How long will the Corona lockdown last?
An app to predict lockdown time & the peak of crisis
Going by current rate a major health crisis awaits us in May. Millions of people would fall critically ill at the same time. Social distancing helps but the number of cases will still overwhelm hospitals.
Why “lockdowns”?
Corona cases are doubling every 3 to 4 days. That is exponential growth. Take “rice and the chessboard” analogy: if you put one grain of rice on the first square, two on the second, four on the third, and keep doubling them on subsequent squares you will need 18 billion-billion grains by 64th square. That is about 400 times higher than global rice production.
The Corona crisis can’t go that crazy because human hosts are limited by population. From the same growth principle, however, we can see how 500 thousand cases will become 16 million in just two weeks (R0 of 3.3–5). Around 6% of them can become critically ill. No country in the world has the capacity to help them at once.
What does a lockdown mean?
It means, social distancing is to be taken seriously — minimizing contacts by four time or more. Only then the number of critically ill patients come close to the handling capacity.
The Mathematical model for the Corona App
It uses compartmental epidemiological model used for viral outbreaks, SEIR. Solving ODE over the model estimates the lifecycle of the Corona outbreak.
In the model:
sd is social distancing factor
N is the population
beta is how often a contact results in new infection
gamma is the rate an infected recovers
sigma is the rate at which an exposed person becomes infective
The social distancing factor is modulating beta parameter, which can be used to implement change in relative social behavior. In the App you can try the best fit between the “Expected” data points given by model and the real “Confirmed” data points, and then move leftwards. How much you had to move left relative to the best-fit slider position indicates how much more we need to social-distance ourselves.
Fatality rate ranges from 1–3%, which is at most a guesstimate right now. The rate of a critical illness depends on health indicators as shown in this extensive study published in The Lancet. The parameters in the model need constant revisions but they are still useful for a gross qualitative estimation. And preparing for semi-worst is not a bad idea.
How to use the Corona App to calculate lockdown and peak infection date?
- Open the left sidebar (if it is hidden)
- Select your lock-down date (or any later stage date)
- Move the contacts slider to bring the “Confirmed” & “Expected” closer
- Insert the number of acute care units (ICUs) available in your country
- Insert critical illness rate. This is difficult because we do not know the number of actual cases (50% can be asymptomatic). In literature it swings from 0.7 to 6.1 % (about half needs ICU).
- On the main window, you can see the peak date with expected infections and the number of critically ill people.
- “Interaction” factor is calibrated to Italy before the lockdown, ~19. (see how this compares with that of yours after lockdown). The later side of the peak date gives an estimate of lockdown time.
- Move the “Interaction” slider to left and observe the relative change compared to the best fit from step 3. This relative change gives you an idea, how much more we should social distance.
Some countries have chosen not to test most of their residents, which means there will be a large disparity between ‘Expected’ count given by the mathematical model and manually ‘Confirmed’ count. This, however, will not change the number of critically ill people coming to hospitals.
Stay safe everyone!
__________________
