Government publishes R estimates for English regions

The Government has today (Friday 12 June) published regional R values for England for the first time.

R estimates, published today by the Government Office for Science, cover each of the NHS England regions in England. The average R value for the UK, as a whole, remains at 0.7-0.9.

As an average value, R can vary in different parts of the country, communities, and subsections of the population. It cannot be measured directly so there is always some uncertainty around its exact value. This becomes even more of a problem when calculating R using small numbers of cases, either due to lower infection rates or smaller geographical areas. This may be due to the uncertainty of the underlying data leading to a wider range for R. 

When there are a small number of cases estimates of R become less robust and less useful in determining the state of the epidemic. Estimates of R for geographies smaller than regional level are less reliable and it is more appropriate to identify local hotspots through, for example, monitoring numbers of cases, hospitalisations, and deaths. R values are shown as the range,  and the most likely estimate is in the middle of this range.

What is R?

The reproduction number (R) is the average number of secondary infections produced by 1 infected person.

An R number of 1 means that on average every person who is infected will infect 1 other person, meaning the total number of new infections is stable. If R is 2, on average, each infected person infects 2 more people. If R is 0.5 then on average for each 2 infected people, there will be only 1 new infection. If R is greater than 1 the epidemic is growing, if R is less than 1 the epidemic is shrinking.

R can change over time. For example, it falls when there is a reduction in the number of contacts between people, which reduces transmission. It is not an exact number, it is a calculated estimate.

R is not the only important measure of the epidemic. R indicates whether overall the epidemic is trending towards getting bigger or smaller but not how large it is. The number of people currently infected with coronavirus (COVID-19) – and so able to pass it on – is very important.

R should always be considered alongside the number of people currently infected. If R equals 1 with 100,000 people currently infected, it is a very different situation to R equals 1 with 1,000 people currently infected.

How R is estimated

Individual modelling groups use a range of data to estimate R including:

• epidemiological data such as hospital admissions, ICU admissions and deaths – it generally takes two to three weeks for changes in R to be reflected in these data sources, due to the time between infection and needing hospital care

• contact pattern surveys that gather information on behaviour – these can be quicker (with a lag of around a week) but can be open to bias as they often rely on self-reported behaviour

•household infection surveys where swabs are performed on individuals which can provide estimates of how many people are infected – longitudinal surveys (which sample the same people repeatedly) allow a more direct estimate of the growth in infection rates

Different modelling groups use different data sources to estimate R using mathematical models that simulate the spread of infections. Some may even use all these sources of information to adjust their models to better reflect the real-world situation. But there is uncertainty in all these data sources, which is why R estimates can vary between different models, and why we do not rely on one model; evidence from several models is considered, discussed, combined, and R is presented as a range. The most likely true value is in the middle of this range.

Who estimates R?

R is estimated by a range of independent modelling groups based in universities and Public Health England (PHE). The modelling groups discuss their individual R estimates at the Science Pandemic Influenza Modelling group (SPI-M) - a subgroup of SAGE. Attendees compare the different estimates of R and SPI-M collectively agrees a range which R is very likely to be within. 

Why does R become less useful?

There are three main reasons for this:

• Firstly, when there are few cases, R is impossible to estimate with accuracy and will have wide confidence intervals that are likely to include 1. This does not necessarily mean that the epidemic is increasing but could be the result of greater uncertainty. 
• Secondly, as incidence decreases, R will tend towards 1, and has to be evaluated in conjunction with incidence. The policy implications of R = 1 when there are 1,000 new infections per day are very different to when there are 100,000 per day.
• Finally, R is an average measure. When incidence is low, an outbreak in one place could result in estimates of R for the entire region to become higher than 1. Conversely, small, local outbreaks will not be detected. Estimates of R based on small numbers may also not capture change in the area fast enough to inform policy in a useful way.