Skip to main content
Fig. 4 | BMC Public Health

Fig. 4

From: Clustering and mapping the first COVID-19 outbreak in France

Fig. 4

Source: Own elaboration based on data from Santé publique France, INSEE. Data sorted out by positive tests conducted in city labs. The variety of local situations is extreme. This chart alone invalidates the “blind” lockdown strategy adopted by the French government. In a limited number of cases (green), the “logical” hierarchy of indicators is respected (positive tests > hospitalizations > deaths). This logic implies that the number of persons tested positive is greater than the number of hospitalized patients, considering that only a small proportion of infected people develop a severe form of the disease. Similarly, the number of hospitalized people should be greater than the number of patients who passed away with/because of SARS-CoV-2, given that mortality is limited within the cohort of hospitalized patients. Most often (black), the overall number of positive tests conducted in city labs is less than the number of hospital admissions (hospitalizations > positive tests > deaths). The position becomes even more curious (red) with positive tests below the level of COVID-19 related deaths (hospitalizations > deaths > positive tests). This situation illustrates the varying reliability of the official database during the emerging phase of the statistical system dedicated to COVID-19, both through time and space. There are several possible reasons for this. (1) The departments registering a small number of COVID-19 cases took more time to implement the statistical platform; (2) Frontline hospitals were already conducting their own tests but these are not reported in the city labs database, which interferes with the results; (3) The reliability and homogeneity of RT-PCR testing procedures might explain certain local discrepancies; (4) The number of reporting public hospital or clinics and city labs has fluctuated over time

Positive tests, hospital admissions, and deaths by department as a proportion of local population, lockdown phase.

Back to article page