As I, and many others, have pointed out for weeks, you can’t calculate an accurate death rate without knowing the denominator. What we are seeing is that there are many people who get the virus who show few, if any, symptoms – and many of those are never even being tested for the virus. In Germany’s care, they are testing more, finding more infected people, and thus their official death rate is lower. 0.5% is still pretty high, but it’s nowhere near the 8+% we have been hearing about in other countries.
Germany‘s rate of new coronavirus infections dropped from 21 per cent to 15 per cent today while the country’s death rate remained strikingly low.
Official figures from the Robert Koch health institute showed the total number of cases in Germany rising by 4,118, taking the total from 27,436 to 31,554.
The death toll climbed by 35, rising from 114 to 149 – an overall death rate of just under 0.5 per cent, far lower than Germany’s major European neighbours.
The low death rate has not been fully explained, but may be linked to more comprehensive testing, younger patients and more intensive care facilities.
Germany has a policy of ‘doing everything to find, isolate, test and treat every case’, unlike in the UK where people with mild symptoms are not routinely tested.
The problem is that nobody really knows why the German mortality rate is abnormally low, even compared to South Korea’s 1.4% mortality rate. South Korea has a similar “test everybody who came in contact with a positive case” regime, and I haven’t seen any reports saying that South Korea’s health care system has been overwhelmed. I suspect that it’s that, at least so far, the lion’s share of infections in Germany have been among those 35-59 (<a href=”https://corona.rki.de/”>source</a>). The mortality rate among infected Germans 80 and older is over 7% as <a href=”https://www.rki.de/DE/Content/InfAZ/N/Neuartiges_Coronavirus/Situationsberichte/2020-03-24-en.pdf?__blob=publicationFile”>79 of the 114 deaths as of yesterday were among that age group</a>.
The Daily Mail story did not directly mention the warnings from the Robert Koch Institute (Germany’s version of the CDC) carried in German media that they still expect cases to spike and that spike would likely push the overall mortality rate up past 1%.
Further, as of Monday, Germany banned gatherings of more than 2 non-family/non-household people, but, for the most part, have not ordered businesses shuttered.
The other issue is: People rarely die from COVID 19. Those that are, tend to have other underlying health complications, so it is not the COVID 19 itself.
It is inappropriate to blame COVID 19 itself for deaths when other competing health factors may be equally to blame. From my perspective, when taking that into account, we are panicking over a flu strain.
People with compromised immune issus and underlying health conditions suffer much greater risk from catching garden variety flu.
I’m going to enjoy my coupon for a free Dilly Bar of Death from Kevin when he realizes that he knew someone who died of COVID.
Says the racist and sexist baby killing advocate.
Remember, you have to buy a dilly without getting one if I am correct.
It seems silly to argue for/against the “denominator”. That number is never going to be known. One has to work with the only baseline available: confirmed infections. That has always been the case and is how historical influenza rates have been derived and described. So if you hear some pundit/mediawhore say, “The death rate is…” such and such without the qualifier of “of confirmed infections” (CI), you can simply ignore the numbers being given as irrelevant.
Speaking of irrelevant, just this morning WI-DHS, using divining rods, claimed WI is going to have 22,000 infections and 440 – 1,500 deaths two weeks from now; a psuedoCIDR of 2% to 6.81% (rather a large spread, wouldn’t you say?).
What is most appalling is that WI-DHS is absurdly basing their models on Wuhan and Italy.
Now, why do you suppose these WI g’vment ‘experts’ chose only these two specific entities – which are closely linked governmentally, finacially, and of course CCPVirus-wise – on which to base their guessing games? Why not mix it up? Why not Italy and Germany, or Wuhan and SKorea? Why not Germany and SKorea?
You know why.
Oh, yes. The WI-DHS can tell us the hammer of the Great Impending Doom is upon us, with every hospital overflowing with hoards of choking phlegm-ejecting victims of the CCPV… well, that virus. But they can’t tell us – right now – now how many have been hospitalized, how many have recovered, or the particulars of the victims. They are on it!
BTW the WI CCPVirus CIDR, as of 9pm yesterday, stands at 1.02%. 6 deaths/585 infections. Meanwhile, there were 92 deaths from other influenza bugs as of March 4th, 31243 infections, a 0.29% CIDR (the three children remaining inconsequential to our state ‘experts’).
And MjM, at what stage in the infection do the deaths happen? Now, or later? Now, or after someone’s been on a ventilator for a week or three?
The only reason people are cherry-picking denominator values is because they want to stop the shutdowns because they think the economy is more important than lives saved.
jjf, based on news reporting, there few, if any, murders in Chicago.
Should continue the martial law they have indefinitely?
It will save the lives of gangbangers and the innocent people they kill.
So, you would support the restrictions in Chicago indefinitely?
Flu A and -B denominators are determined by best-practice statistical inference, but that requires a very large base–which the Flu has, being in the 100’s of thousands annually in the USA.
We don’t have that large a base on the Chinese Clap yet, so valid statistical inferences cannot be drawn. That’s partly due to lack of testing, and partly due to -frankly- a very small (in comparison) # of cases.
Assuming test kits will be available, authorities can/should run tests on larger “sample” populations which will show anti-bodies; that will enable determination of the statistically-valid denominator.
No, Mar, these restrictions are in place temporarily to reduce the load on the hospitals, in order to save lives. And I know the government doesn’t like to tell you this, but I predict there will be waves of restrictions, region by region, over the course of the next year or so. That’s how suppression will work. If a vaccine or treatment arises, that can change the game. But I predict a few more waves of lockdown, falling infection, removal of the restrictions, and then as infection rises again, new restrictions. That’s what’s in the federal plan.
Meanwhile, BuzzKill…er…BuzzFeed is alleging the numerator (i.e. the number of deaths attributed to COVID-19) is too low.
As for Foust’s prediction, that is what has happened in Singapore and Hong Kong.
Daddio sez: Flu A and -B denominators are determined by best-practice statistical inference
The CDC compiles reported influenza cases, not chunks of randomized samplings. True, they make predictions. But they are, usually, extrapolations based on observable trends in the reported data. It’s also true that with statistics that you can say, “the average of x over the last five years is y”. Call me conservative, but it is utterly silly to make statistics-based global, let alone localized, predictions given 4 months worth of hardly impeccable data. See Owen’s newest post, for example
The other issue is, well, statistic-based predictions in general. Statistics represent past history, not the future observed. Can anyone say with certainty that Ryan Braun will recover in 2020 the .013 points lost in 2019 to his lifetime .298 batting average?
And so, while we will know confirmed cases no one will ever know the actual number of CCPV infections there are or will be (unless you test every human on the planet every day), and the statistics will always be be off by likewise-unknown degrees.
International travel is going to be problematic for quite awhile.
Seeing the USA in a Chevrolet might make a comeback. More recreational dollars staying at home. A week(end) in Bayfield certainly has started looking a whole lot more appealing.
That’s what I said, but you said it more elegantly. And yes, the flu has a larger sample–something else I suggested MUST be had to make any sense at all of the Chinese Clap.
This morning’s RETRACTION by the Brit at Imperial has the interesting nugget that there are FAR more cases than he estimated, making the denominator yuge.
Daddio claims: “…but you said it more elegantly”
Elegance? Heh! Now THAT is something no one has ever accused me of before! I’d buy ya a brew for the compliment but Herr nEvers says I can’t.
I understood what you wrote, but I’m not sure you understood me. CDC death rates are based on the reported infections. The CDC predictions I am talking about, the year-to-year, coming season ones, are based on a complete (as complete as can be) data set, not data picked from just, say, Billings, Topeka, and Houston. The data used is the previous year’s dataset ( and if I remember correctly, a mid-season prediction update comes based on current data mixed last year’s outcomes). This may make the short-term fortune telling slightly more accurate but it still has the problem I mentioned above that ALL predictions have: history. IOW, even the complete data set is, at the time of the prediction, a snapshot in time past. Consider the high dynamic nature of biology. Now add in the fact that the reported IA/B infections are just that. The ‘as complete as can be’ data set is STILL missing the probable millions NOT reported (meaning, confirmed).
Aaaaand we’re back to square one.
To me, the only relevant denominator to use for any of this would be population. Yes, it too is always in flux, but it’s the most encompassing number available.
As you can tell, I am not a fan of statistical Zoltarism. That it is far more often abused to proclaim definitive outcome rather than what it actually represents is why.
Statistical Zoltarism? Hah! Wait’ll you hear about economics and the stock market.
And around and around we go.
Not-good news – the mortality rate in Germany is starting to spike. Depending on whether one uses the 3/31 RKI numbers compiled at midnight local time this morning (61,913 cases, 583 deaths) or the current (as of this typing) Johns Hopkins numbers (71,690 cases, 774 deaths), the mortality rate is between 0.94% and 1.08%.
From the RKI reports, the median age of infected has also risen since the 24th from 47 to 48. While the number of cases among those 59 and under, both in aggregate and in the groups defined by the RKI, slightly more than doubled over that time, the number of cases among those 60 and older almost tripled. The 3/31 RKI report also notes that reports of outbreaks in nursing homes is increasing, with high death rates in those outbreaks.