Glimmer of hope IV
The UK COVID-19 peak is in sight!
Part IV of Glimmer of hope. (10-April-2020)
Summary
To summarise today’s and the previous three blog posts. The UK COVID-19 peak looks to be in sight.
If you are elderly, or vulnerable, then this is the critical period. In my view the next ten, or so, days is the period to firmly self-isolate while the COVID-19 peak passes through. This is the time to use your food supplies. Stay out of shops. Get food delivered. Stay away from everyone. Stay safe.
United Kingdom
My forecast for peak COVID-19, during its first phase in the UK, is becoming clearer. Today I forecast that the peak (daily death rate) will occur in 5.6 days time (purple curve). However many pandemic trajectories (cyan curves) are still quite possible.
Data and model fits at left, forecasts at right. Red – Cases. Blue – Deaths. Brown – projected Cases. Purple – projected Deaths. Pink – illustrates range of plausible Cases trajectories. Cyan – illustrates range of plausible Deaths trajectories.
Model
My model suggests the UK is close to the peak in daily Cases, and most likely 5 to 6 days away from the peak in daily death rate. However, it also shows there are still great uncertainties. I estimate a peak daily death rate of 1109.
I should have mentioned somewhere in my previous three posts how the swathe of model curves and projections in the above diagram are generated. The classical statistical technique of bootstrapping is used. The multiple models are fitted using the maximum likelihood method. My full (generalised) 5-parameter model is:
Typically I only need the first three parameters (ie fix b3=0, b4=2). The longer Chinese deaths series needed all five parameters. I have found that my computer code runs quite happily in Rstudio on the Cloud. All the data that I have processed has been uploaded from the excellent European Centre for Disease Prevention and Control web-outlet.
M. King Hubbert
Here is a prophetic quotation from M. King Hubbert (1903 – 1989), whose maths has figured highly in this sequence of blogs.
“We don’t know how to cope with it. Our ignorance is not so vast as our failure to use what we know. Growth, growth, growth — that’s all we’ve known . . . ; human population growth is like nothing that has happened in all of geologic history”
I’m glad the data and insights were useful. Good specific predictions. And the test will be soon, after Thursday 16th – conditional that Easter Sunday does not bring a mass population breakout. Very encouraging.
Stuart, I’ll try to firm up even more on the predictions I plan to post this evening by adding 95% confidence limits.
Well done again. More than a glimmer now I think. Thanks for including the mathematics.
Bruce, Thanks for your original comment suggesting that a better description of the underlying maths would be useful.