SIR model
The SIR model is the foundational compartmental model of infectious disease dynamics, dividing a population into three classes: Susceptible (individuals who can be infected), Infected (individuals who are infectious), and Recovered (individuals who have recovered and are now immune). Introduced by Kermack and McKendrick in 1927, the model describes how an infectious agent propagates through a population as individuals transition between compartments according to differential equations.
The classical SIR model assumes a well-mixed population of size N, with S(t) + I(t) + R(t) = N at all times. The dynamics are governed by two parameters: the transmission rate β (the rate at which susceptible individuals become infected through contact with infectious individuals) and the recovery rate γ (the rate at which infected individuals recover). The governing equations are:
- dS/dt = −βSI/N*
- dI/dt = βSI/N − γI*
- dR/dt = γI*
The ratio R₀ = β/γ is the basic reproduction number — the expected number of secondary infections produced by a single infected individual in a fully susceptible population. When R₀ > 1, the disease spreads exponentially in the early phase; when R₀ < 1, the infection dies out. The herd immunity threshold — the fraction of the population that must be immune to prevent sustained transmission — is 1 − 1/R₀.
Extensions and Limitations
The basic SIR model is a mean-field approximation: it assumes homogeneous mixing and ignores network structure, spatial heterogeneity, and individual variation. Extensions include the SEIR model (adding an Exposed compartment for latent infection), the SIS model (for diseases without lasting immunity), and network-based SIR models where transmission occurs only along edges of a contact network.
The model's predictions are sensitive to the well-mixed assumption. In structured populations — where individuals interact primarily with a small number of contacts — the effective R₀ can be much lower than the naive estimate, and the epidemic dynamics are governed by the network topology rather than the population-average parameters. Propagation of chaos — the statistical independence of individuals in the large-population limit — justifies the mean-field approximation for well-mixed populations but warns that it fails when contact structure matters.
The SIR Model as a Universal Mechanism
The SIR model is not merely an epidemiological tool. It is a universal mechanism for threshold-driven contagion processes. The same equations — with reinterpreted variables — describe the spread of rumors, the diffusion of innovations, the adoption of technologies, and the propagation of financial distress. In each case, there is a susceptible population, an infectious mechanism (transmission), and a removal mechanism (saturation, recovery, or abandonment). The mathematical structure — exponential early growth, peak, and decay — is invariant across substrates.
This universality is both a strength and a danger. The strength is that insights from epidemiology transfer to other domains: herd immunity thresholds exist for innovation adoption, and superspreading events in epidemics are analogous to viral cascades in social media. The danger is that the analogy can be overextended: financial contagion involves strategic behavior that the SIR model cannot capture, and opinion dynamics involve reinforcement and memory effects that the basic compartmental structure does not include.
The SIR model is the hydrogen atom of contagion dynamics: the simplest system that captures the essential physics of threshold-driven spread. It is wrong in every specific case — no real population is well-mixed, no real disease has constant rates, no real recovery confers perfect immunity — but it is right in the way that matters: it identifies the control parameter (R₀), the threshold condition, and the qualitative dynamics that every contagion process must obey.