Explaining rapid reinfections in multiple-wave influenza outbreaks: Tristan da Cunha 1971 epidemic as a case study

Camacho, Anton; Ballesteros, Sébastien; Graham, Andrea L.; Carrat, Fabrice; Ratmann, Oliver; Cazelles, Bernard

An extensively used epidemiological model for influenza dynamics is of susceptible -exposed -infectious -removed (SEIR) form [8, 11] . After exposure to the virus, susceptible hosts (S) pass through an exposed state (E) of latent infection, become infective (I) and are finally removed (R) from the infectious pool as they simultaneously recover and acquire permanent immunity. However, our immunological hypotheses motivate a more accurate description of the different stages from infection to development of long-term protective immunity. We incorporate several known [8, 11, 12, 17, 18] and novel features to the classical SEIR model in order to mechanistically translate the six biological hypotheses into six stochastic state-space models (see figure 2 and electronic supplementary material, text S1, for further details). Particular emphasis is given to ensure that each model combines enough parsimony to enable parameter inference and enough complexity to match unambiguously to a single hypothesis.