Statistics: In the first time about covid-19, the newspaper repeated repeated HV errors

Statistics on the course of the disease in the Czech Republic need to be taken in two different ways, as is not the case in the Czech Republic, according to statistics Karel Helman.

In the public space of both public and well-known public officials, the so-called error repeatedly appears, which makes it difficult to understand the situation around the SARS-Cov-2 coronavirus infection. I don’t think this is a change, especially as a result of the misunderstanding of some aspects of my field.

Let’s just say that the introduction not only for the economy of devastating government measures in the middle of the ordinary was understandable. Uncertainty about the condition and development of the disease was enormous at the time. It is also important to emphasize that as long as the measures are in force, they must be observed. What this does not mean is that everything is right.

One big and at the same time understandable relatively simple problem of the discussion of the disease in the cities and elsewhere is the systematic ordering of two completely different indicators of the state of the disease. Namely, the sweat of the disease and the sweat of people with proven covid-19 (according to the terminology of the Ministry of Health).

It is a very different quantity. The number of nakaench was, is and always will be unknown. We could find out it accurately only if all citizens were (at one point) tested. But that, of course, is not the case. One can only try to estimate the number of nakaench based on the extensive representative survey that is taking place these days. Even after the survey, however, it will only be an estimate, the actual number of nakaench and the development of this sweat in time will remain unknown.

It is known only to the number of people with a proven disease, which is constantly employed in the media and elsewhere with sweat. The two may not be related in any way. It is calm that, even at the time when vude talk was about the number of nakaench sweat, the number of nakaench among the many citizens of the Czech Republic decreased. Only the number of cases detected increased.

I understand that it is a shortcut to communication, but the difference between how it can be, the girls, the listeners and the listeners can share the number of rising steps! and the information that the increase in the number of people with a proven disease, the development of sweat as unknown is enormous.

Work on sweat will give a completely wrong impression that it has something to do with the condition (and development) of the infection. The informed value in the fact of the annotated indicator of the number of people with a proven disease in terms of the actual condition and the time of development of the infection is completely zero. Let’s not learn anything from him. The dominant space is mixed and fills the mark.

Equally, there are a large number of deaths on covid-19 (which is still completely unknown) and the number of deaths for which covid-19 was proven at the time of their death. In the city and in public accounts, all those who had a positive test for the presence of the virus in their bodies before death are very often described as victims.

This does not mean that the virus was the cause of death or something. One noticeable analogy: if a hunter lands sweat, and if someone stabs him in the chest, there will be both defects. However, we are figuratively witnessing the fact that a hunter with cancer, who was killed by someone with a bad finger in his finger, is also called the victim of the betrayal.

Role vbru

On the other hand, we forget about how many factors affect the number of countries and the number of people with proven disease (and not only them). In a crucial way, for example, they influence the decisions of those who follow them. A good example is the choice of people for the test room (so-called test strategy). For example, if we only test people in the open air, it will be relatively high by country (according to some information, it is possible that this is a case of data from Italy); if we tested only people who are completely healthy at first glance, it would be zero, according to the earths.

We could do something like that in those days. A significant change in the testing strategy is currently taking place in the Czech Republic, for example, the so-called smart quarantine will be introduced. As a result of this change, if the smart quarantine meets the declared duty, a change can occur according to the test with a positive result.

The values ​​from the time when the quarantine smart will really work will therefore not be directly comparable with the old values ​​(only the development trends will be comparable). Thanks to the new system, the number of people with a proven infection will increase.

If the situation from the last few weeks is repeated, some of me may conclude that the curve is up, or that it has been added – but this is a mistake, in fact, the capture of infected people has just stuck. (If a clever quarantine were not detected, the accuracy of the identification of infected persons would not be increased, it would be introduced in a very debatable step.) Faulty barrels of this type would be the result of insufficient understanding of the impact of the testing strategy.

Neastn likes absolutely

Take for example about a week old prof. Roman Prymula that the criterion for the release of measures could reach the limit of 400 nakaench. Professor Prymula undoubtedly had in mind 400 positive testers a day, and his year is probably just a murderous acronym, not much sense.

If it is possible to get enough daily number of tests performed, it would be possible to achieve the number of sweat soaked. If we passed 20,000 tests per day (which is the ken capacity according to the publication of Andrej Babie), we would achieve a sweat of 400 people with a proven infection, even if, according to the positive results, it was only 2 percent. Which would indicate that the infection is probably very small.

It is generally a problem with the absolute data around the virus that most of them are affected by the number of tests performed. This is especially true of the number of people killed and the number of people with confirmed covid-19.

Of the absolute selves, tusk is to monitor the sweat of a patient with an external course of the disease due to the capacity of the health system. If the goal is to monitor the status and time development of the disease, the benefit of most other indicators expressed in absolute values ​​is either zero or very doubtful.

As a statistician, of course, I do not think that even with the day of lack of data, there is no day of quantity that would not make sense to monitor. Many of them, unfortunately, do not pay attention to the public, either because they had attention.

The best of the range of limited possibilities we have to monitor the development of the test is the daily percentage of positive tests performed according to the sweat of people with a proven disease to sweat (on a given day). Even with the knowledge that this floor is, of course, still influenced by the testing strategy we are testing. This indicator is a bag in public space either completely defaced or lack of accent. And what about the state and development of contagion in the Czech Republic?


Daily according to the sweat of people with proven disease and sweat test. The red points indicate the restriction of free movement of persons (16.3.) and the obligatory noen rouek (19.3.).

Let’s take it chronologically. After that, the values ​​are burdened by relatively uncertainty, because the test was successful. Nevertheless, we can state that the piblin from the 11th run (458 test) to the 20th run (2085 test) did not increase significantly (nor did it decrease) and fluctuated around the value of 7%. During this period, in the same way, the sweat in Lombardy ranged from about 30-60%.

On Saturday the 21st of March (1880 test) was according to 8.4%. Since then, a gradual decline can be observed. In the first days of April, the daily number of people with (newly) proven infections in R was 3-5% (Lombardy 10-25%). Thus, throughout the follow-up of the disease covid-19, the situation in the Czech Republic was completely incomparable in terms of the extent of the spread of the disease with Lombardi, even before the introduction of governmental measures.

In the period from 13 to 22 April (ie after the spontaneous release of Easter, their influence, epidemiologists and some government officials said) did not deviate from the range of 1 to 2.5% and was very close to its theoretical minimum, ie zero. If the project is a smart quarantine, it should come from outside.

But the main thing is that we have been in a situation for a long time when, from a mathematical point of view, there is nowhere to fall. From this point of view, there is nothing to wait for when releasing the government measures.

The author works as an assistant professor at the Department of Statistics and Probability of the University of Economics in Prague.