With the R number across the UK and Ireland less than one, what lessons can be learned from the United States on re-opening safely?

The Northern Ireland Executive announced that there will be further loosening of lockdown restrictions next week, following the Department of Health announcing that the R number (the average number of people an infected person will pass the virus to) in Northern Ireland is currently between 0.7 and 0.9.

It appears that R is holding steady at less than one across the UK and Ireland. The chart above shows estimated values for R across the Republic of Ireland, Northern Ireland, Scotland, Wales and the English regions until the end of May (it excludes the sporadic cases found in the UK before the middle of February). The calculation looks at case numbers over a rolling seven-day period.

From the estimated regional R numbers, London was the first region to have R rise above 1, followed by the South East and the West Midlands of England shortly afterwards. Scotland and Northern Ireland appeared to be the last places where R rose above 1, in the week leading up to the 10^th of March.

Massive spikes in the R number can be seen in the South West of England around the start of March, and in Wales around the middle of March, suggesting that there could have been “super spreader” events around these times in these areas.

It can be seen from the data that the R number started to fall when lockdown restrictions were implemented across the UK and Ireland. It took around a month from the introduction of restrictions for the R number to fall below one, where it has (except for a brief period in mid-May in Wales) remained ever since.

The data seems to confirm that the lockdown measures did what they were intended to do; the R number started to fall as soon as they were introduced and now seems to be holding steady at less than one.

Whilst the events of March 2020 were shocking and unprecedented, from a policymaking perspective the correct course of action (implementing lockdown measures) was clear enough and had widespread public support.

The choices to be made from here onwards, as tentative steps are made towards re-opening the economy, will be much more difficult. It was always going to be the case that lockdown could not be sustained forever, and that extraordinarily difficult choices need to be made to balance protecting lives and protecting livelihoods.

We simply do not have enough data and information to know how safe it is, for example, to reopen schools. Or restaurants. There is now enough data to suggest that COVID-19 transmission is highly over-dispersed, i.e. transmission is caused by a comparatively small number of infected people infecting a large number of people.

This means that randomness and luck will play a significant role in determining outcomes as re-opening measures are implemented. We lack data on the relative safety of different re-opening measures.

Also, we are dealing with “fat tails”. One superspreading event, such as a house party, could infect many people and cause a cascading chain of infections that causes the R number to rise back above 1.  One could scarcely imagine a more difficult problem for policymakers to address.

Still, there is now enough data on regions that have reopened to provide some potential insight on what re-opening might bring. Much of this will come from the United States, where there has been a zeal to re-open the economy that has not been seen in other developed economies that have been badly impacted by the virus.

For example, take the state of Wisconsin. The chart below shows how cases (on the left hand side), and the associated R number (estimated using the EpiEstim app), have progressed since March.

Wisconsin is an interesting case. The state’s Democratic governor was extremely quick to react to the epidemic in the state, calling a state of emergency on the fourth day of there being new cases in the state (excluding one isolated case in January), when there were only seven cases in the state. Schools were closed the next day, with a stay at home order introduced eleven days after that. The R number is calculated over seven days, which is why the boxes in the chart above are seven days wide.

The R number in Wisconsin dropped below 1 around the 13^th of April, well before R was below one anywhere in the UK or Ireland. However, following determined efforts by Republicans in the state legislature, the state supreme court ruled by a 4-3 ruling that the stay at home order was unconstitutional.

With the state effectively re-opened, the R number has drifted above 1. There were a record 665 new cases on the 3^rd of June. By comparison, the Republic of Ireland, with a comparable population had 47, up from 10 the previous day.

The chart below shows how case numbers and the estimated R number has changed over the last three months in the state of Texas.

Texas has re-opened gradually, which can provide some insight into how R has changed in response to different policy decisions, albeit under different circumstances than the UK and Ireland because R never really dropped significantly below 1 at any point.

As is the case elsewhere, the R number dropped when the remain home executive order was implemented. When the order expired the R number increased above 1, and whilst it has fallen below 1 since then it appears to have increased above 1 in the last week following the reopening of offices and childcare, and bars and restaurants (albeit at reduced capacity). The 1,885 new cases identified in the Lone Star State on the 2^nd of June was a new record.

There are caveats with the case data in the United States. For example, the number of tests has increased over time, so some of the increase in cases could be attributable to this. However, the timing of R increasing at the same time as reopening policies does suggest that there could be a causal link between relaxing restrictions and the R number increasing.

Data from other parts of the world that have loosened lockdown restrictions are a key piece of the puzzle for policymakers seeking a safe way to restart the economy, alongside other policies such as rigorous contact tracing, universal mask wearing, and other unconventional data gathering efforts such as sewage monitoring.

Many difficult choices will have to be made in the months ahead in the battle against this terrible disease.