[Paper Review] Modelling Covid-19 epidemic in Mexico, Finland and Iceland

Rafael A. Barrio, Kimmo Kaski+3 more|arXiv (Cornell University)|Jul 18, 2020

COVID-19 epidemiological studies27 references4 citations

TL;DR

This study proposes a spatially explicit SEIRS model with stochastic mobility to simulate and predict the COVID-19 epidemic in Mexico, Finland, and Iceland. By coupling local compartmental dynamics with inter-regional transmission via human mobility, the model accurately reproduces observed case trajectories and forecasts the impact of reopening, revealing that countries with strict early confinement and testing (Iceland, Finland) face smaller second waves than Mexico, where widespread early transmission reduced susceptibility.

ABSTRACT

Over the past two decades there has been a number of global outbreaks of viral diseases. This has accelerated the efforts to model and forecast the disease spreading, in order to find ways to confine the spreading regionally and between regions. Towards this we have devised a model of geographical spreading of viral infections due to human spatial mobility and adapted it to the latest COVID-19 pandemic. In this the region to be modelled is overlaid with a two-dimensional grid weighted with the population density defined cells, in each of which a compartmental SEIRS system of delay difference equations simulate the local dynamics (microdynamics) of the disease. The infections between cells are stochastic and allow for the geographical spreading of the virus over the two-dimensional space (macrodynamics). This approach allows to separate the parameters related to the biological aspects of the disease from the ones that represent the spatial contagious behaviour through different kinds of mobility of people acting as virus carriers. These provide sufficient information to trace the evolution of the pandemic in different situations. In particular we have applied this approach to three in many ways different countries, Mexico, Finland and Iceland and found that the model is capable of reproducing and predicting the stochastic global path of the pandemic. This study sheds light on how the diverse cultural and socioeconomic aspects of a country influence the evolution of the epidemics and also the efficacy of social distancing and other confinement measures.

Motivation & Objective

• To model the geographical spread of COVID-19 across countries with vastly different demographics, policies, and social structures.
• To evaluate the effectiveness of social distancing, confinement, and testing policies in controlling pandemic spread.
• To predict the impact of reopening mobility on epidemic resurgence, particularly the risk of a second wave.
• To separate biological transmission parameters from mobility-driven spatial dynamics for improved forecasting.

Proposed method

• The model uses a 2D grid overlaying each country, with cells weighted by population density to simulate local SEIRS dynamics via delay difference equations.
• Infection spread between cells is modeled stochastically, representing human mobility as a driver of geographical transmission (macrodynamics).
• The model decouples biological parameters (e.g., incubation and recovery rates) from mobility parameters (e.g., travel frequency and routes).
• It incorporates time-varying mobility to simulate policy changes, such as lockdowns and reopening.
• The model was calibrated using early epidemic data and validated against observed case counts in Mexico, Finland, and Iceland.
• Scenario simulations were run to predict outcomes under full mobility restoration, comparing the three countries.

Experimental results

Research questions

• RQ1 How do differing national policies on social distancing and testing affect the trajectory of the COVID-19 pandemic in countries with distinct demographics?
• RQ2 To what extent can a stochastic spatial model with mobility accurately predict real-world epidemic dynamics with limited data?
• RQ3 What is the risk of a second wave following the reopening of international and domestic mobility?
• RQ4 Why do countries with similar early responses (e.g., Finland and Iceland) show different rebound patterns after reopening?

Key findings

• The model accurately predicted the epidemic trajectory in Mexico using only early April 2020 data, despite limited information.
• Iceland and Finland, with strict early confinement and testing, showed minimal rebound when mobility was restored, due to low population susceptibility.
• Mexico, where the first wave infected nearly all, showed an extremely small rebound upon reopening, as most were already immune.
• Finland experienced the largest rebound upon reopening, as a large portion of the population remained susceptible due to effective early containment.
• The model demonstrates that timely, well-followed confinement and testing measures significantly reduce the risk of a severe second wave.
• The study confirms that mobility-driven spatial models with stochastic transmission are effective for forecasting even with sparse data, especially when calibrated early.