URL:
https://www.medrxiv.org/content/10.1101/2020.04.15.20062752v2

Type d’article :
Preprint

Type de contenu :
Modèle, Estimation de paramètres, Données épidémiologiques, Prédictions,

Thème :
Epidémiologie

Que retenir ? :
Development of a modified SEIR model to account for the situation of Mayotte, and use of data from Mayotte Agence Régionale de Santé. Estimation of R0 before and after lockdown measures.

Description de l’article :
The authors introduce a modified deterministic SEIR model, that divides the total population into susceptible (S), exposed (E), infected (I), simple or mild removed (RM), severe removed including hospitalized (RS), and death cases (D). They use data from the Agence Régionale de Santé in Mayotte. First, they provide an estimation of the constant transmission rate parameter to the epidemic data in Mayotte during an early exponential growth phase. Then they predict the epidemic without any control in order to understand how the control measures and public policies influence the control of the epidemic. They introduce a temporally varying decreasing transmission rate parameter with a control (quarantine) parameter. Then they investigate numerically the influence of this control parameter: results, in term of predictions, seem to be very sensitive to this parameter value, probably due to the fragility of Mayotte health infrastructure and the significant fraction of the population without access to water.

Commentaire : The focus is on Mayotte island and data show very interesting trends affecting small numbers (less than 250 infected cases), and include cases imported through international travel (before the airport was closed). The novelty of the paper is the modified SEIR model applied to real data collected in Mayotte with initial conditions such that the initial exposed population relying on the real population density and observed number of initial infectious. The first results concern the deterministic model applied to real data, and seem to be strongly dependent on small variations in the numbers of cases/infected/removed/dead. Yet, reproduction numbers before and after safety measures are computed and seem in line with other estimations. A stochastic model is introduced (the rate of infection follows stochastic variations) but not applied to the data.