Question of the day: how much does mathematical prediction of coronavirus cost?

Andrey Zlobin, mathematician, candidate of technical sciences, participant of international conferences on bioorganic chemistry, biotechnology and bionanotechnology

It is necessary to understand that there cannot be cheap mathematical models of the incidence of coronavirus.

At the end of December, the President of the Russian Academy of Sciences A.M. Sergeev recalled the adequacy of the models for the spread of COVID-19. The TASS news agency quotes the president of the Russian Academy of Sciences: “This should be a lesson and understanding that there are so many unknown numerical characteristics in covid. That whatever the best, most powerful model for the most powerful supercomputer you have, it won't work. Because empirical numerical characteristics are inserted into any model to analyze, to look at dynamics. We do not have them, and no one has ”.

The President of the Russian Academy of Sciences pointed out the main problem of mathematical models of the coronavirus. Predicting a pandemic is difficult because, firstly, it requires a large amount of empirical data, and secondly, a very complex mathematical apparatus and highly efficient computer technologies. At the same time, reading the current scientific publications on the mathematical modeling of COVID-19, one never ceases to be amazed at their primitiveness. It seems that the articles are written formally, that the main thing for the authors is to work out an extremely modest salary and increase the citation indices. In this case, the adequacy of the models goes far into the background.

Is another approach possible? Of course. There is the experience of the 90s, when the difficult economic situation in the country demanded the rapid saturation of the Russian pharmaceutical market with the necessary drugs. At that time, the author worked at the Analytical Center of a large commercial pharmaceutical company, which was one of the leaders of the Russian market. The responsibilities of analysts (professional mathematicians) included the study of morbidity throughout the Russian Federation. It was vitally important for commerce to know where and in what quantity specific drugs are needed, and at what prices. At the same time, the company did not skimp on research and a lot of money was allocated for scientific work. We (in the 90s!) Have created not only a powerful computer system worth tens of thousands of dollars, but also developed a methodology for obtaining the necessary empirical data on the incidence of various diseases. At that time, no one had this data either, there was nowhere to take them, the company received all the figures through the efforts of its analysts. Each figure was generously paid and, as a result, the commercial approach gave an objective picture of the incidence in Russia. The recognition of this work was international, with the largest pharmaceutical companies in the world eager to use our mathematical developments. The price bites. Despite this, however, there was a line of global pharmaceutical giants at the doors of the Think Tank. Selected results were published in professional pharmaceutical publications and were aimed at informing about the unique analytical tools.

Is it ethical to talk about a commercial approach when the coronavirus claims many lives every day? It is not only ethical, but also necessary, since only in this case the required volume of investments in science and mathematical modeling becomes clear. It is necessary to understand that cheap mathematical models of the incidence of coronavirus cannot exist in principle.

Predicting COVID-19 is very expensive math. In this case, the model is either expensive or bad. There is no third. Moreover, almost 30 years have passed since the 90s. Many mathematical methods have radically developed, the capabilities of computer technology, telecommunications have changed, revolutionary changes have taken place in the field of information and biotechnology, digitalization, big data technologies, and artificial intelligence are rapidly being introduced. An effective tool of the 90s for predicting disease incidence, then worth tens of thousands of dollars, today looks like an ancient fossil. The price for a modern similar development must be orders of magnitude higher.

It costs a lot of money to obtain empirical data, but this is the only correct way in the current situation of the global pandemic. Complex mathematics, which will have to "fork out" also does not imply simple solutions. To predict a dangerous pandemic was not fortune-telling on the coffee grounds, the mathematical model must currently be developed by a large interdisciplinary team of scientists. Let's leave the irrelevant question "who will give the money?" It is more important to answer: is it possible to create a commercial adequate mathematical model of the incidence of coronavirus? Yes, you can. There is no doubt about that.