The tool was created by Vision Consulting Engineers to help users understand how the Covid-19 virus may spread through a population over time.
In particular, we want the public to be able to test different scenarios and understand why 'flattening the curve' is so important.
The Model is calculated using a statistical package in R called EpiModel (details provided in the 'Technical Information' tab).
Though there are a variety of modelling options that can be used, Vision chose one which used the R0 (R nought) values.
Covid-19 is still a young virus, so specific information on infection probability and exposure are limited.
However, the R0 is a value which can be observed through how quickly the virus has spread in places already infected (China, Italy, Japan, etc.).
In essence, the R0 is the measure of how many new infections will be created over time for each person currently infected in the population.
On the 'Model' tab you will see two different types of data to be entered, Rates and Population figures.
The following list provides a brief explanation of what each parameter means and some aspects to consider when entering the numbers.
R0 (R nought)
The relative rate at which the virus will spread; 1 would mean almost no transmission by infected, 5 would be considered high (such as Measles)
This is perhaps the most influential parameter for changing the curve, and is also the variable over which we have the most control.
In a population where good hygiene is observed, hands and surfaces are washed, and distances kept, the R0 can stay relatively low.
More broadly, early detection and quarantining infected is also effective at bringing the value down.
This represents the number of people that are expected to be hospitalised from the infection.
It is important to understand that this is not necessarily linked to the number of people who need hospital care.
Due to many factors such as demographic composition, cultural norms, or availability and affordability of treatment; not everyone who may need care will visit hospitals.
Intensive Care Rate
This represents the rate of people that are hospitalised which are likely to end up in Intensive Care.
This represents the number of infected that are likely to die from the infection.
Note that this is a core component of the model calculations, and not directly linked to the calculations for hospitalisation or intensive care.
It can be assumed that populations with older age composition, pre-existing conditions, and limited health care access will experience elevated mortality rates.
This is the total population that is at risk of being infected.
At current, Covid-19 is believed to be an 'equal opportunity' illness, meaning most of a given population can be considered susceptible.
However, tactics such as isolation can effectively reduce the population at risk, as they will be removed from the infection chain.
The model can technically be started at any stage of the infection, meaning a small or large number can be set as the starting count.
The total number of hospital beds available to the given population.
Remember that other illnesses, accidents, and medical events will continue to occur.
We recommend investigating what the remaining capacity is for the medical system after accounting for the existing demand.
Intensive Care Beds
The number of intensive care beds available for treating infected.
This indicator requires the same considerations as hospital bed capacities.
Time Scale Slider
By default, the model will start at day 0 and create results out up to 600 days.
Changing the start or the end date of the slider will change the view of the charts to only the selected day range.
This application was developed by Vision Consulting and Engineers using R Shiny and R Server.
The core modelling package used, EpiModel, was developed by members of Emory and the University of Washington.
New Zealand data on Covid-19 counts is sourced from the Ministry of Health website, while the figures for beds has been sourced from Otago University.
While the model used is a widely accepted method for modelling the spread of disease, it must be stressed that this application is intended only as a communication tool.
The primary object is to allow users to understand the importance of slowing the spread of Covid-19.
Figures and charts shown on this page should not be used for any form of official decision making.
A Deterministic Compartmental Models (DCM) was used for the underlying calculations, specifically the Susceptible Exposed Infected Recovered (SEIR) model.
This model assumes that once infected and recovered a person cannot be re-infected or infect others.
The 'Exposed' period is set to a default of 5 days. This is a non-infectious incubation period.
The 'Infected' period is set to a default of 14 days. This is the stage were an infected person will show symtoms and become infectious.
The number of assumed 'Exposed' at the start of the model is set to the total infected x the R0. This may be conservative, as
the current assumptions are that or for every infection detected, at least 2 to 10 have gone undetected.
Hospitalisation and ICU rates are not a core component of the SEIR model.
Vision has derived the curves by assuming the number of people transitioning into the infected state.