When the Wuhan outbreak was brought under control, it turned out that around 86,000 registered infections had resulted in around 4,700 deaths. In week 11 of the pandemic (around March 18), I roughly estimated that the expected percentage of fatal infections in the Netherlands could be around 5%. Is that correct? Why didn’t I know and why couldn’t I know for sure?
2. Uncertainties
In 2020 a lot is known about pandemics. A generally accepted approach to understanding how a pandemic will turn out in populations is to use SIRD models based on compartments: individuals belong to either the S category (say: Susceptible) or the I category (say : Infective), or to the R category (say: Resistant), or have died of the infection (say: Deaths). Wikipedia provides a structural sketch of how the numbers develop over time (see Fig. 1).
There is a lot to criticize about the S-I-R-D model, but it has proved useful in many cases. It is therefore obvious that during the first period of the COVID-19 pandemic, I sought its guidance. It’s a clear model, but it also leads to continued panic. Because it is clear that the model eventually leads to a situation in which everyone is either resistant or dead. Then, everyone is or has been infected. In the Netherlands alone, with an expected number of 5% fatal infections, I expect this to lead to 850,000 COVID deaths. (When I adopt an estimation of 0.7% deadly infections, that number still amounts to a substantial 120,000 fatalities in the Netherlands.)
But those numbers can be way too high, especially when defensive actions are thought of and implemented.
Such can be deduced from another source of knowledge that became available in week 11 (around March 17): reporting on the experiences in Wuhan, where the pandemic had already been almost fully brought under control in week 11. In China, with a population of around 1.4 billion, there were about 86,000 infections and 4,700 COVID deaths at the time.
The conclusions that could be drawn from those experiences are the following nine:
- the reproduction number of the virus in a still uncontaminated area (R0) is around 2
- hygiene + lock-down + social distancing + quarantines + tests + masks are effective defensive actions against the pandemic
- so it is possible to get the virus under control
- even when some infected people may remain asymptomatic
- even when asymptomatic infected people can spread the virus
- there are large differences in vulnerability per age category
- IC (intensive care) treatments often require ventilators (breathing machines)
- there are no effective medicines available (yet)
- no effective vaccines are available (yet)
I will refer to them as the nine preliminary conclusions. I considered them defensible, then (in week 11) and now (in week 42).
This is the situation as I found it in the Netherlands in mid-March. At the time when the Dutch Prime Minister held his press conference and declared an intelligent lock-down.