The SVEIRS Epidemiological Model

🦠 The SVEIRS Epidemiological Model: A Popular‑Science Introduction

🧬 Abstract

Vaccines and fading immunity are central themes in modern public‐health campaigns.
To capture these processes mathematically, epidemiologists extend classic compartmental models with vaccination and waning immunity compartments.
The SVEIRS model – short for Susceptible–Vaccinated–Exposed–Infectious–Recovered–Susceptible – is one such extension.
It combines the incubation stage of the SEIR model with vaccination (V) and allows both vaccine‐induced and natural immunity to decay over time, returning individuals to the susceptible class.
This article introduces SVEIRS through an intuitive lens, discusses its assumptions and applications, explores key variants, and provides guidance on how this model informs real‐world vaccination policies.

🌍 1. Introduction

Compartmental models like SIR, SEIR and SIRS divide a population into categories based on disease status and describe how people flow between them.
The SIR model, for example, has three compartments – Susceptible (S), Infectious (I) and Recovered (R) – and assumes that recovered individuals become permanently immune.
Real diseases rarely behave so simply.
Some infections (e.g., pertussis and seasonal coronaviruses) confer only temporary immunity; others can be prevented or mitigated via vaccination.
When vaccination moves people from the susceptible to the recovered class, the fraction of the population that must be vaccinated to prevent an outbreak is at least 1 – 1/R₀.
Compartmental models can incorporate vaccination and waning immunity by adding a vaccinated compartment or tracking partially immune individuals separately.
The SVEIRS framework does exactly that.

📦 2. Model description

The SVEIRS model divides the population into five dynamic compartments (plus a return to S):

• Susceptible (S): individuals who can contract the disease.
  
  
• Vaccinated (V): individuals who have received a vaccine, gaining temporary immunity and possibly partial protection.
  
  
• Exposed (E): individuals who have been infected but are not yet infectious (latent period).
  
  
• Infectious (I): individuals who can transmit the disease.
  
  
• Recovered (R): individuals who have recovered and are temporarily immune.
  
  
• Back to Susceptible: both vaccine‐induced and natural immunity can wane, returning individuals to S.

The flow of people between compartments can be described by the following ordinary differential equations:

dS/dt = μ N − β S I / N − v S + ρ R + ω V − μ S

   (susceptibles)
dV/dt = v S − β_v V I / N − ω V − μ V

   (vaccinated)
dE/dt = β S I / N + β_v V I / N − σ E − μ E

   (exposed)
dI/dt = σ E − γ I − μ I

   (infectious)
dR/dt = γ I − ρ R − μ R + ω V

   (recovered)

🔍 Parameter definitions and typical values

Symbol	Description	Typical range	Notes
β	Transmission rate between susceptible and infectious individuals	0.3 – 1 per day	Higher β means more infectious contacts per unit time.
β_v	Transmission rate from vaccinated individuals (often β_v = c β with 0 ≤ c < 1)	0 – 0.5 β	Reflects vaccine efficacy in reducing transmission.
σ	Incubation (exposed → infectious) rate = 1/(average incubation period)	0.2 – 0.5 per day	Latent periods of 2–5 days give σ ≈ 0.2–0.5.
γ	Recovery rate = 1/(average infectious period)	0.05 – 0.3 per day	Infectious periods of 3–14 days give γ ≈ 0.07–0.33.
v	Vaccination rate (per capita)	0.01 – 0.1 per day	Fraction of susceptibles vaccinated each day; depends on campaign intensity.
ρ	Waning rate of natural immunity (R → S)	0.001 – 0.02 per day	Inverse of the duration of natural immunity (1 year → 0.0027/day).
ω	Waning rate of vaccine‐induced immunity (V → S or V → R)	0.002 – 0.05 per day	Booster campaigns often respond to waning over months or years.
μ	Per capita birth/death rate	0 – 0.00004 per day	Optional demographic turnover; can be set to zero for short epidemics.
N	Total population	Fixed	Sum of all compartments (S+V+E+I+R).

The equations capture the following processes:

1. Infection: susceptible and vaccinated individuals become exposed through contact with infectious individuals at rates β and β_v, respectively. β_v is smaller than β because vaccination reduces susceptibility and/or infectiousness.
   
   
2. Vaccination: susceptibles are vaccinated at rate v and move to V.
   
   
3. Incubation: exposed individuals progress to the infectious class at rate σ.
   
   
4. Recovery: infectious individuals recover at rate γ and move to R.
   
   
5. Waning immunity: recovered individuals lose natural immunity at rate ρ, and vaccinated individuals lose vaccine‐induced immunity at rate ω. Waning returns them to S (or in some models, to R).
   
   
6. Demography: births and deaths occur at rate μ, replenishing S with newborns and removing individuals from all compartments. For short epidemic time frames, μ is often set to zero.

🧪 Basic reproduction number and threshold

In the absence of vaccination and waning, the basic reproduction number for an SEIR model is R₀ = β/γ, which is the average number of secondary cases produced by a single infectious individual in a fully susceptible population.
With vaccination and waning, the effective reproduction number becomes more complex; it depends on the fraction of vaccinated and recovered individuals and the reduction in transmission due to vaccination.
For example, if vaccination reduces transmission by a factor c = β_v/β, the effective reproduction number in the presence of vaccination is roughly:

R_eff ≈ R₀ [ S/N + c (V/N) ].

An epidemic can be prevented when R_eff < 1, which can be achieved through sufficient vaccination coverage. For the SIR model, the vaccination coverage required to prevent an outbreak is at least 1 – 1/R₀.

🧠 3. Assumptions and applicability

The SVEIRS model builds on the assumptions of the SIR and SEIR frameworks and introduces additional considerations:

1. Homogeneous mixing: Individuals mix randomly; contact rates are uniform across the population. This is a common assumption in compartmental models but can be relaxed in network variants.
   
   
2. Fixed population size (optional): Births and deaths can be ignored over short time scales; if demography is included, births replenish the susceptible class at rate μ and deaths occur uniformly across compartments.
   
   
3. Temporary immunity: Both natural immunity after infection and vaccine‐induced immunity decay over time, at rates ρ and ω.
   
   
4. Vaccination is imperfect: Vaccines may not prevent infection entirely; vaccinated individuals can still become infected (albeit at reduced rates β_v).
   
   
5. Exposed individuals are not infectious: The E compartment reflects a latent period during which individuals are infected but cannot transmit the disease (incubation), as described in SEIR models.

🧾 When to use SVEIRS

The SVEIRS model is appropriate for diseases with:

• An incubation period (latent stage) and temporary immunity (natural or vaccine‐induced).
  
  
• Vaccination programs where coverage, waning protection and booster strategies matter.
  
  
• Waning immunity observed after infection or vaccination, as in pertussis, influenza, measles, and COVID‑19.
  
  
• Moderate to large populations where deterministic models approximate average behaviour well.

🔧 4. Variants and extensions

Like other compartmental models, SVEIRS serves as a building block for more sophisticated frameworks. Here are several common extensions:

4.1 SVEIR Model (no waning)

If natural immunity does not wane (ρ = 0) and vaccine‐induced immunity is lifelong (ω = 0), the model reduces to SVEIR. Vaccinated individuals remain protected indefinitely, and recovered individuals remain immune. This variant can be used for diseases where immunity is long lasting but vaccination roll‐out dynamics are important (e.g., measles in highly vaccinated populations). The equations simplify by omitting the waning terms (ρ = ω = 0).

4.2 SIRS / SEIRS / SVIRS

• SIRS: Susceptible–Infectious–Recovered–Susceptible; used for diseases with waning immunity but no latent or vaccine compartments.
  
  
• SEIRS: Adds an exposed stage to SIRS.
  
  
• SVIRS: Similar to SVEIRS but without an exposed stage (no latency); appropriate when the incubation period is negligible.

4.3 Age‑structured and network SVEIRS models

Real populations are heterogeneous: children, adults and the elderly have different contact rates and immune responses. Age‐structured SVEIRS models divide the population into age groups, each with its own parameters (βᵢⱼ, vᵢ, σᵢ, γᵢ). Network models replace homogeneous mixing with a contact network; vaccinated individuals may cluster, affecting herd immunity thresholds. Such models capture school/workplace structures and social networks.

4.4 Multi‐dose vaccination and booster models

Vaccines often require prime–boost dosing or boosters after waning. Multi‐dose models divide V into subcompartments (e.g., V₁, V₂, V₃ for first, second and booster doses) with different waning rates and efficacies. These variants help plan roll‐out strategies and booster schedules.

4.5 Time‑dependent parameters and seasonality

Transmission rates β and β_v may vary seasonally or due to new variants. Time‑dependent functions β(t) and β_v(t) allow the model to simulate waves of infection. Similarly, vaccination rates v(t) can reflect rollout campaigns, and waning rates ω(t) may differ by vaccine type.

4.6 Stochastic SVEIRS models

Deterministic models use average rates and assume large populations. Stochastic SVEIRS models incorporate random events, capturing uncertainty and extinction probabilities, especially in small populations or early outbreak stages. They can simulate the chance that an epidemic dies out even when R₀ > 1.

4.7 Spatial and metapopulation models

Epidemics often spread through multiple regions with varying vaccination coverage. Metapopulation SVEIRS models link several SVEIRS systems via migration or commuting flows. They are crucial for evaluating travel restrictions and vaccine allocation across regions.

🎯 5. Conclusion

The SVEIRS model unites vaccination programmes and waning immunity with the incubation dynamics of SEIR. It captures how vaccines reduce transmission, how immunity fades, and how booster campaigns shape long‑term disease trajectories. By adjusting parameters like β, σ, γ, v, ρ and ω, public‐health analysts can simulate scenarios ranging from rapid vaccine roll‑out to slow booster uptake and emergence of new variants.
The SVEIRS framework also highlights that no model is a perfect mirror of reality. Real populations are heterogeneous, contact patterns are structured, and immunity is complex. Nonetheless, simplified models like SVEIRS provide valuable insight into which control strategies are likely to succeed, and they help quantify the levels of vaccination needed to suppress outbreaks.
Future research will refine these models to incorporate individual variation, different vaccine types, and the dynamic evolution of pathogens.

📚 References

Anderson, R. M., & May, R. M. (1991). Infectious diseases of humans: dynamics and control. Oxford University Press.
Brauer, F., & Castillo‑Chávez, C. (2012). Mathematical models in population biology and epidemiology (2nd ed.). Springer.
Diekmann, O., & Heesterbeek, J. A. P. (2000). Mathematical epidemiology of infectious diseases: model building, analysis and interpretation. John Wiley & Sons.
El Khalifi, M., & Britton, T. (2023). Extending susceptible‑infectious‑recovered‑susceptible epidemics to allow for gradual waning of immunity. Journal of the Royal Society Interface, 20(206), 20230042. https://doi.org/10.1098/rsif.2023.0042
Hethcote, H. W. (2000). The mathematics of infectious diseases. SIAM Review, 42(4), 599–653. https://doi.org/10.1137/S0036144500371907
Kermack, W. O., & McKendrick, A. G. (1927). A contribution to the mathematical theory of epidemics. Proceedings of the Royal Society A, 115(772), 700–721.
Keeling, M. J., & Rohani, P. (2008). Modeling infectious diseases in humans and animals. Princeton University Press.
Park, S. W., Bolker, B. M., Funk, S., & Metcalf, C. J. E. (2019). Roles of generation‐interval distributions in shaping relative epidemic strength, speed, and control. PLOS Pathogens, 15(12), e1007815.
World Bank. (2021). An introduction to deterministic infectious disease models【537062538160086†L295-L340】.
U.S. Centers for Disease Control and Prevention (CDC). (2022). Technical explainer: Infectious disease transmission models【338848892261824†L176-L208】.
Lopez‑Herrero, M. J., & Taipe, D. (2024). Disease incidence in a stochastic SVIRS model with waning immunity (preprint). https://doi.org/10.48550/arXiv.2410.14883