How to Flatten the Curve: Should We Communicate on Incidence or Prevalence to Encourage Self-Isolation?

In this talk, I will first present an ODE model of an infectious disease coupled with self-isolation. People are assumed to compare payoffs to both strategies (isolation or not) once they learn they are infected, on a game theory basis. The payoff not to self isolate is correlated with two metrics commonly used to describe the dynamics of a disease: prevalence and incidence. Next, I will detail the equilibria of the model and the conditions for their local stability. Surprisingly, any linear combination of such data lead to the same prevalence at equilibrium. However, in some contexts, one of them is clearly better than the other during the transcient phase. These results also hold faily well for the stochastic counterpart of the ODE model, ran on empirical networks.