Covid 19 Spread Modelling
This page contains details of how SDMA requires the support of data scientists to help in modeling the spread of Covid 19 in Kerala.
Goal: To create a data driven decision-making framework for public authorities to actively intervene and manage the spread of COVID19 outbreak.
COVID-19 cases will soon explode into unprecedented levels and we need to flatten the epidemic curve to reduce the burden on our current healthcare system. In order to facilitate effective response, we require our data-science volunteers to help us model the epidemic curve for COVID-19.
"To minimize the impact of the coming pandemic, we need to slow its spread and thereby keep our healthcare systems from being overwhelmed. Through aggressive measures we can flatten out the epidemic curve, keeping the number of people simultaneously infected at a low enough level to be manageable." - Carl Bergstorm, Professor of Biology, University of Washington
“If you look at the curves of outbreaks, they go big peaks, and then come down. What we need to do is flatten that down.” - Anthony Fauci, Director of the National Institute of Allergy and Infectious Diseases
Tentative Modelling Task List
Evolution curves will be used for decisions related to capacity planning
Predict estimated time to breach of capacity from current day
ICU Capacity
Hospital Beds
Medical supplies
How fast is COVID-19 outbreak growing?
How many cases should we expect in next days/weeks?
Plan the prioritization of hospital supplies/facilities to diffuse crowding
Predict when COVID-19 cases peak and start receding
Identify current phase of the outbreak
Dataset Specs (In Progress)
Requirements for data collection from each location for modelling purpose
Mandatory Patient Data to be collected while admitting
Day of occurence of symptoms
Admission date
Dates of exposure / infection
Severity of condition - Critical/Moderate/Mild
Dates of outcome: death / recovery
Metadata on patient: age, gender, location, occupation, etc
Contact data: exposure (who infected you?) and Contact tracing (who could you have infected?)
For Modelling Epidemic Curve
Data collection need to be done in this format
Current Hospital Capacity Data
Require current available hostpital beds in each location
Require current available ICU capacity for each location
Estimate of supplies needed to treat one person in
Critical condition who needs an ICU bed
Moderate condition who needs a hospital bed
Mild or Stay-at-home condition
List of items require to treat a person
Data Scientist Reading Resources (In Progress)
TODO
We need to issue call for Data Scientists
We need to collect dataset of reported cases
Need to create a format for dataset and a central repository which can be accessed publically
We need to identify appropriate dataset to be passed to multiple teams for modelling
Consult with SME to identify suitable geographical granularity at which epidemic curves needs to be created
Create predictive models with Data Scientists
Apply this to a spatial data set and visulize future projections
Hypotheses testing of "what-if" scenarios
Suggestions:
Model to predict when elective surgeries can start.
Last updated