Abstract:Transmission rates in epidemic outbreaks may vary over time depending on the societal response. Non-pharmacological mitigation strategies such as social distancing and the adoption of protective equipment aim precisely at reducing transmission rates by reducing infectious contacts. To investigate the effects of mitigation strategies on the evolution of epidemics, nonlinear transmission rates that are influenced by the levels of infections, deaths or recoveries have been included in many variants of the classical SIR model. This class of models is particularly relevant to the COVID-19 epidemic, in which the population behavior has been affected by the unprecedented abundance and rapid distribution of global infection and death data through online platforms. This manuscript revisits a SIR model in which the reduction of transmission rate is due to knowledge of infections. Through a mean field approach that assumes individuals behave like molecules in a well-mixed solution, one derives a time-varying reproduction number that depends on infection information through a negative feedback term that is equivalent to Holling type II functions in ecology and Michaelis-Menten functions in chemistry and molecular biology. A step-by-step derivation of the model is provided, together with an overview of methods for its qualitative analysis, showing that negative feedback structurally reduces the peak of infections. At the same time, feedback may substantially extend the duration of an epidemic. Computational simulations agree with the analytical predictions, and further suggest that infection peak reduction persists even in the presence of information delays. If the mitigation strategy is linearly proportional to infections, a single parameter is added to the SIR model, making it useful to illustrate the effects of infection-dependent social distancing.

A feedback SIR (fSIR) model highlights advantages and limitations of infection-dependent mitigation strategies