Wednesday, October 20, 2021

Most Covid deaths are not due to Covid now

A simple calculation showing that Covid-delta is far less lethal than flu

One of the points that those of us who opposed the uncontrolled Covid hysteria emphasized from the beginning was the fact that
the people who got a positive PCR test were not necessarily ill.
In practice, what matters are the fractions. It is a quantitative question. Some of the Covid-positive people have been ill, some of those were hospitalized or they died. But some of the PCR tests are false positives and many truly infected people with a positive PCR test are asymptomatic (or their illness is so mild that it is not worth mentioning). A related statement was that
In most countries, a Covid death only means a "death with a positive PCR test" which doesn't mean that the death was fully or mostly caused by the Fauci virus.
Again, that is an important point to have in mind. But a year ago when the Czech daily casualties became substantial, I finally became certain that the distinction didn't matter in the 2020-2021 winter season. One can look at the total weekly numbers of deaths. In Czechia (10.7 million people), there were numerous weeks in which the normal weekly number of deaths (2100) was doubled (4000+). In fact, the 2020 excess deaths (17k) surpassed the number of Covid deaths by dozens of percent. Some of the other deaths could have been either due to the lockdown and a less inaccessible healthcare; or they could have been deaths that were secretly due to Covid, too.

But this situation ("Covid deaths are mostly real deaths due to Covid" and "the excess of deaths is even higher than the official Covid deaths") has profoundly changed when delta and derived variants have become the dominant versions of Covid. Four months ago I emphasized that Delta SARS-CoV-2 is strictly a common cold virus because it causes a running nose (this is true for all the derived variants, it is not useful for epidemiological purposes to distinguish them). A running nose is a cosmetic trait that makes the disease more infectious (you shoot droplets on others by your nose) but it also makes the disease far less serious because the nose's reaction helps you to reduce the dose in your body quickly, perhaps before it becomes substantial in the lungs.

When the total number of deaths was doubled for a week almost a year ago, the excess clearly couldn't have been explained through "deaths caused by something else than Covid" which "accidentally" occurred along with a positive test. The fraction of PCR-positive people was never anywhere close to 100% but even 100% would fail to be enough to explain the actual excess of deaths. But the ratio deaths/cases dramatically dropped relatively to the last winter season. This drop was both due to the less severe, cold-style delta variant; and because of better treatments that almost 100% work for classes of patients (Renegeron or Bamlanivimab). I have followed all the daily Czech numbers for more than a year and I believe that I remember them and understand their implications more than anyone else. Most of the lessons are obviously relevant outside Czechia, too.

Note that the people who pay attention to the "number of cases" have a new reason for hysteria because the autumn-related growth of the cases has approached a "tripling in 2 weeks". Yesterday, more than 3,000 PCR tests were positive. That is a big increase from the lows in the early summer which were around 100 positive PCR tests a day. Near these lows, we had as little as 0.1 Covid deaths a day. The most recent daily Covid deaths (which means deaths with a positive PCR test) are
4,5,7,6,6,10,8,10,10,1
The most recent 7-day average is around 8 Covid deaths a day. That is way below the peak 261 on November 3rd, 2020 (in March, we repeated numbers close to 250 again). However, what is interesting are the percentages of the PCR tests that are positive. On the statistical page, they may be found under the headline
Vývoj denního poměru (v %) počtu osob s nově prokázaným onemocněním a celkového počtu provedených testů podle indikace
In that section, you may choose a "15-day graph", "full time graph", or "tables"; and then you can choose according to the indication of (reason for) the tests, "diagnostic", "epidemiological", and "preventive". It is not hard to understand these three groups and it is very useful to have the 3 separate numbers. Diagnostic tests are those on the people with symptoms who want to know whether they are caused by Covid. Epidemiological tests are the "tracing" tests of the contacts of recently positively tested people. And "preventive" tests are tests of the people who have no reason to think that they have Covid (neither symptoms, nor contacts with positive people) but who would be at risk if they happened to be infected. It is a good idea to treat these people as early as possible which is why an early positive test may help.

Let me simplify these three percentages of positive PCR tests as "diag", "epid", and "prev".

Considering the 7-day averages, the fraction "diag" peaked near 50% in December 2020 (every other person with flu-like symptoms indeed had Covid!); in October 2020 and March 2021, near the other two big subwaves, it was close to 40%. "Epid" peaked at 25% in October 2020 (a quarter of people around the positive folks could have expected to become positive as well), was close to 20% in March 2021. And "prev" peaked near 20% in October 2020.

The fresh figures are 13%, 8.6%, 2.4% (two significant figures) while their newest 7-day averages are 9.0%, 5.0%, 1.3%.

If you have symptoms, you can directly refine the probability that you have Covid. The latest positive fraction of the "diag" tests is 13% so if you have Covid-like symptoms, you basically belong to this group and your probability of getting a positive PCR test (assuming that you will have one soon) may be estimated to be 13%. If you met someone with a positive test, your probability of getting a positive test (if you decide to have it) is 8.6%. These numbers are actually useful and directly answer pretty common questions that rational people may want to know! They are much more useful than the "number of positive tests in a day in a country" which depends on the intensity of testing and other social conventions that say nothing about the medical situation.

The last fraction, "prev", is even more useful for us because we want to understand the deaths. The viral load in the society is going up in the recent months and the fraction of positive preventive tests goes from 1.3% (a 7-day average) to 2.4% (the latest figure). Let us say that it is 2% in "these days" (it is the lowest number among the 3 because it comes from an ensemble of people who have no reason to think that they have an increased chance of having Covid). We mustn't forget that there is some error margin here. What does the number 2% tell us? It tells us a lot.

Just focus on the subset of the Czech nation that is at significant risk with Covid. Illuminate them with a big light bulb, forget about the rest. Great. We actually have a pretty good estimate of the fraction of these people who would get a positive PCR test today if all of them were tested today. It is those 2%. The fraction may be a bit different (lower or higher, we don't know) in the overall population but we don't really need that fraction (I will assume it is very close to those 2%, too, but that will only affect one minor result below). Most of the deaths are from the group of "people at risk", anyway, and that is why the percentage optimized for these people, for the people who "deserve a preventive test", is most accurate.

So the virus (overwhelmingly delta and derivatives) may also be PCR-found in 2% of the hospitalized people, 2% of ICU patients, and 2% of the people who died yesterday. For 16 weeks in a row, weeks 21-36, the total number of deaths in Czechia was below the normal (2015-2019 average). You may check that the deaths are below the normal even in a half of Europe right now. At any rate, we get about 300 deaths every day. If you assume that Covid is just spread randomly among these people (Covid doesn't deliberately avoid the people who have serious cancer or heart attack etc.), it is unavoidable that 2% of 300 i.e. 6 people a day are dying with Covid even if Covid doesn't contribute to the death at all!

In reality, we get 8 deaths with Covid a day or so these days. 6 is a majority of 8. There are error margins both in the numerator and the denominator. But I think it is more likely than not that fewer than 50% of the deaths with Covid are actually due to Covid now (for the first time; this "minority claim" didn't hold in the 2020-2021 cold season). The number of deaths caused by Covid every day may very well be 1-2 and we can't reliably rule out the result 0, either!

To compare the case fatality rate with that of flu, what is the fraction of the ill people who die because of the Covid virus? Well, it depends how you count the "ill" (how you count the "dead" people is more well-defined). If you defined everyone with a positive PCR test to be "ill", you would naively get 1-2 deaths due to Covid out of the 3,000+ positive PCR tests that are getting per day. So the CFT would be between 1/1,500 and 1/3,000. That is already lower than flu's 1/1,000.

However, the actual fraction of "deaths from Covid" over "people who would produce a positive PCR test" is even smaller than that. Why? Because most of the people who would get a positive PCR test (they are infected) don't actually go to that test while they are positive. In 14 days like yesterday (14 days because that is the time for which a PCR-positive person is considered positive, we don't end it by a negative test because that is often hard to get for months), the number of actually demonstrated PCR cases is comparable to 2,500*14 = 35,000. However, the current number of Czechs who would get a positive test is closer to the 2% of the nation i.e. 200,000 (here I extrapolated the "preventive" positive fraction). Something below 1/5 of the infected people are proven by a PCR test in Czechia.

That gives us a spectacularly different CFR. 200,000 people are positive today, that number may be divided to the 14 days which define the period when they are considered positive, and you get 14,000 people who would get a positive test every day if everyone were tested. Out of those, only 1-2 people die from Covid, as I have calculated above. So the actual CFR is actually close to 0.01% or 1/10,000. It is just 10% of the CFR commonly quoted for flu!

You could argue that the number of "people who are ill with flu" is counted differently during flu epidemics and the result is actually smaller than the number that would get a positive test. People are self-reported to have flu if they have the symptoms etc. But that can produce at most a factor of 5 because flu often leads to the expected symptoms and workers want to stay home when they have flu. Even if you reduce the number 10,000 from the previous paragraph by a factor of 5, you get a smaller CFR than flu's 1/1,000. While the error margins are substantial, the case fatality rate of Covid-delta is closer (on the log scale) to the cold than the flu, to say the least.

So Covid delta simply is less dangerous than flu! Right now, despite the growth from the early summer by 1.5 orders of magnitude, the number of active people is still just 150 per 100,000 or so, below the number 1,600-1,800 per 100,000 that normally defines a flu epidemics. So we are one order of magnitude below a flu epidemics threshold now; and the flu-like disease, Covid delta, is less dangerous than flu, too! Those are reasons why the non-standard and emergency policies (and the hysteria) are absolutely unjustifiable by the facts. Thank God, at least the 3 Germanic Scandinavian countries have understood that it is silly to "fight against Covid now" and they ended all restrictions. In fact, you don't even need face masks inside airplanes, during flights inside Scandinavia.

Meanwhile, tons of other countries controlled by hopeless unhinged crackpots who actually believe that Covid delta is an exceptionally serious disease; or evil people who know very well that it is not but who just want to exploit the fabricated fear for their personal benefits are leading their nations into a suicidal, war-like behavior. The situation is most shocking in Australia but countries like Lithuania, Slovenia, and even Italy etc. have gotten close to this Australian insanity. In practice, if whole nations may be brainwashed to believe that the situation deserves a state of emergency now, they may be brainwashed even when the numbers are going to be even smaller than today (even if it is by extra 3 orders of magnitude smaller!). Because the "number of cases" is almost certain not to go to zero in whole affected countries for a year or many years (and reasonably likely, never), these brainwashed nations may expect quite a long, dark future.