Channel 4 Dispatches: Earlier lockdown modelling

6 May 2020

—–

This modelling and the results below were undertaken for Channel 4 Dispatches “Coronavirus Did the Government Get It Wrong?“, which was aired on the 3rd of June 2020.

—–

In mid-March, the number of reported Covid-19 infections were doubling every two- to three-days. So, 225 reported infections on the 10th of March quickly became 422 on the 12th of March taking the running total to 1,580.

The UK lockdown was announced on Monday the 23rd of March.

An additional 2,016 lab confirmed infections were reported on the lockdown day taking total reported infections up to 11,087.

Not everyone that catches the Covid-19 disease is tested, so reported figures are an underestimate of total infections. For example, on the 29th of March we estimated that up to 1.6 million people had the infection – this contrasts with national figure of 25,439 reported infections.

Anticipatory behaviour and perhaps concern over the seriousness of Covid-19 led to people social distancing ahead of the lockdown. But the lockdown remains the most effective measure at reducing the spread of the disease. (See chart below.)

Estimates differ, but between 0.5% and 1% of the people that are infected with the Covid-19 disease die. The progression of the disease means that it takes around two weeks from infection until death. This combined with some post-lockdown spreading of the disease in households and care homes means there is a lag between lockdown, hospital admissions and peak deaths.

Peak deaths occurred around the middle of middle of April. There are some oddities with the reporting of deaths that make it hard to give a precise date and number, but as of the end of April total reported deaths were 26,771 of which 21,031 were in hospitals.

There is a possibility that an earlier lockdown would not have been taken as seriously by the population, particularly without strong enforcement or the same level of credibility that came with 359 deaths. It may also have been harder to sustain in the long term, although with fewer people infected it may not have been necessary to maintain the lockdown for as long.

Our modelling of the impact of an earlier lockdown is set out below. This assumes that anticipatory social distancing, as with the actual lockdown, happened ahead of the lockdown. Had this not happened then the impact may have been lower. Based on this assumption, our estimates for the impact of an earlier lockdown on reduced hospital deaths relative to the 21,031 hospital deaths are as follows:

March 20th, 2020

• Estimated deaths of 18,992
• Estimated reduction of 2,039

March 16th, 2020

• Estimated deaths of 12,617
• Estimated reduction of 8,414

(earlier estimates are inherently hard to model accurately due to uncertainty about people’s willingness to social distance or enter lockdown)

These estimates are for hospital deaths only. So may underestimate lives that could have been saved in the community, although currently less is known on the spread of the disease in the community due to the lack of testing.