COVID-19: Does France underestimate the number of cases?

This blog talks about the actual COVID-19 situation of France, the definition of confirmed cases in France and some predictions of confirmed cases in France.

Today I read an article on franceinfo that discusses if France underestimate the number of cases, and expressed that the difference between the number of tests carried out in different countries is explained by different approaches to the health crisis; in France, a “preselection” of patients is carried out upstream: tests are therefore less frequent, but they better target people potentially carrying the virus. I also searched some information on the website of French government, and will talk about this subject with the following points:

  • Actual COVID-19 situation in France
  • Definition of confirmed cases
  • Prediction of confirmed cases in France

Actual COVID-19 situation in France

Yesterday, the COVID-19 confirmed cases of France passed 1000 cases. As of 2020-03-07, there are 1009 confirmed cases, 12 cured and 11 death cases in France.


The plot below describes the new confirmed cases per daily, published by Sante Publique France.

France daily new cases

Definition of confirmed cases

According to the publication on Sante Publique France, we know that

  • The definition of close contact now includes all contact from 24 hours before the onset of symptoms in a confirmed case of COVID-19.

  • Possible case

a) Anyone with clinical signs of acute respiratory infection with fever or feeling feverish, AND Having traveled or stayed in a risk exposure zone within 14 days before the date of onset of clinical signs:
- On a case-by-case basis and after consulting Public Health France, a confirmed or potential exposure to a cluster-type event, documented outside these risk exposure zones, may also be considered.

b) Anyone with clinical signs of acute respiratory infection within 14 days of any of the following:
- Close contact of a confirmed case of COVID-19;
- Co-exposed person, defined as having been subjected to the same exposure risks (that is to say a trip or stay in a risk exposure area) as a confirmed case.

c) Any person, even without the notion of travel / stay in a risk exposure zone or close contact with a confirmed case of COVID-19, presenting:
- Pneumonia for which another etiology has been excluded beforehand on the basis of clinical, radiological and / or virological criteria and whose clinical state requires hospitalization, OR
- Signs of acute respiratory distress up to ARDS (Acute Respiratory Distress Syndrome) in a context possibly viral and without any other etiology obvious from the start.

  • Confirmed case

Anyone, whether symptomatic or not, with a sample confirming infection with SARS-CoV-2.

  • The list of risk exposure zones
Continent Country
Asia China
Asia Singapore
Asia South Korea
Asia Iran
Europe Italy: Regions of Lombardy, Veneto, Emilia-Romagna and Piedmont

Prediction of confirmed cases in France

Method 1: predicting with France’s actual daily evolution

For the first approach, we can calculate the mean evolution value of daily confirmed counts. Considering that there are continuously new cases since 24 Feb. 2020, we calculated the mean evolution between 2020-02-24 and yesterday (2020-03-07), got 43.3%, then predict the confirmed cases with this value for the following week.

Prediction with France history

Method 2: simulating with Italy’s different segment evolution

You may ask why we simulate Italy’s evolution for France ? Since actually France tests only the people who have been to the risk zone or who have the symptom of COVID-19, which risks missing some cases. However, Italy tests all people from the epidemic area of northern Italy or who have close contact with the diagnosed person, with or without symptoms.

This method calculates firstly Italy’s mean evolution for the segments greater than and less than France’s cumulative confirmed cases, then calculate the ratio and apply on data of France. In my case, there are 1009 confirmed count in France; in Italy, the mean value of daily evolution for the period less than 1009 cases is around 32%, the mean value for the period greater than 1009 cases is around 26%, so the ratio between 32% and 26% is 1.23. Now we calculate the mean value of daily evolution of France, get 17.8%, then we use 0.178/1.23, get 14.5%, which is our predictive daily evolution. The result is shown as below:

France simulates with Italy avg daily evolution

Easter egg

I just saw a comparison between COVID-19 and flu, which is pretty detailed and worth sharing with you:

COVID-19 vs. flu