How many people are actually infected with Covid-19?
Published 23 March 2020
The lack of social distancing seen in public this weekend suggests that the majority of people are not yet taking social distancing too seriously. This is undoubtedly the case on the London Underground where there have been photos of people crammed into carriages - approximately 120 by our count. Tonight, the government has announced a range of measures to ensure (and enforce) greater social distancing.
The official count of individuals infected with Covid-19 is heavily dependent on testing strategies and will be a significant underestimate of the actual number of cases. So how likely is it that you will be queuing next to, or sharing, a tube carriage with, someone that is infected?
Without a substantial increase in testing, it is hard to get an actual number of total infected people. The most reliable proxy measure is the crude mortality rate and knowledge of how quickly the disease is spreading. The chart below shows the number of deaths by country (data from 22nd of March).
The evidence is that the crude mortality rate is similar across countries at about 1 in 100 (1%).* A growth rate of 22% per day, slightly below the rate seen recently would mean that cases double every 3.5 days. Or four times every two weeks. That suggests there are around 1,600 people infected with Covid-19 for each reported death (there being a two week lag)**. Taking into account those that have recovered or died, this would suggest that there are currently 531,200 people infected with Covid-19 in the UK - 1 in 125 people.
These numbers are probably higher in London. The evidence suggests mortality is lower (e.g. 0.5% which is the case for younger people) and the speed of infection is higher (e.g. testing policy means fewer identified diseases). Based on this and current reported deaths in London, the number of people infected could be around 800,000***. Or 1 in 11 people - so 11 people on any crowded tube carriage.
The chart below summarises the number of cases by area and relative growth rates.
This contrasts to the situation in China, which was picked up in an earlier post, where strong enforcement came into effect more rapidly. ----
* There is no clear evidence that mortality is different between countries when their health services function effectively - most of the reported differences are due to differences in testing (e.g. community-based in Germany, hospital-based in the UK).
** The average time from infection to death is two weeks and with a crude mortality rate of 1%, it means that one death today resulted from 100 people becoming infected two weeks ago. A growth rate of infections of 22% means a doubling of cases every 3.5 days, so x4 every 14 days. That means within two weeks 100 people grow to 200 to 400 to 800 and then to 1600.
*** Increasing the rate of growth to 28% per day means that cases double every 2.8 days, so x5 every 14 days. This results in 100 growing to 3,200 in 14 days. This increases further if we assume the mortality rate is lower at 0.5%. Based on this, the ratio of deaths is 1 to 6,400. This does not take into account the people that have recovered (= people infected two weeks ago).