The coSIR model predicts effective strategies to limit the spread of SARS-CoV-2 variants with low severity and high transmissibility

Longchen Xu, Haohang Zhang, Hengyi Xu, Han Yang, Lei Zhang, Wei Zhang, Fei Gu, Xun LanNonlinear 2021 · DOI

Abstract

The online version contains supplementary material available at 10.1007/s11071-021-06705-8.

Introduction

SARS-CoV-2 is a dynamic pathogen. As it spreads, it mutates into variants with varying degrees of transmissibility and severity. Traditional epidemiological models, such as the standard SIR model, often treat viruses as static entities, failing to account for how public health policies might influence the evolutionary trajectory of the virus. This paper addresses this gap by introducing the coSIR (competitive SIR) model to study the competition between viral strains under different policy interventions.

Methodology: The coSIR Model

The researchers developed the coSIR model by extending the SIDARTHE framework. Unlike the standard SIR model, coSIR categorizes infected individuals into five distinct subgroups for each competing strain based on symptom severity and detection status:

  • $i_j$: Asymptomatic, undetected.
  • $d_j$: Asymptomatic, detected (via testing).
  • $a_j$: Ailing (symptomatic), undetected.
  • $r_j$: Recognized (symptomatic), detected.
  • $t_j$: Threatened (life-threatening symptoms, detected).
Schematic of the coSIR model illustrating the transitions between susceptible (S), infected (I), and recovered (R) states, including the internal sub-states for two competing strains.

The model utilizes a system of Ordinary Differential Equations (ODEs) to simulate the competition between two strains. It also accounts for partial cross-immunity through a "risk reduction rate" ($\omega$), which represents the reduced likelihood of reinfection after exposure to a different strain.

Key Insights and Results

The study identifies a "Stealth Advantage" for variants with lower severity. Because these strains produce fewer symptoms, infected individuals are less likely to be detected and isolated, allowing the virus to circulate longer and reach more hosts. This leads to a "Mortality Paradox": a variant that is significantly less lethal but more transmissible can cause a much larger outbreak and more total deaths than a highly lethal but slow-moving strain.

Simulation results demonstrating the impact of Cure and Screen policies on variant competition (b, d), the relationship between infection levels and severity (f), and the temporal progression of different infection states (h).

Key findings include:

  • Evolutionary Direction: The model predicts SARS-CoV-2 will likely evolve toward a "flu-like" state—highly transmissible but with lower individual pathogenicity.
  • Policy Robustness: "Block" policies (e.g., masks, social distancing) are the most effective across all variant types.
  • Policy Synergy: Combining "Isolation" (quarantining the sick) and "Screening" (mass testing) can match the effectiveness of lockdowns without the same economic impact.
  • Mobile Cabin Hospitals: These facilities are highlighted as effective "dual-threat" interventions that simultaneously increase isolation and cure rates.

Limitations and Future Work

The model assumes a well-mixed population and significant cross-immunity between strains. Future research could extend the model to include vaccination dynamics, age-stratified populations, and more complex feedback loops between policy changes and hospital capacity.

Summary updated

Abstract

The online version contains supplementary material available at 10.1007/s11071-021-06705-8.
BibTeX
@inproceedings{xu2021cosir,
  author = {Longchen Xu and Haohang Zhang and Hengyi Xu and Han Yang and Lei Zhang and Wei Zhang and Fei Gu and Xun Lan},
  title = {{The coSIR model predicts effective strategies to limit the spread of SARS-CoV-2 variants with low severity and high transmissibility}},
  booktitle = {Nonlinear},
  year = {2021},
  doi = {10.1007/s11071-021-06705-8},
}
Figure