                    Like I said, if we’re pulling numbers out of our ass, anybody’s guess is good as gold.

                    1. You’re wrong, but damn if you aren’t enthusiastic about it. How about you drop a single citation backing up anything you say. Every fucking person working in hospital or morgue in NYC right now is posting shit about how overwhelmed they are, but according to you, it’s all overblown. So show us, don’t tell us.

                    2. Try to follow the conversation, you worthless fuck. We’re discussing Bailey’s unsupported numbers for infection rates and projected CFR, which is several orders of magnitude higher than even the most hysterical official figures we are seeing.

                      How about you post a fucking citation about “every fucking person working a hospital or morgue in NYC right now,” asshole.

                      Fuck off and die, faggot.

                    3. Dollars comes I swinging his retarded chicken brain around. Holy fuck are you dumb.

                2. The CDC is only looking at April and early May. THIS wave. Same with the IHME.

                  Anyone who thinks this will be over with this wave is simply delusional. The vaccine is at least a year away. Until then (or herd immunity is reached), we are going to be having waves and playing whack-a-mole because – the vast majority of the population is still vulnerable to getting infected and the virus ain’t going on vacation.

                  1. Which means anyone who supports lockdowns until we have “essentially no new cases, no deaths,” is a fucking moron.

                    1. Either a moron or a strawman

                    2. Fauci actually said it, so moron.

                  2. And considering IHME’s plunk-and-crank model is already shown to be fraudulent trash, there’s no reason to take anything they say seriously at this point.

                    1. Red, I’d like to respond to that = IHME’s plunk-and-crank model
                      I do not know of any model that perfectly replicates ‘real life’. FFS, Wall Street hires the absolute best and brightest and they do not even get the direction of treasury interest rates right more than half the time (let alone the rate itself, which they almost never get right).

                      Models tend to be linear. Meaning, nice orderly straight line projections with definite trend breaks. Guess what? Real life is not linear. Real life is non-linear. We say, ‘life has its ups and downs’ and that is 100% true. Linear models are a very imperfect projection of a non-linear world. This is the problem we have.

                      I don’t think the IHME model is fraudulent trash. First, much of the initial projection data was based on non-US data, which I think we can agree is suspect. Second, as we started ramping up testing, we added our more reliable US data, and the model adjusted. Third, the model parameters changed in response to what we saw in our non-linear world (real life experience). This last point I think is a very fair criticism – parameters are a big deal. That tells me the model developers really did not think through implications at the model development stage. Fourth, we just don’t have practical experience modeling global pandemics. It is a total unknown.

                      The IHME model will improve over time.

      6. Maybe I’m a cold hearted bastard, but balanced against fucking up everyone’s lives for probably years to come, I don’t think it’s an unacceptable number of deaths. It sucks. I don’t like it. But it’s a natural disaster. Shit happens and you can’t always fix every problem.

        1. Z: Thank you for taking cost-benefit seriously. The problem is that epidemics take place in open-access health commons. As a libertarian my go-to response for overuse in an open-access commons is to assign property rights and let the rights owners, who now bear the costs and earn the benefits, protect the resources for the rest of us. I have not been able to think of a way to do this in an open-access health commons and I have done considerable research to see if anyone else has. So far no luck. If you or anyone can point in the direction of proposals on how to enclose an open-access health commons, please send me a link at my Reason email.

          1. I get your point here, but what if there is no solution? Is it not possible that this the absolute best that can be done, and that the world simply tosses curveballs sometimes?

            1. M: Except that non-pharmaceutical interventions to blunt epidemics in the open-access health commons have been used before successfully. The question is: Are the costs worth benefits? Check out the reporting my Reason colleague Jacob Sullum who has done great work on this topic.

          2. Please define the term = open-access health commons

        2. The thing is, those people were dead LONG ago, as soon as this bug got out of Wuhan. Flattening the curve isn’t about saving EVERYBODY, it’s about preventing the EXCESS deaths that an overloaded system would cause.

          The game has subtly shifted to “save everybody”, which is impossible no matter when we get permission to go outside.

      7. GG: The 25 percent infection rate is derived from what it was in US during the 1957-58 flu pandemic and the Spanish flu pandemic. Seems like a conservative estimate given that the coronavirus is invading an immunologically naive population. With respect to your bet, the number of deaths in the US will be much lower – no thanks to you – because of the social distancing policies that were adopted.

        1. “GG: The 25 percent infection rate is derived from what it was in US during the 1957-58 flu pandemic and the Spanish flu pandemic”

          So completely unrelated.

          “Seems like a conservative estimate ”



        2. GG: The 25 percent infection rate is derived from what it was in US during the 1957-58 flu pandemic and the Spanish flu pandemic. Seems like a conservative estimate

          As More Death Data Becomes Available, COVID-19 Looks Less and Less Like the Flu
          Death data from New York State demonstrates a stark difference between the two contagious viruses

          RONALD BAILEY | 4.9.2020 12:20 PM

          1. “The data says it’s not like the flu, except when I need it to be to make up numbers” -more honest Ron Bailey

            1. Thank you.

              Bailey is a hack, and the more he tried to defend himself against the charge of being one, the more of a hack he reveals himself to be. The man has absolutely no shame.

            2. $: I will explain again. You need to distinguish COVID’s lethality (the article you cited) from how many people (the proportion of the population) it will eventually infect. To get a handle on the latter, the attack rates of previous epidemics respiratory illnesses like the flu are probative. It’s really not that hard to understand.

              1. Why are they probative? Are all pathogens that infect the respiratory system presumed to be equal in terms of infection rates? MERS and SARS didn’t spread widely … why not? There was no lockdown. Does the ease of transmission mean that COVID is not as fatal as believed? Or, is it spreading in some heretofore unidentified manner?

                There are so many variables involved between different pathogens that your perfunctory approach to comparing them is really coming off as wildly unscientific. You *are* just picking random numbers, Ron.

                1. Indeed, whatever is convenient at each singular moment

                2. MERS and SARS didn’t spread widely … why not?

                  Because both of them only shed virus (were contagious) well after symptoms got serious. Which meant that it spread widely inside hospitals among medical staff – and very quickly only those staff spread it outside hospitals. Once they were all wearing PPE and tested, it ceased to be effectively contagious.

              2. Why did you choose two of the worst infection rates and not the ones with lesser populations infected? You cherry picked two bad ones and then called it conservative. You’re a hack.

              3. No Ron

                I will explain again

                You’re human garbage.

        3. No thanks to me? What do you think I’m doing? Running around town licking strangers?

          Next question science guy: How do you actually prove that social distancing lowers the rate of infection? Because, as far as I can tell, I’ve been cooped up in my house for nearly a month and the infection and death rates in NYC keep growing.

          Pray tell, wise sage, why is this so?

          1. I also live in NYC. I have some pretty strong doubts that the lockdown is actually doing anything. “Essential” stores around me have reduced their hours. Combining that with the fact that the only reason people go out is for groceries, that in NYC most people don’t buy a ton of groceries to last a month (not driving to the store), I think there’s a chance you spread the disease even more.

            1. I was saying this in another thread. There are about 6 major supermarkets within a mile of where I live, always packed with people. Some are doing serious grocery shopping, others are running in for snacks, beer, and ice cream. But the stores are always packed with people.

              I’ll tell you for sure that nobody is “social distancing” in the cereal isle, or on line while they’re waiting to pay. Indeed, since coffee shops, bars, and other grab and go establishments are shuttered, there are more people in these supermarkets now than ever before.

              So, NYC is on lockdown; nobody can go to work. But, jamming hundreds of people into a supermarkets, day after day after day, is perfectly okay, apparently.

              We’ve simply shifted the scenery, and now people are getting infected in supermarkets rather than subway cars. It’s an absurd joke but questioning it is now considered to be beyond the pale.

              1. It’s all theater

            2. Have the grocery stores put in things like plastic shields or have staff wearing masks or enforce physical distancing of customers? It took about a week post-closure here in CO for that to happen (and they still haven’t really given medical staff the supplies they need to protect themselves or to protect from cross-infection in hospitals) – but that marks the point where there will be flattening.

              I agree a lockdown itself doesn’t do much of anything until/unless those places that remain open become a better anti-contagion environment.

              1. One of the grocery stores here is sanitizing the check-out conveyor belt between customers, supposedly to protect them. Think of how often you actually come in contact with the belt while putting groceries on it. For me, that’s a rarity.

                However, the same grocery store does _not_ similarly disinfect the keypad or any other part of the credit/debit terminal, the one everyone touches to answer a couple of yes/no questions and, if they’re using a debit card, enter a PIN. The same is true for the local Lowe’s, which recently added improvised plastic shields at the check-outs, but everyone still touches the credit/debit terminal. Yes, about 50% of shoppers, by my own casual observation, now have gloves on as of last weekend (the last time I was out of the house). I watched one of them touch the terminal, then scratch their nose – but, hey, the glove will protect them…

                1. the same grocery store does _not_ similarly disinfect the keypad or any other part of the credit/debit terminal, the one everyone touches to answer a couple of yes/no questions and, if they’re using a debit card, enter a PIN.

                  That might be a good place to have – like you know – hand sanitizer. Apparently however they decided to get rid of their inventory of that rather than use some portion of that where it is actually most effective. There’s not really a lot of learning going on from what I can see.

          2. No thanks to me? What do you think I’m doing? Running around town licking strangers?

            Next question science guy: How do you actually prove that social distancing lowers the rate of infection? Because, as far as I can tell, I’ve been cooped up in my house for nearly a month and the infection and death rates in NYC keep growing.

            Pray tell, wise sage, why is this so?

            “We have always been at war with COVID-19.”

        4. Source for that infection rate? You link to your own earlier article, which has no citation or link for that number.

          Also, the ’57 pandemic was only immunologically naive for people under 67 (admittedly a smaller group back then). People older than that had partial immunity from earlier exposure in 1889 – which just happens to be the people most at risk to these sorts of illnesses.

        5. How the fuck are two of the worst flu I fection rates conservative instead of pessimistic?

      8. Don’t forget the guy who had a heart attack, was shot in the head and died from covid-19. Half of your hypothetical number of deaths will be > 70 years old.

        1. I: You will be happy to hear that I have been doing considerable research on what turns out to be the really complicated problem of assigning the cause of death in each patient’s case. Will report my findings soon. But in the meantime consider that when someone is at death’s door (especially for older folks), it comes down to a question of deciding just exactly which health insult pushed them through it. For example, in decades past pneumonia was sometimes referred to as “the old man’s friend” since it was a relatively peaceful way to die of cancer, congestive heart failure, etc. In any case, the CDC’s 2003 revised manual for listing cause of death gives an example of just how complicated these decisions can be.

          From CDC manual:
          Multiple conditions and sequences of conditions resulting in death are common, particularly among the elderly. When there are two or more possible sequences resulting in death, or if two conditions seem to have added together, choose and report in Part I the sequence thought to have had the greatest impact. Other conditions or conditions from the other sequence(s) should be reported in Part II. For example, in the case of a diabetic male with chronic ischemic heart disease who dies from pneumo­nia, his certifying physician must choose the sequence of conditions that
          had the greatest impact and report this sequence in Part I. One possible sequence that the certifier might report would be pneumonia due to dia­betes mellitus in Part I with chronic ischemic heart disease reported in Part II. Another possibility would be pneumonia due to the chronic is­chemic heart disease entered in Part I with diabetes mellitus reported in Part II. Or the certifier might consider the pneumonia to be due to the ischemic heart disease that was due to the diabetes mellitus and report this entire sequence in Part I. Because these three different possibilities would be coded very differently, it is important for the certifying physician to decide which sequence most accurately describes the conditions caus­ing death.

          1. Everything is a COVID-19 death. If you have doubts, resolve in favor of it being a COVID death.


            If this was truly as deadly as the panic merchants are making it out to be, there would be no ambiguity. That they have to fudge the numbers this badly gives away the game.

            1. And as we saw with the 13-year-old yesterday, even deaths without coronavirus!

              1. CO has recorded zero deaths attributed to covid19 to anyone under 20

                Course I’m sure Fox News or Facebook has far better data sources

                1. No one would be surprised if Fox were wrong about this, but according to the graphs at your link, “Data reported daily around 4pm, subsequent days data may show differences as cases are reported after this time”. Give it a couple of days. Also, as this specific case has received national attention, it might be reported differently than if it were just another needle in the haystack.

          2. Some links you might find useful and maybe already know about:

            CDC death stats are not timely

      9. 300k is a little higher than I’d like, but not ‘scary’ territory yet. It’s also almost certainly an overestimate, considering the 220k deaths prediction which assumed *no action* was based on an IFR greater than 1%.

        Have you bothered to calculate how many deaths from suicide or stress-induced medical conditions the lockdown is likely to cause? How would you like to incorporate purely economic harms of unemployment greater than the great depression? How many lost QALYs do those deaths likely represent? Just one number isn’t enough to make a utilitarian judgement on, you need to know what the alternative world looks like.

        1. Those are good questions. Bailey is not qualified to answer any of them.

        2. S: Check out the reporting of my colleague Jacob Sullum – he has looked at some of these issues.

          1. I’ve read it. But to make good policy, we need a comprehensive analysis that incorporates all this stuff, projects deaths and other consequences in both worlds, and can make an informed judgement between them.

            You being somewhat alarmist over 300k deaths suggests that, even if Mr. Sullum has done some analysis, you haven’t even tried to incorporate that work into your own thinking (much less the more expansive analysis needed). Also, I don’t want to see Mr. Sullum’s conception of these costs, I’m asking about yours – I want to know what you’re weighing 300k deaths against to come to the conclusion that such an outcome should be avoided with such draconian measures.

      10. Once assumptive numbers you pulled put of your ass. Let’s wreck the economy uneducated guesses!

      11. We would AT LEAST have to break the population down by age. Not everybody has the same amount of risk.

      12. 305,250 deaths
        x30? hospitalizations
        x5? ICU
        x?? Long term lung damage

        1. should be ~x2 ICU (half of critical cases die), and x5-10 hospitalizations.

          Also, those aren’t spike numbers, those are total pandemic numbers. There will certainly be a second wave, after all – the ’57 pandemic flu came back in ’58 and ’59, and the infection percent and deaths are for its entire duration, not just ’57. If we’re saying that 25% of the population infected number is a model, then it’s the sum total of the pandemic duration, not the first wave.

    2. Leave JFree alone. He still thinks he is right.

  5. But they also believe these findings show that lockdowns can begin to be lifted, as long as people maintain high levels of hygiene to keep COVID-19 under control.

    I’m not sure I would count on the people who invented scheisse porn to maintain high levels of hygiene but for those of us who aren’t German this is god news. Can our governments stop being stupid now? Rhetorical question, I know answer is “No. Not now, not ever.”

    1. What part of “this is an election year” did you not understand?

    2. “scheisse porn”

      I’m unfamiliar with this term, and I refuse to look it up. Please elaborate.

      1. Porn for SQRLSY.

      2. That’s German for defecation. And while there are some strange people into that particular fetish, there is no evidence that it originated with or is any more or less prevalent among the Germans than everywhere else.

        1. But it’s so much fun to imagine that that’s what buttoned down Germans like to do with their spare time.

        2. Another hun Scheiße porn denier…

          1. Unless I see some good physical evidence, I’m not going to believe it originated in Germany.

            1. are you *sure* you want to *see* that?

            2. Literally no one cares.

        3. The cannibalism thing is backed by hard evidence though.

        4. Germany is the main place where most toilets have a shit inspection shelf

    3. They need those high levels of hygiene exactly because they invented scheisse porn.

  6. By the way, I’m still waiting for JSlave and all the other hysterical ninnies around here who were initially saying millions of people were going to die to have some balls and admit that they were way off.

    1. Ron revised his estimate to a mere 300k…

    2. I never said millions. That is the worst case scenario. I estimated roughly a million – roughly 2/3 over 65 and 1/3 under-65. Hell – I even estimated fatalities by political affiliation in my city – 1.1% of R’s; 0.9% of D’s; 0.5% of independents; 0.3% of L’s. So I see a lot of positives with a high death count here.

      When we either have roughly herd immunity or a vaccine, that is when we can tally up and see who’s right or wrong. Until then, fuck off.

      1. When we either have roughly herd immunity or a vaccine, that is when we can tally up and see who’s right or wrong.

        Kind of hard to get herd immunity when you put your populace on house arrest.

      2. No we can tally you as wrong right now.

        1. “No we can tally you as wrong right now.”

          Pretty easy to do a couple of weeks ago.
          JFree’s a lefty imbecile and screams about MASS DEATHS!!! are a pretty good indicator that there’s a loose connection someplace.

      3. The average age of death is over 80. And because the death rate skews old, that means more than 50% of the deaths are older than 80, probably significantly more. There’s no way 1/3 of deaths end up being under 65 unless that average shifts dramatically. When you take an average of a non-linear function, the average is not the median – relative outliers (the few deaths under 40 in this case) exert undue pull on the average, so to maintain an average that close to the end of a limited range (there’s not many people over 100), you need a lot of datapoints on the other side to counter-act that pull.

        I’d make a rough guess at 2/3 of deaths are over the average, at least half the remainder (1/6) are over 65, and at least half of that remainder are over 50 (1/12). And that’s probably still overestimating the fraction under 50 that die, but some quick napkin math suggests its in the plausible ballpark if the average age of death ends up between 75 and 80. (If it’s 85, like some data i’ve seen, those fractions need to skew older yet).

        1. My estimate was based on 2nd wave exceeding medical capacity everywhere very early. Within the first few days. That wave is not about a disease starting in one place halfway around the world. It’s about starting nearly everywhere at the same time – based on the weather getting a bit colder so people start sneezing/coughing and being indoors more.

          Once you exceed capacity, fatalities are more about who needs hospitalization not who benefits the most/least from treatment in hospital. Those who need hospitalization and can’t get it cuz they’re all full become fatalities. That hospitalization rate is far less age-skewed – and obviously the % of peeps is much much younger.

          1. JFree
            April.11.2020 at 2:12 pm
            “My estimate was based on 2nd wave exceeding medical capacity everywhere very early….”

            So more bullshit guesses from the cowardly piece of shit who has been 100% WRONG.
            Stuff your PANIC!!! flag up your ass, stick-first.

          2. “Once you exceed capacity, fatalities are more about who needs hospitalization not who benefits the most/least from treatment in hospital. Those who need hospitalization and can’t get it cuz they’re all full become fatalities.”

            This is a seriously flawed argument for the current solution. I have looked at every study I can find, and the prognosis for severe cases is very, very poor. So far, survival of just 33% of those on ventilators is the norm. And if you look at longer-term survival of non-covid patients on ventilators, it is also very poor (over 60, 2/3 die within a year). Which means that as things stand now, 10% might survive the year, and even those might have permanently impaired lung function.

            Which means that much of what we have witnessed to date has been ventilator theater. Doctors are used to doing everything they can for seriously ill patients. They are used to throwing massive teams at patients who arrest on their watch. The medical system is in full freak out over not being able to apply their current standard of care to seriously ill COVID patients.

            The only sound argument for the current extreme lockdown is to buy time for an effective therapeutic. With each passing day, the lockdown becomes much harder on public health than the virus.

  7. Did you not include the ’68 pandemic flu in the IFR comparison for a reason? It had an IFR of 0.5%, higher than covid.

    1. And likely to have comparable number of deaths as well. I would be shocked if total deaths from covid-19 passes the Hong Kong flu totals.

  8. What is the hospitalization rate for Covid vs the typical flu?

    1. Roughly 500K hospitalizations on the average (a lot of variance though), with 25MM average infected (also a lot of variance). Comes to ~2%.

      Covid-19? It is a crapshoot. No one really knows – yet. If you assume 50% are asymptomatic, you can do some very rough math. Assuming 50% are symptomatic, then 20% of that number (10%) overall would require hospitalization. The other big ‘but’ is that 20% rate is from China, one of their studies. I can’t really trust their numbers.

      1. The flu infected number by CDC is all symptomatic. They no longer include the asymptomatic infected in public data. When they did it ranged from 80-100% of the symptomatic number – so 40-50% of the total infected. Roughly same range as covid19.

        The covid19 ‘confirmed case’ numbers are more equivalent to the ‘doctors visits’ subset of flu infected. There’s not the same history of testing antibodies and developing vaccines for following year

        And CO for one does include a hospitalization % – hospitalizations/confirmed cases – breakout by age for covid19.
        4.5% for 20’s
        9% for 30’s
        14% for 40’s
        20% for 50’s
        27% for 60’s
        34% for 70’s
        26% for 80’s (lower because deaths are very fast for this group)

        1. If you want to compare those above numbers roughly to 2017-2018 (a bad year) flu season – hospitalizations/doctors visits:
          18-49 age – 1.5%
          50-64 age – 2.5%
          over 65 age – 16%

          covid19 is actually far more relatively serious for the younger groups. For the older group everything is serious and covid19 doesn’t stand out as much

      2. One of the doctors characterizing the china data suggested 84% never got diagnosed (asymptomatic or mild symptoms such that they never sought medical help). The 20% hospitalized is only of those diagnosed, not those with any symptoms. So it could very well be 20% of much less than 50%.

        1. Squirreloid, you’re right. And I’ve been pretty upfront saying the data outside the US is very suspect. Even EU data is not so good. China data is just crap. Japan and SK present different issues (mostly due to test reliability). Our data set is getting there, and I am getting more concerned, government happy-talk aside.

          In the People’s Republic of NJ, Phailing Phil Murphy is demonstrating that Goldman Sachs rejects cannot manage their way out of jack-shit. Utterly terrible performance. Absolutely TERRIBLE at rolling out testing. Phil phucked up very badly.

          At this stage, we need to get that finger-prick serologic testing out there is a real hurry. I think we all know it ain’t just flu. It’s serious, and yes…we need to open up our economy soon. When? My best guess is Independence day. Why? I don’t see evidence of a steep downslope in cases like the models project. I still cannot believe I am saying that…basically throw out Q2 from the economic history books. Incredible. Could you even imagine something like this?!

          1. I think it’s more likely Memorial Day. I agree there isn’t any real downslope. There is the appearance of a downslope but that is because a lot of states have plateaued in testing once they caught up and locked down. That is actually more of a lag. Downslope can’t happen until public health actually goes out and finds contacts of positives and offers those people a way to isolate without necessarily infecting their family or bankrupting them. I’m not sure we are capable of doing that here. Actual public health is like 500th in line at the trough.

            But there’s no way they can keep throwing around $1 trillion in new debt every few weeks – and there is transparently no interest in mobilizing the newly unemployed (like the militia) to deal with the laundry list of stuff that could be done to actually help deal with the virus. So we’re going to pretend the problem away in early/mid May – and hope that there is a summer slowdown in contagion.

          2. Data from Japan and S. Korea are qualitatively different.
            The Japanese data should a steady exponential rise in infections and a variable death rate near 1% that does not account for 1) the time between infection and death and 2) the number of uncounted infections.
            Korea’s infections grew at the same rate as Italy as reckoned from the date of first cluster of community infections. After two weeks, Korea imposed a heavy regime of tracking and tracing (based on 0.4% of the population being tested.) That regime transformed spread to a slow linear growth in the number of infected persons.

  9. If the average IFR in Gangelt is less than 0.4%, how does that break down by age groups and preexisting conditions? And how do those numbers compare to typical influenza?

  10. I want to point out that people should take these IFR studies with a grain of salt. There are multiple reasons as to the reason why. First I look at the China study which Ferguson did and came up with an IFR of about 0.66%. Now while this may indeed be correct, the Chinese government freely admits that quite a few people that died were not included in the overall death count. If that is the case than the IFR would be higher You end up with a numerator v.s. denominator issue in which the numerator is far more sensitive than the denominator. For example 2/100 is much different than 1/101. The death totals are far more sensitive than the case totals. In the US right now it’s apparent that they are missing deaths even though they are trying their very best to account for all of them.

    Now on the German study, the population being studied is so small. What are the co-morbidies? I know southern Germany typically has a longer longevity than the rest of the country. The population in this area may be much more healthy than the typical population. Did a few die that wasn’t accounted for? Because in a sample that small, it would make a large difference in the overall outcome. These early IFR studies are informative, but they are not definitive in the least bit. Some of the statistical data such as that from the CEBM assumes the IFR to be below .2%, but they also assume and R0 of around 3.5(and a doubling rate of about 1.8 days). This greatly increases the denominator. This would also mean that it’s going to spread to spread more which would increase the death toll.

    Calculating IFR is a very complicated issue, and I think most studies right now have a bit of their math wrong using faulty assumptions. It’s still far too early in the process to same for certain what is going on.

    1. On the other hand, they aren’t actually doing the work to attribute the deaths they do count to covid-19. Few autopsies are actually being done. And they’re patently counting people who died from things other than covid-19, just because they tested positive.

      There are biases in both directions on the covid-19 death counts.

  11. “For comparison, the U.S. infection fatality rates for the 1957–58 flu epidemic was around 0.27 percent”

    Because there was widespread testing back then?

    What those trying to downplay the mortality rate of COVID-19 conveniently pretend to forget is that many with the “seasonal flu” may also be “asymptomatic cases or undiagnosed people who recover at home.”

    On a side note: Notice how Conservatives inflate the number of dead that Stalin sacrificed to turn a nation of peasants into a super power (in a lame attempt to whitewash the murders committed by alt-Right hero Hitler) – Yet want to emulate Stalin’s actions by sacrificing older workers in order to keep the Conservative Donor/Parasite class in comfort.

    1. “On a side note: Notice how Conservatives inflate the number of dead that Stalin sacrificed to turn a nation of peasants into a super power (in a lame attempt to whitewash the murders committed by alt-Right hero Hitler) – Yet want to emulate Stalin’s actions by sacrificing older workers in order to keep the Conservative Donor/Parasite class in comfort.”


  12. Were this study true. Leaders of nations have done close to criminal damage to drive the world into the deepest economic downtown in many decades with the burden falling most heavily on the working class.

  13. Believe this study at your peril. The apparent death rate in Germany has been rising steadily for the past 3 weeks.

    1. The apparent death rate is a meaningless number. Population testing or it’s just fear-mongering.

    2. “…The apparent death rate in Germany has been rising steadily for the past 3 weeks.”

      Your cite fell off, chicken little.

  14. I think we are going to have to put up with more deaths than we currently imagine so that the survivors can have a semblance of a decent post pandemic life. This is a tragedy that we are just going to have to live with for the next couple of years. Minimize deaths via protocols such as those enacted by the Taiwanese while getting back to the business of life.

    1. 1 million or so people dead in U.S. from COVID-19 vs 349 million spiraling into an economic, psychological, cultural, and social despair that will last decades or more.

      Why is this even a question?

  15. English translation of the study result summary is available here.

  16. It would be more noteworthy if they used those numbers as an opportunity to admit that everyone lost their shit over NOTHING.

  17. When comparing the German numbers to past pandemics, is it apples and apples? Did we test for antibodies and rates of infections for the Spanish Flu, Swine Flu, etc.

    PS: There is a similar antibody study in the works at Stanford University.

  18. Hi Ron. This might sound “truthery” but Those numbers for Spanish flu are wrong. That’s the approximate global death rate due to the pandemic, not the case or even infection fatality rate.

    Infection fatality rate looks to have ranged from 3-20 percent. The bad number comes from a WHO slide that cites a vaguely worded abstract. More info in the Wikipedia citations and if you look at pre-COVID WikipediaPage history.

  19. 1 million or so people dead in U.S. from COVID-19 vs 349 million spiraling into an economic, psychological, cultural, and social despair that will last decades or more.

  20. One microgram of fine particulates per cubic meter is associated with a 15 percent increase in the Covid-19 death rate.

    So to extrapolate 0.37% from Gangelt to somewhere else, you’d have to know the air quality as well. Average age and co-morbidities would be required as well.

  21. “But they also believe these findings show that lockdowns can begin to be lifted, as long as people maintain high levels of hygiene to keep COVID-19 under control.”

    I vote for this.

Please to post comments

Comments are closed.