Hard to say that no one had warned.
"The arrival of the variant shows that we will find ourselves in mid-March in a health situation with significant consequences for people entering intensive care," warned on January 28 the president of the Scientific Council, Jean-François Delfraissy, questioned. by the parliamentary office for the evaluation of scientific and technological choices (OPECST).
"I have no rational argument to tell you that we are not going to undergo a progression of the epidemic which could begin in March-April", pointed out on January 17 on LCI the epidemiologist Arnaud Fontanet, professor at the 'Pastor Institute.
Less than two months later, we are there.
The health situation linked to Covid-19 has deteriorated markedly in recent weeks, with an incidence rate increasing by nearly 20% in one week.
"All indicators, already at high levels, are on the rise," alerted Public Health France Thursday evening in its weekly epidemiological update.
In Île-de-France, the number of patients in intensive care (1,201 on March 18) has exceeded the peak of the second wave.
After having rejected it for a long time, the government has resolved to take new containment measures (lightened) seven days a week. The Ile-de-France region, Hauts-de-France as well as the departments of Seine-Maritime and the Alps -Maritimes and Eure will be concerned from midnight this Friday evening.
The majority British variant since the end of February
For several weeks, several models have been carried out by researchers and epidemiologists.
Often criticized by some for failing to predict the evolution of the epidemic in the short term, they are mainly intended to be projected in the medium term.
Even if they can of course be wrong - the aim of the government being besides to make them "lie" -, this time they were rather right for the timing.
READ ALSO>
Explosion of the Covid-19 epidemic: what credit should be given to these models who imagine the worst?
This is especially true for the propagation of the B.1.1.7 variant, known as “British”, estimated to be between 30 and 70% more contagious.
In a note of January 16, epidemiologist Vittoria Colizza and her Inserm team estimated "that the variant would become dominant in France between late February and mid-March".
According to data from Public Health France, the switch took place during the week of February 17 to 23.
Preliminary results from the 1st large-scale sequencing analysis in France indicate the presence of 1-2% of VOC 202012/01 in w01 in the country.
This suggests that VOC could become dominant by the end of February in absence of interventions.
https://t.co/vfV56gaDwj pic.twitter.com/mbXA0DMlzf
- Vittoria Colizza (@vcolizza) January 18, 2021
This same Inserm model projected different scenarios of the epidemic, each time with a more or less significant impact on the hospital.
With a so-called “British” variant 50% more contagious and a reproduction number “R” (that is to say the number of people that a positive case will infect on average) of 1, we had to expect approximately 15,000 hospitalizations in week 11 (March 15 to 21).
However, from March 12 to 18, this number was "only" about 10,500 (up 7% in one week).
In this case, the model probably did not take into account the effectiveness of measures taken in early or mid-January, such as the generalized curfew at 6 p.m., and it may also have overestimated the actual overcontagiousness. of the "British" variant.
Morning essentials newsletter
A tour of the news to start the day
Subscribe to the newsletterAll newsletters
"We did well not to follow the forecasts of certain models which indicated a runaway epidemic from mid-February", rejoiced Jean Castex Thursday evening.
“By the time the model is made, you can't think you know everything about the variant.
So we have to make choices based on known scientific data, ”defends Mircea T. Sofonea, lecturer in epidemiology and evolution of infectious diseases at the University of Montpellier.
"All models are wrong, but some are useful"
With his team for modeling the epidemiology and the evolution of infectious diseases, this teacher-researcher is also used to making projections.
One of them, dated January 21, predicted that the number of Covid patients hospitalized in critical care would reach 4,000 in mid-February and 5,000 in early March nationally.
A month later, at the end of February, the same work, updated, established "that in the absence of a change in health policy", this number could exceed 5000 "at the beginning of April".
There are ultimately “only” 4,300 in mid-March, but this number has increased by 17% in two weeks.
The increase is more or less strong depending on the region, Île-de-France being one of the most affected.
And the lightened reconfinement which begins this Saturday could make this new projection “lie”.
"Our models anticipated what happened in the sense that they fulfilled their function of warning of a rebound in March, even if the fluctuations of the epidemic shifted the curves", indicates Mircea T. Sofonea .
VIDEO.
Covid-19: France has entered "a form of third wave", says Castex
Two researchers from the Lille University Hospital, Philippe Amouyel and Luc Dauchet, have also drawn up their own models but looking at the number of new cases identified.
Their first results, published by the JDD on February 6, initially showed an explosion for the beginning of March with 50,000 people positive per day.
Three weeks later, this deadline was postponed to mid-March.
“The curve has shifted over time but we still find the same.
It goes back ”, then explained Philippe Amouyel to our colleagues.
The actual number in mid-March is rather around 25,000 new cases per day on average, but it has increased by 20% in one week.
The model therefore "got it fairly" that there would be an increase during this period, but it was denied regarding the magnitude in mid-March.
READ ALSO>
"This is not what we expected": epidemiologists confused by the drop in cases
This is the whole difficulty of modeling work, which is often perceived as forecasting.
Modelers consistently insist on the use of the less engaging term “projections”.
“These scenarios are made on the basis of incomplete data and uncertain assumptions.
The spread of the SARS-CoV-2 virus remains difficult to anticipate and the dynamics of the epidemic can change rapidly, ”the Institut Pasteur indicated as a“ warning ”in a note dated November 30.
To use the phrase of the great British statistician George Box: “All models are wrong, but some are useful.
"