/COVID-19-Making-Decision

Our goal in this project is to advise Italian policymakers on the optimal timeline for interventions to prevent hospital overcrowding, protect citizens by curbing the virus’s spread at the right time, and reduce public health costs for the government.

Primary LanguageJupyter Notebook

COVID-19-Making-Decision

R badge GitHub Jupyter Notebook

Team Members

Annette - Emmanuel A. - Jay - Oumarou - Shama - Tiwonge - Yvan

Title

The Impact of Timing of Non-Pharmaceutical Interventions (NPIs) on ICU Beds Occupancy During the 2020 COVID-19 Outbreak in Italy

non icu

Objective

Our goal in this project is to advise Italian policymakers on the optimal timeline for interventions to prevent hospital overcrowding, protect citizens by curbing the virus’s spread at the right time, and reduce public health costs for the government.

Research Question

How does implementation of non-pharmaceutical interventions on days 5, 10, 30, and 50 after the first detected case of SARS-CoV-2 in Italy affect ICU bed occupancy?

Dataset Description

The dataset containing historical data on COVID-19 cases, hospitalizations, and hospital bed occupancy has been sourced from the Italian National Institute of Health. Our dataset, named hospitalization.csv, includes 150 rows and 4 columns. Each row corresponds to a specific day starting from February 24, 2020 and includes the following columns:

  • Time: The date of the recorded data (daily).
  • Excluded_ICU_cases: The number of patients hospitalized in normal beds (non-ICU).
  • ICU_cases: The number of patients hospitalized in ICU beds.
  • Total_cases: The total number of hospitalized patients, combining both normal bed and ICU bed hospitalizations.

non icu icu

Model

The stochastic SEIR model will be implemented using historical data from Milan's COVID-19 outbreak. The parameters ($\beta$, $\sigma$, $\alpha$, $\theta$, $\gamma_1$, $\lambda_1$, $\delta$)

The model will simulate the progression of the pandemic over 150 days for the first wave, analyzing the impact of different non-pharmaceutical interventions (NPIs) on hospital bed capacity.

SEIR Model for Milan's COVID-19 Outbreak

Model Diagram

model

SEIR Model States and Parameters

States

Symbol Description
$S$ Number of susceptible individuals
$E$ Number of exposed individuals
$I$ Number of infectious individuals
$B_N$ Number of non-ICU hospitalized individuals
$B_{ICU}$ Number of ICU hospitalized individuals

Parameters

Symbol Description
$N$ Total population
$\beta$ Transmission rate
$\sigma$ Rate at which exposed individuals become infectious
$\alpha$ Rate at which infectious individuals are hospitalized
$\theta$ Rate at which non-ICU hospitalized individuals recover
$\gamma_1$ Rate at which non-ICU hospitalized individuals move to ICU
$\lambda_1$ Rate at which ICU individuals recover
$\delta$ Death rate in the ICU

Differential Equations

The model is governed by the following system of differential equations:

$$ \begin{align*} \frac{dS}{dt} &= -\beta \frac{SI}{N} \\ \frac{dE}{dt} &= \beta \frac{SI}{N} - \sigma E \\ \frac{dI}{dt} &= \sigma E - \alpha I \\ \frac{dB_N}{dt} &= \alpha I - (\theta + \gamma_1) B_N \\ \frac{dB_{ICU}}{dt} &= \gamma_1 B_N - (\lambda_1 + \delta) B_{ICU} \end{align*} $$

Usage

To run the model, follow these steps:

  1. Clone the repository: 
    git clone https://github.com/Vilmo18/COVID-19-Making-Decision
  2. Navigate to the project directory: 
    cd COVID-19-Making-Decision

Analysis

The simulation will provide insights into the impact of various NPIs on hospital bed capacity. Results will be visualized to show the progression of the pandemic under different scenarios.

Contributing

If you would like to contribute to this project, please fork the repository and create a pull request with your changes.

License

This project is licensed under the MIT License