A SEIR-SEI model calibration for the Zika virus epidemic in Brazil

Eber Dantas (1) , Michel Tosin (1) , Americo Cunha Jr (1)

1 UERJ - Universidade do Estado do Rio de Janeiro [Brasil] = Rio de Janeiro State University [Brazil] = Université d'État de Rio de Janeiro [Brésil]

The Zika virus has been reported as autochthonously transmissible in over 30 countries and has reached more than 200,000 probable cases in Brazil. Thus, the use of mathematical models for epidemics become of great importance, since they are useful tools to predict the underlying outbreak numbers and allow one to test the effectiveness of different strategies used to combat the associated diseases. In previous works (Dantas, 2017; Tosin, 2017), the authors have identified reasonable parameters and initial conditions (IC) that fit a SEIR-SEI model to epidemic data of the Brazilian outbreak, by the deterministic solution of an inverse problem. This work uses Bayesian Inference to explore the uncertainties associated to the initial number of infectious humans, an IC that lacks reference data on the literature related to the epidemic, updating the associated stochastic model with experimental data provided by the Brazilian Ministry of Health. Posterior probability density functions are generated assuming uniform and gamma prior distributions over different conditions.

Domaines: Systèmes dynamiques [math.DS]; Bio-informatique [q-bio.QM]; Santé publique et épidémiologie; Probabilités [math.PR]; Statistiques [math.ST]

Eber Dantas, Michel Tosin, Americo Cunha Jr. A SEIR-SEI model calibration for the Zika virus epidemic in Brazil. 4th International Symposium on Uncertainty Quantification and Stochastic Modeling (Uncertainties 2018), Apr 2018, Florianópolis, Brazil. ⟨hal-01711056⟩