Ordinary Differential Equations (14): Epidemic Models and Epidemiology

In early 2020, a novel coronavirus swept across the globe, and mathematical models became key tools for understanding the pandemic and formulating policies. When epidemiologists predicted that "if no measures are taken, the number of infections will grow exponentially," they were relying on differential equations. In this chapter, we will start from the simplest SIR model and gradually build a mathematical framework for understanding infectious disease transmission, validating these models' predictive power with real data.

The SIR Model: Cornerstone of Epidemiology

Basic Assumptions

The SIR model divides the population into three categories: - S (Susceptible): Not yet infected but can be infected - I (Infectious): Infected and contagious - R (Recovered/Removed): Recovered (immune) or deceased

Core assumptions: 1. Total population remains constant (ignoring births, deaths, migration) 2. Homogeneous mixing: Each person has equal probability of contact with others 3. Lifelong immunity after infection 4. Constant transmission and recovery rates

Model Equations

Let,,represent the number of people in each category at time:

Parameter meanings: -: Transmission rate (effective contacts per infectious person per unit time) -: Recovery rate (is the average infectious period) -: Total population

Basic Reproduction Number

Definition: (basic reproduction number) is the average number of people one infected person can infect in a fully susceptible population.

Key threshold theorem: - If: Epidemic will break out - If: Epidemic will naturally die out - If: Critical state

This is one of the most important results in epidemiology!

Effective Reproduction Number

During an epidemic, not everyone is susceptible. The actual reproduction number is:Whendrops below,, and the epidemic begins to recede. This is the herd immunity threshold:For example, if, we needof the population to be immune.

The SEIR Model: Adding the Latent Period

Why Do We Need a Latent Period?

Many infectious diseases have an incubation period— the time from infection to becoming infectious. For example: - COVID-19: 2-14 days (median 5 days) - Influenza: 1-4 days - Measles: 7-14 days - Ebola: 2-21 days

SEIR Model Equations

Introducing a new category E (Exposed): Infected but not yet infectious

New parameter: -: Rate of transition from exposed to infectious (is average latent period)

COVID-19 Modeling: Extensions of SEIR

Special Challenges of COVID-19

COVID-19 brought new challenges to modeling: 1. Asymptomatic transmission: A large proportion of infected individuals are asymptomatic but contagious 2. Reporting delays: Significant delay from infection to confirmed diagnosis 3. Parameter uncertainty: Early estimates ofand other parameters were inaccurate 4. Interventions: Quarantine, masks, vaccines changed transmission dynamics

Extended SEIR Model

Model considering asymptomatic cases: Where: -: Asymptomatic infected -: Symptomatic infected -: Proportion asymptomatic -: Relative infectiousness of asymptomatic cases

Mathematical Analysis of Interventions

Quarantine and Social Distancing

Quarantine measures reduceby lowering contact rates:whereis the proportion of contact reduction.

Goal: Make, requiring:For example, if, we need to reduce contacts by at least.

Vaccination

Vaccines convert susceptibles to immune at a certain rate: whereis the vaccination rate.

Herd immunity: When the immune fraction reaches, even without continued vaccination, large-scale outbreaks won't occur.

Network Epidemic Spread

Why Do We Need Network Models?

The SIR model assumes homogeneous mixing, but in reality: - Social networks are highly heterogeneous - "Super-spreaders" exist - Community structure affects transmission

Effect of Degree Heterogeneity

On a network with degree distribution,is modified to:For scale-free networks (),may diverge, leading to— disease can almost always spread!

Summary

In this chapter, we learned the core content of infectious disease modeling:

1. SIR model: The cornerstone of epidemiology, describing susceptible-infected-recovered dynamics
2. Basic reproduction number: The key threshold determining whether an epidemic can break out
3. SEIR model: Adding incubation period, better matching many real diseases
4. COVID-19 modeling: Handling complex factors like asymptomatic transmission and interventions
5. Parameter estimation: Fitting model parameters from data
6. Network models: Considering heterogeneity of social contact structure
7. Spatial spread: Geographic diffusion of epidemics

These models, though simplifications of reality, provide powerful tools for understanding and controlling infectious diseases. As the COVID-19 pandemic demonstrated, mathematical models play an increasingly important role in public health decision-making.