One should always be wary of perception bias when it comes to figures or graphics. An excellent example concerns the epidemic of the coronavirus “NCoV” which is particularly prevalent in China. Researchers from the American John Hopkins University publish in real time the main figures available on this subject. An educational tool provided that it does not fall, in this case, into two traps: if the curve of the total number of infected people is looked at on a computer screen, of horizontal shape, this one gives the impression of be fairly flat. It is certainly not reassuring because the increase is very real, but does not invite catastrophism.
Read also: Coronavirus: in the Kafkaesque hell of a hospital in Wuhan
Conversely, on a smartphone screen, of vertical shape, the same curve, with the same data, seems to rise much faster at the extremes because the scale of the abscissas is more compact. At the time, it seems much more worrying. Another phenomenon: this curve gives all the data, including those for the same day, even though these are only provisional.
2019-nCOV Global Cases by John Hopkins CSSE. John Hopkins University
As a result, the slope of the curve seems to abruptly decrease on the last day: on social networks, several internet users have concluded that the epidemic was slowing down ... wrongly because it is almost certain that the final results, up on the results temporary will straighten the curve. It is therefore necessary to observe the curves and read the figures carefully. With Professor Arnaud Fontanet, researcher at the Pasteur Institute, where he heads the Center for Global Health, let's decipher several graphs that allow us to understand the coronavirus epidemic.
See also - Coronavirus: French returnees "are doing well", assures the government
Coronavirus: French returnees "are doing well", assures the government - Watch on Figaro Live
● Where is the epidemic?
Conflicting figures exist on the subject. On the one hand, China publishes daily the number of cases diagnosed, the number of deaths, as well as the number of cures. With regard to this first source, the number of people infected on February 3, 2020 stood at almost 20,438. This first figure has increased steadily since January 20 (278 cases). Other state sources add 194 cases to the rest of the world. Not only does the curve obtained increase, but its slope increases, that is to say that the number of new cases recorded each day is also increasing. There were 17 new cases on January 17, 688 on January 25, 3278 on February 3. The graph below measures this increase in the number of daily cases.
But here again, be careful when observing this curve: these are cases officially diagnosed by China and not cases in absolute terms. On January 31, the British journal The Lancet , one of the medical references, published a study based on a mathematical model according to which 75,815 individuals were infected in Wuhan - the main Chinese province affected by the epidemic - as of January 25, 2020. Officially, on that day, the number of people diagnosed was ... 685. Or a ratio of more than 1 to 100.
● What comparisons with Sras?
How to understand these figures? Monday February 3, on Europe 1, the Chinese ambassador in Paris, Lu Shaye, declared that he was " convinced " that the figures published by his government were " transparent ", adding that in addition to " 17,238 confirmed cases " , it was also necessary to take into account " 21,558 suspected cases ". “ As the article in The Lancet suggests , we are probably far beyond the number of officially declared cases. The most likely thing is not that China is trying to hide the real numbers, but rather that they are unable to keep up with the spread of the epidemic and update their diagnoses in the face of the real increase in the number of case ”, analyzes Professor Alain Fontanet.
If we observe this time the official number of deaths, the daily increase appears, there again, in a significant way: There were 16 deaths on January 24 against 64 on February 3, which brings the total to 425 dead in China (427 worldwide). Symbolically, the first figure exceeds the number of deaths caused by the SARS epidemic - "severe acute respiratory syndrome" linked to another type of coronavirus - in mainland China in 2002/2003.
Read also: Is the coronavirus more dangerous than the flu, SARS or Ebola?
Above all, as the graph below shows, for SARS, the maximum number of deaths in China - 425 dead - had been reached on July 1, more than three months after Beijing had revealed to be affected by the epidemic. With the NCoV, this record was reached just two weeks after the first daily report published by China. On the other hand, the 427 deaths worldwide of the new Coronavirus do not yet reach the number of deaths caused by the entire SARS epidemic in 2003, in other words worldwide (812 deaths) and no longer only in mainland China.
Relative to the number of cases of infections, this number of deaths gives a mortality rate for SARS (from 9 to 10%) at least three times higher than for the new Coronavirus (between 2 and 3%). A difference that is easily understood by observing this time the comparative evolution of the number of people officially infected with the two viruses. With more than 20,000 official cases, the NCoV affected more than twice as many people in less than a month as the SARS in more than three months (8,400 cases).
“ Again, you have to be careful when looking at these data. There was almost no mild form of SARS, which seems to be the case with the new coronavirus. In other words, depending on whether you take into account with more or less precision all these mild cases, you will see your mortality rate logically decrease or rise ”, comments Professor Fontanet, who prefers not to compare the new coronavirus with that of Sras : “ We see that the NCoV spreads much faster than the SARS. From this point of view, it is closer to the flu. In the event of an influenza pandemic, it has not been uncommon in history for a quarter of a population to be affected ”.
See also - Coronavirus epidemic: where are we?
Coronavirus epidemic: where are we? - Watch on Figaro Live
● Can we foresee the future?
Can we extend the curves in one direction or another? Can we predict the peak of the NCoV epidemic from which the number of new cases registered each day would begin to decrease? Some Chinese researchers have estimated that this peak could be reached quickly, as early as next week. Other researchers have estimated that the peak is more likely to be reached in April or May. " No one really knows. It is premature, it seems to me, to anticipate these evolutions ”, slices professor Fontanet, who evokes several uncertainties like the possibility that a“ new epidemic of NCoV starts in other countries than China ” . On RTL this Monday, Professor Christophe d'Enfert, director of the Fungal Biology and Pathogenicity Unit of the Pasteur Institute, also spoke of the preponderance of these uncertainties: " Viruses, in particular this type of virus, mutate enough easily. And so we can imagine that it acquires mutations that will make it perhaps more contagious, or more pathogenic. It's impossible to predict . ” All the more reason to observe the curves, but without drawing hasty conclusions.