Center for Strategic Assessment and forecasts

Autonomous non-profit organization

Home / Politics and Geopolitics / / Articles
Coronavirus: how we deceive ourselves
Material posted: Publication date: 21-03-2020
I understand that the degree of hysteria in the background of the coronavirus caused society to a conflict between followers and deniers of the General panic. Sharp enough for group polarization did not cause any special desire to Express their point of view on the General public. But when my companion began to pursue conspiracy theories and mindless interpretation of the data, I realized that the comforting arguments are very few and decided to fill their deficit.

Distortion of mortality

I am sure for many will be a revelation the fact that the number of deaths with the presence of coronavirus recorded in the who is the number of deaths from the coronavirus.

What do you think? New to human infection is still not fully understood, but the hospital from the suburbs of some Livorno already know how to define death on her? To call the cause of death is not a bug when compiling to catch. It's much more difficult.

A final clinical diagnosis. The underlying disease may involve multiple nosological forms. There is the concept of competing diseases, which both suffered dead and each of which individually could lead to death.

There is still a conclusion about the cause of death by results of pathoanatomical dissection, as well as his differences with the final clinical diagnosis that are resolved by the expert Commission.

That is why who is in black and white writes:

The definition of true mortality from COVID-19 requires additional time. Today's data indicate that the overall mortality rate is 3-4%, while the mortality rate from infection would be lower.

However, most people never delves into the methodology of any study. It is enough to see in one sign, the words "cases" and "deaths" to voluntarily bring forward their interpretation of the data — "death from coronavirus".

Such surface disturbing study of the issue leads to a distorted perception of reality. Infected with the coronavirus of a man who jumped out the window or died from stage IV cancer, millions of inhabitants of our planet will unconsciously come to regard as the victim of a terrible epidemic. The situation is so absurd that I wouldn't be surprised if someone my direct quote from the who indicating that the source is in question or doubt.

Why bother to publish such mortality? — ask some. As an analyst will tell you that in statistics, a proxy variable. In our case, for example, are given the opportunity to compare them with the total or natural mortality. Yes, there will be a high proportion of errors, but to learn a new infection can be long, but for action world health organization need to quickly.

Italian precedent

Being within Chinese borders coronavirus bothered much smaller number of people. A new surge of hysteria has arisen on a background of frightening information field in Italy and drastic action by the local government.

More than 12 000 infected, with about 1,000 deaths and rigid quarantine. Italy makes us especially worry about the health of our parents and the elderly, because in Italy the greatest number of deaths among those infected occur among people okolovenoznoe age and older.

I have been to Italy and, in comparison with Russia, the numerical superiority of older citizens in the streets is noticeable even to the naked eye. The memory led me to thoughts about the necessity of studying demographic data.

It turned out that Italy has the oldest population in Europe. Almost 22% of people older than 65, the median age is 45 years. By the way, in Russia only 15% of people older than 65 years and the median was set at 40 years in the USA — 16% and 38 respectively. Now let's look at the distribution of coronavirus by age:

It is obvious that the high mortality rate among infected Italians primarily due to the age factor. The belief in reverse indicates a lack of understanding of basic error percent. For the doubters of my words I recommend you to read the article in Scientific American.

At the same time, looking at the chart, we see that older people are significantly more susceptible to infection even Pro rata. Hence comes the hypothesis that the immune system of the elderly becomes less able to coronavirus.

But such phenomenon is not typical for coronaviruses in General. In Korea, for example, the main group infected between the ages of 20 to 29 years — 29% of the total. In China, about 15% of the infected is a group of 70+, which is almost two times less than in Italy or France.

Recall that the coronavirus is the true cause of death. Therefore, we cannot say that the risk of mortal danger coronavirus for the elderly is a multiple higher. But the fact remains — elderly people with confirmed coronavirus die much more often.

It is impossible not to agree that the histogram of natural mortality by age looks to be very similar to what you see above, for the older people, the more likely the risk of death. So let's try to compare mortality in the presence of infection with a natural mortality.

For this we need to compare mortality data in terms of a thousand men from the national Bureau of statistics of Italy , together with data on the number of deaths among infected from dall Istituto Superiore di Sanità.

The mortality rate from the first table is equal to 1.05%, second — 5.8%. Does this mean that the lethality of the infection itself is somewhere at the level of 4%? Not at all. You do not forget that 76% of infected over 51 years? Such a sample is very different from the demographics of the Italian population and now I will illustrate.

When a uniform distribution of natural mortality in category of 90+ will be equal to 24% ((180.7+299.6)/2). And the value for the same group in the table COVID-19 is equal to 19%. That is, the mortality rate among infected is smaller than natural.

Hardly anyone will think that coronavirus still heals people, but the fact that he kills every fourth man many believe. Although this is the same degree of absurdity of the conclusions.

Too bad that we can't determine even an approximate fatality rate of infection comparing the overall value. We will assume the error distribution, as the table below shows the different age cohorts.

Moreover, data on natural mortality are based on hundreds of thousands of records, while the sample for the dead Italians with confirmed coronavirus does not exceed and thousands of people. We don't know how representative this sample.

For children, for example, the difference in mortality will also be negative, because the death of children happen, and infant deaths with confirmed infection have not been recorded.

Look also at gender. The grannies are much more resistant to the virus or perhaps a virus nothing to do with it and we observe the season of increased mortality among men, which increases overall mortality? — Unknown.

Specific figures on the mortality rate of the virus does not exist. They will be able to identify only employees who. Existing private speculation is irresponsible nonsense. My aim was to convey the idea that the risk of the deadly threat of coronavirus in the minds of most multiples are inflated and has nothing to do with reality.

Global collapse

We found that the mortality rate from coronavirus is not the deaths from coronavirus, and numerical values of natural mortality can look much worse than those from a new epidemic. However, on habré there are article with nearly five million views and it says the following:

Known mortality. In this scenario, 1% (details will be discussed later). This means that somewhere on February 12 in the region already had about hundreds of cases, and only one of them ended in death 17 days later.

Now use the average doubling time for cases of coronavirus — this is 6.2 days. Thus, for 17 days before the death of this man, the number of cases was multiplied by ≈8 (=2^(17/6)). It turns out that if you do not diagnose all cases without exception, only one confirmed death in the day means 800 of the true cases in the same day.

In Washington today, 22 dead. Using our estimates, we obtain ≈16000 true cases of coronavirus today. This is the same as the official cases in Italy and Iran together.

Where did the 1% mortality? — It turns out that the author, a Thomas Pueyo, to identify this factor focused on the data on the situation in the cruise ship Diamond Princess: with 706 ill, 6 deaths and 100 recovery.

Looking at these values Thomas concludes: "the final mortality is in the range from 1% to 6.5%". After that he takes the most gentle percentage, with a deft movement of brains shows how the virus will engulf the world and leaves the reader with the idea that it was demonstrated the best scenario when mortality is only 1%.

In my head just don't understand how this stuff could grow on habré. You do not know the caveat about the lethality of the virus and mortality when it is normal. But as programmers, mathematicians, analysts and other habracha made a projection 6 deaths with 706 cases in the population of the entire planet?

Is your representative sample? But looking at the somersault of Thomas with the approval of the 16,000 cases of infection on the basis of 22 of the victims, does not anyone understand that the probabilities don't work that way? You are not embarrassed that the youngest passenger Diamond Princess among the dead were 70 years old?

And this despite the fact that the cause of death of the passengers is unknown. For those who still don't see anything strange in arithmetic Thomas, I will quote a Professor of epidemiology at Stanford University's John Ioannidis:

Projecting mortality "Diamond Princess" on the age structure of the U.S. population, the mortality rate among people infected with COVID-19, will amount to 0,125%. But since this estimate is based on very small data among the 700 infected passengers and crew there were only 7 deaths, the real mortality rate may be as five times lower (0,025%) and five times (of 0.625%).

I believe that Mr. Pueyo published his Tsidulko with the best of intentions. But all his work is impregnated with the errors of induction and implication of data. I'm too busy to write a detailed refutation of such amateurish work. Such articles must be blocked due to misinformation of the population.

Infinite exponent

I believe that most of you have already seen such graphics online. But much less people are reading the official records of the who. In one of them, published on March 6, says:

One of the main differences between the Chinese virus and the normal flu is transmission speed. Influenza has a shorter incubation period and serial interval (time between successive cases) or three days. For COVID-19 this period is five to six days. This means that the flu is spreading faster coronavirus.

So what if all these exhibitors who persuades the reader to associations with the bubonic plague? Some may tell me that the deaths with confirmed coronavirus varies from 3% to 4%, while direct mortality from seasonal flu is less than 0.1%.

Again, do not confuse total mortality with mortality rates of infection. Such from the coronavirus humanity is unknown. Therefore, you will find claims who that the high mortality rate with a margin compensates for the speed of propagation and this makes Covid-19 multiples is more dangerous than the flu.

The world has long learned to determine the lethality of the flu. Therefore, I could not find data on total mortality for him. But still I managed to meet the who publication 2012 on influenza-associated mortality in China between 2003 and 2008.

Look carefully at the graph. In North China, the overall mortality rate for the flu ranged from 1.5% to 3%. I want to draw your attention that before you data on the period in 5 years tens of millions of records. Moreover, we clearly see seasonality. Now think about the representativeness of data Covid-19.

Having experience in public communications, I know that common sense reasoning is not enough and sometimes it is necessary to polacoat authorities. So I bring you a fragment of conversation, Washington Post Melissa Nolan, a virologist at the University of South Carolina:

Which virus is more deadly? That's a difficult question to answer for many reasons. First, health officials are not comparing analogous data sets between the viruses. They have years of influenza data but just months of covid-19 numbers - which are evolving by the day.

And, of course, I can't blame the attempt to project the outbreak of the flu in 1918 to today's pandemic from the article "Coronavirus: why we must act now". Perhaps a comparative analysis would be useful to understand the situation, but not comparative analysis of Thomas Pueyo conducted such a clumsy method.

The world's population over the last century has increased almost 3 times, the number of people over 65 has increased 10 times and 30 times increased the number of those over 85. Taking into account these factors, Mr. Pueyo? — No. He only had one schedule for Philadelphia from St. Louis to produce the approximation.

However, to solve this problem you need to determine the index weight the variety of parameters to calculate errors and calculate the correlation, not allowing asymmetry. This can handle only seasoned analysts and not Mr. Thomas, whose call to panic is replicated millions of hits.

The pandemic of fear

Under the circumstances, the most common two types of behavior: indifference and panic. However, the correct attitude towards the coronavirus is somewhere in the middle : in accordance with the recommendations of the who.

Discrediting the pandemic of fear, I have never belittled a real danger of coronavirus. The world health organization meet intelligent people, and I advise everyone to listen to their recommendations: wash hands, avoid public events, to abandon the cash and stuff.

The paradox is that when these same people from who or Rospotrebnadzora warn us about the outbreak of seasonal flu or encephalitis, we don't really respond to the messages. At least in comparison with the current situation.

The explanation of this paradox have long known: humans are irrational by nature. Besides, not all know the story of coronaviruses that have been already opened in 1965. Mind you, a lot of them. We can assume that some of you have already suffered a mild form of SARS caused by one of the members of the family of coronaviruses.

That irresponsible press has laid into the mass consciousness as the coronavirus — is SARS-CoV-2, which contributes to disease Covid-19. The last two loud coronavirus were reasons for the spread of atypical pneumonia (SARS) and middle East respiratory syndrome (MERS).

Why the concern in the information field from Covid-19 significantly more than that from other coronaviruses? Difficult question, the answer to which, perhaps, can only give epidemiologists. Personally, I want to mention two potential causes of this phenomenon.

First: this is a multiple increase of foreign tourism in China. Second, the growth of the Internet audience and the emergence of social networking, where everyone allows himself to think that, without competencies and without going into the methodology of the research, he knows the truth about the pandemic and the right to Dole out advice about how should act of power.

Any calls to follow the Italian General scenario should be despised. It is up to the government of each country in particular. The United States, England and several other States, for example, act in a different way. But, what about this says epidemiologist John Ioannidis:

Mortality in the whole population at the level of 0.05% — lower than that of seasonal flu. If is a real number, then the isolation of the world with potentially enormous social and financial consequences can be absolutely irrational. It is as if the elephant was attacked by a house cat. And, scared and trying to avoid the cat, the elephant accidentally jumps off the cliff and dies.

Many do not even imagine how can affect one week partial freezing of the economy on the GDP. Damage will be measured not only in monetary terms but in human lives. So please Faybusovich and other experts to put their opinions into oblivion.

The apotheosis

The emotional content is most popular, while anger is the most common emotion in social networks. The probability of prolongation of false information 70% higher, and more than half of readers share the news, not reading it on the header.

Unfortunately, fear wireline the truth. Not just Facebook, Google, LinkedIn, Microsoft, Reddit, Twitter and YouTube have teamed upto fight the fakes Сovid-19. However, the proof of the fallacy is often less popular than the original approval. I wish that this text was an exception.

If you read this writing completely and agree with my arguments, then take the time to share this article with your friends. Thank you all for your attention.

Sincerely, Ilya Pestov, the author of Groks.


RELATED MATERIALS: Politics and Geopolitics