Mathematical Properties of Strategies to Control Epidemic Outbreaks in the Context of SEIR Models with Multiple Infectious Stages

Abstract

In this work we analyze mathematically the consequences and effectiveness of strategies to control an epidemic in the framework of classical SEIR models with multiple parallel infectious stages. We define the mathematical concept of a control strategy, showing that it implies turning classic epidemiological models into systems of non-autonomous differential equations. The analysis of these non-autonomous systems is based on the two main results obtained in this work: the first establishes a condition that implies a dynamic without epidemic outbreaks; the second establishes a maximum value for the susceptible population associated to the fixed points that are attractors, moreover, we proof that any trajectory converges to some of these attractors. An important consequence of this last result is the existence of an insurmountable limit on the number of infected individuals after the end of a given control strategy. This restriction can only be mitigated by changing the maximum value of susceptible population associated to the system attractors, which could only be done with permanent control action, that is, without returning to normality. Another interesting result of our work is to show how the moment to start and the way how the control strategy ends strongly impacts the asymptotic value for the total number of infected individuals. We illustrate our analysis and results in a SEIR model (with two or three parallel stages) applied to describe the COVID-19 epidemic.