The number of infected is unknown. Tests are not enough, we must react "exaggerated"

People all over the world are watching the number of people infected in their country rising. However, this number represents the number of people who were confirmed to be infected with the new coronavirus. It is a small detail at first glance:

But the difference can be huge. And the problem is, now we don't know how big the difference is. Maybe because testing is slower than it should.

In an ideal (hypothetical) situation, all the people of the world would be tested. Those who confirmed the infection would stay at home for 14 days. Problem solved, pandemic stopped. But this is completely unrealistic. In fact, only a small percentage of people can be tested quickly with current technologies. This means that there are definitely many more infected than we know about. That is why measures such as restricting free movement, washing or disinfecting hands and wearing a veil are recommended to slow down the epidemic.

How many tests, so much information. Beware of "estimates"

Imagine a country that does not test a single person for the presence of coronavirus. Her official infection number would be zero. At the same time, the information value of such a number would be zero. With each new test, the amount of information we have increases.

In the Czech Republic, we have very accurate statistics on how many people are tested. In the last two weeks, the number of tests per day has been increasing, and the number of confirmed cases has been increasing every day.

It would be a good idea to calculate the percentage of tests that return positive, multiply that percentage by the entire population, and "estimate" the total number of infected. But beware, such a calculation would only make sense if the individuals tested were selected at random. However, because people who are more likely to become infected are logically selected for testing, such a triple is unusable and misleading. The inaccuracy of the "mechanistic estimate" is greater the better people are selected for testing.

Coronavirus

Races in testing speed and availability

Obviously, testing is key to finding out the real situation. But different states have different approaches to testing. Unfortunately, this also means that the numbers are difficult to compare. OurWorldInData server has probably shown the most detailed comparison so far.

Right after China (dates only until the end of February), when the new coronavirus epidemic broke out for the first time, South Korea follows, which can boast a high number of tests per capita: out of a million people, 4,831 people were tested here. In other words, almost half a percent of the population has been tested. The test sites, where people don't even have to get out of the car, also have credit for this: the workers take a sample of their saliva and find out the result by the next day.

Testing for SARS-CoV-2 virus in South Korea

In the Czech Republic (until March 18), 7,664 people have been tested so far, ie about 719 Czechs tested per million inhabitants. This number has doubled in the last four days, and now the number of tests is growing exponentially, even a little faster than the number of infected.

The United States is worse off. Originally, they only tested the official laboratories of the CDC (US Federal Institute of Health), and by March 12, only 4,255 had been tested. At the end of February, state laboratories joined them, testing a total of nearly 28,000 people. Altogether, the official numbers in the US per million inhabitants are only about 97. The New York Times also included numbers from the COVID Tracking Project in the calculation, but only 125 tested per million inhabitants in the USA.

Only in the last week in the USA, test sites following the example of South Korea begin to start operating. In America, they are used to drive-thru, you can even visit the bank without getting out of the car. Testing is so convenient and one does not come into contact with anyone along the way. Therefore, it can be assumed that the number of tested will increase significantly in the coming days.

Drive-thru testing in Denver, USA

With that, the number of infected will increase. Paradoxically, more intensive testing will naturally bring unfavorable news: The situation is worse than we thought. The virus has infected more people than ever before.

But real information about the extent of the infection is needed. Not just for doctors and epidemiologists. They are important for the public and for politicians. On the basis of higher numbers (ie numbers closer to the real number of infected), it is possible to enforce measures that would be difficult to enforce in the context of lower dozens of infections.

Our fantasies cannot work with exponents

It is the rising number of confirmed cases that has forced the United States to increase the number of cases tested. This is the paradox of the current situation. Until nothing happens, it may seem too early for any action or intensive testing. But for every person we know about, there are many others we don't know. He is infected with an unwelcome infection, so he spreads it on.

In fact, a spread of a virus such as Ebola could not be overlooked. Ebola kills more than half the people who get it. It's too "conspicuous." The new coronavirus, on the other hand, kills a much smaller number of people (mortality is estimated at around 4.5%), as well as older, usually less active people. Thus, the virus can spread exponentially.

Intuitively, we can imagine that out of fifty infected people, one hundred and fifty can be infected in a week. But our imagination can't work with big numbers, and certainly not with exponents.

And exponential growth is something the human imagination can't handle. Intuitively, we can imagine that out of fifty infected people, a total of one hundred and fifty people can be infected in a week - when each of them infects two more. Maybe we'll remember that in the next week it could be almost five hundred. But our imagination can't work with big numbers. Or do you guess how many will be infected in our hypothetical situation in just 14 weeks?

As we can see, the human brain can't deal with exponents. This is the first cognitive error that awaits us in assessing the epidemic. The second is our subconscious preference for normalcy (in English normalcy bias). In the face of disaster, people often refuse to believe that disaster could occur. According to psychologists, longer than healthy, they think that everything will be fine, that everything will "return to normal" soon. Usually yes. But sometimes not. Sometimes a hurricane hits the city. And even then, people don't want to admit that they should change anything.

The reaction must seem exaggerated to be successful

Let's go back to Italy. At the end of February, the graph of the number of confirmed new coronavirus infections looked like this:

It was a lot, but in a country of 60 million people, there are still "only" dozens of cases. In addition, the optimist could see a downward trend in the chart. There might seem to be no reason to panic. Perhaps only a look at the death toll (Italy had 12 deaths by February 27) should have suggested to an untrained observer that the epidemic was definitely not under control.

It's like a view of a slowly flooding river. If the level has risen by only a few centimeters, you can say that nothing is happening and there is no need to evacuate the city. But if your city is downstream from the dam and you find out that the dam has ruptured, it doesn't matter how fast your water rises now. You know for sure that you will have a flood in a few hours. Europe has been watching the "flood" in Wuhan, China, live since mid-January.

Nevertheless, it took almost two weeks for Italy to introduce a strict quarantine on March 10. Since then, the development of the infected has looked like this, and has not yet stopped.

The Italian government did not introduce strict quarantine until it was clear to everyone that the epidemic required significant measures to prevent tragic loss of life. It was the day after the number of victims in a single day exceeded one hundred. Until then, the virus had spread among humans and was growing exponentially.

Every measure - even if it is effective - will only be reflected at a later stage. People who became infected a week or ten days ago are now entering crowded Italian hospitals.

At a time when it is clear to everyone that restrictive measures are needed, it is ten days late for such steps. But when the time is right, the measures seem too restrictive. The irony is that if they work and the epidemic can be slowed down, it could be water on the mill for those who protested against the measures.

But it's like those exponents. Absolutely unintuitive. In addition, we work with numbers that do not express the true state. Measures that act in a "reasonably" perceived situation are thus in fact too mild. Conversely, restrictions that seem exaggerated at first glance are more likely to succeed in slowing the spread of the disease.

When shooting at a moving target, you must aim where that target will be, not where it is. And this target is accelerating exponentially. Which, fortunately, is because we can learn from the mistakes of other countries. But it will all return to normal, just wait. No?

In 1969, American psychologists Bibb Latané and John Darley conducted a wonderful experiment (PDF). They sat the participants in a room and let them fill out a questionnaire. Suddenly the room began to fill with smoke. If the participant was alone in the room, he got up in a moment and went to report a strange smoke (75% of them did). However, if the participant saw other people in the room who did not notice the smoke, he shrugged and continued to fill nervously. But when there were more people in the room, there was a much lower chance that one of them would get up. In groups of three, only 12 percent of the experimenters did so.

The response to unexpected and abnormal situations does not take place individually. We make decisions based on incomplete numbers, we face the non-intuitive exponential spread of the infection, and we also constantly look at what others think about our reaction. The room is filled with smoke.