We knew the regional and departmental scales, but the Covid-19 epidemic can also be examined at a much finer level.
This is particularly useful, at a time when Île-de-France is in a critical situation with an incidence rate of over 600. For the whole of France, it stands at 380.
This indicator corresponds to the number of people tested positive in a territory, compared to a population of 100,000 inhabitants.
It is greater than 1,000 in 1,818 of the 35,000 French municipalities, according to the latest data from Public Health France for the week of March 20 to 26.
Three weeks earlier, only 740 municipalities were concerned.
The exact figures are not known at this very local scale, but only the slices are available.
And 1000 is already four times the maximum alert threshold (250) set last fall.
At least one in 100 inhabitants positive this week
Of course, measuring the incidence rate does not necessarily make much sense in villages with 200 or 300 inhabitants.
It suffices then that two or three residents are tested positive during a week for the incidence to climb to 1000. But such a level is much more significant in the cities.
Fourteen of the 1,818 municipalities concerned have more than 10,000 inhabitants.
Not surprisingly, almost all of them (thirteen out of fourteen) are located in Île-de-France, a region both by far the densest and one of the most affected by the epidemic.
Four towns are located in Val-d'Oise, three in Seine-et-Marne, two in Seine-Saint-Denis, two in Essonne, one in Val-de-Marne and one in Yvelines.
Concretely, an incidence rate greater than 1000 means that out of a group of 100 inhabitants, at least one of them has tested positive during the past week.
“It is both symbolic and very important, it shows an unfavorable development of the situation.
But it is different if it is related to clusters that can be traced or if the virus circulates there everywhere ”, underlines epidemiologist Mahmoud Zureik.
In theory, if you go to a place with 20 people, there is at least a one in three chance that one of them will be infected, according to the CoviRisque tool of the CovidTracker site.
This site assumes that the actual incidence is twice that calculated from positive tests, then a mathematical formula mills both the incidence rate and the number of people present to display a percentage risk of falling. on a contaminated case.
“Doubling the incidence rate known to have the true incidence is a commonly accepted working hypothesis, based on seroprevalence surveys and the correlation between confirmed cases on the one hand and hospitalizations and deaths on the one hand. other ”, indicates the mathematician Jean-Stéphane Dhersin, researcher at the CNRS.
On the other hand, he sees two possible biases in this calculator.
The first is the fact of "taking a group at random from the population, when the people who meet somewhere are perhaps more contaminated than the others".
The second assumes that “if you organize a gathering, there may be other previously infected people within that subgroup”.
But all this is "very difficult to calculate", he admits.
The cluster effect
Positive cases are classified by place of residence, which means that an inhabitant of Bondy (Seine-Saint-Denis) traveling to Paris and tested on site will be counted in Bondy.
But with the ban - unless there is a compelling reason - of trips of more than 10 km, everyone is supposed not to go on vacation or go to live temporarily in another region of the country.
READ ALSO>
Covid-19 in France: most of the country is in the red
Morning essentials newsletter
A tour of the news to start the day
Subscribe to the newsletterAll newsletters
In these kind of small towns, an incidence rate can very quickly climb.
It suffices for a cluster of ten cases to be detected at once, for example.
Moreover, the larger the population of a territory, the lower the probability of reaching or even exceeding an incidence rate of 1000.
The two most affected Ile-de-France departments, Val-d'Oise and Seine-Saint-Denis, are nevertheless almost at 800. A level never reached so far and well above the peak of the second wave at the end of last October (respectively 620 and 540).
Several metropolises had already exceeded 1000 in the past, such as Dunkirk at the beginning of March or Saint-Étienne last November.
If we broaden the spectrum of municipalities to all those with more than 2,000 inhabitants, some located in regions other than Île-de-France also show an incidence of more than 1,000 currently.
This is the case, in particular, of Mandeure (Bourgogne Franche-Comté), Arleux (Hauts-de-France) and Anould (Grand Est).