WintonCentre
Published in

WintonCentre

What are the risks of COVID? And what is meant by ‘the risks of COVID’?

As COVID-19 changes from being seen as a societal threat to a problem in risk management, it is essential that we get a handle on the magnitudes of the risk we face, and try to work out ways to communicate these appropriately. Note that I am only covering the lethal risks, not the potentially important consequences of illness or treatment.

When discussing the risks surrounding COVID, it is very important to carefully distinguish —

  • The risks of dying from COVID-19, among people who get it (known as the Infection Fatality Rate IFR ). This will depend on your health, medical care and so on.

Fortunately new data and analyses permit more insight into these quantities, and point the way to communicate these risks. I’m going to cover the Population Fatality Rate now, finish off with a rant, and return to the IFR in a future blog.

So everything from now on refers to the risks faced by people who were not currently infected.

Population Fatality Rates

The latest data from the Office for National Statistics covers deaths registered in England and Wales up to May 1st, and the table below presents COVID population fatality rates over the peak 5 weeks of the epidemic.

For example, among nearly 6 million people aged 65–74 in England and Wales, over 5000 were registered as dying from or with COVID in five weeks, a rate of around 87 in 100,000, or 1 in 1,143. We would normally expect around 8,300 deaths in this age-group in this period, and so COVID represented a 62% increased risk. This is equivalent to around ‘22 days extra risk’ during this 5 weeks of the epidemic. A COVID death is one in which COVID-19 is mentioned on the death certificate.

It’s a complex table but worth studying.

  • Starting at the first row, 2 deaths have been recorded among over 10 million children under 15 (around 1 in 5 million), an extremely low risk that that represents only 1% of the normal risk faced by this group over this period (and it is worth noting that 95% of COVID deaths in hospital have underlying health conditions).

The risks differ for men and women, and a full table is provided at the bottom of the blog, and shown in the graphs below.

COVID population death rates for females and males on a logarithmic scale — data from ONS.

The amazing linearity of the data on the logarithmic scale shows that COVID rates have a fairly precise exponential increase with age, increasing at around 11–12% each year, corresponding to a doubling every 6–7 years. This means that a 20-year age-gap increased the risk by around 8-fold. So, compared to a 20-year-old, an 80-year-old had 8 * 8 * 8 ~ 500 times the risk of dying.

Men had roughly double the risk compared with women of the same age.

The extra COVID population death rates are roughly proportional (ie parallel on a logarithmic scale) to ‘normal’ death rates for over 45s, but well below normal rates for younger ages. Note these are IN ADDITION to the normal rates.

The same data can be shown on a linear scale, which better displays the huge variation of population risk with age.

COVID population death rates for females and males on a linear scale — data from ONS.

A simple interpretation of the population fatality rates.

Over this 5-week period covering the peak of the epidemic-

  • Men had roughly double the risk of getting the virus and dying, compared with women of the same age.

The lesser relative effect of COVID on younger groups could be partly because risks their ‘normal’ risk will be more strongly influenced by accidents and non-natural causes, whereas COVID seems to multiply your risk of ‘natural causes’ — it just seems to take any frailty and multiply it.

These are observed historical rates in the population, and cannot be quoted as the future risks of getting COVID and dying. In particular the risks of infection will be altered by factors that limit your exposure, and will be expected to drop massively as the epidemic is brought under control. In contrast we might expect the fatality rate if you become infected to remain fairly stable over time.

What do we mean by ‘the risks of COVID’?

Please permit me a rant. It is vital for journalists and everyone else (including me) to try and avoid phrases like ‘the risks of dying from COVID-19’, as this is deeply ambiguous. As I said at the start, it crucially depends on the group it refers to, as it could mean-

  • The risks of dying from COVID-19, among people who get it (the Infection Fatality Rate IFR ).

These are so easily confused. An analysis last week by the Office for National Statistics reported that Black, Asian and minority ethnic (BAME) groups were about twice as likely, after adjusting for some contextual factors, of dying from COVID. But this clearly referred to the population fatality rate — in other words BAME groups had a higher risk of both getting the disease and then dying from it, and an unknown part of this excess risk could come from an increased risk of catching the virus, perhaps through coming in contact with more people in their daily lives. But in the BBC 10pm News on May 7th, this was reported as BAME individuals being “90% more likely to die, if they became seriously ill with COVID-19”, which is not at all what was being claimed and could be very misleading: ethnicity was not an important risk factor for COVID patients who were hospitalised.

Data sources:

Additional tables

Observed population fatality rates for deaths registered in England and Wales over the 5-week period 28th March to 1st May.
Observed population fatality rates for deaths registered in England and Wales over the 5-week period 28th March to 1st May.

--

--

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
David Spiegelhalter

David Spiegelhalter

4.2K Followers

Statistician, communicator about evidence, risk, probability, chance, uncertainty, etc. Chair, Winton Centre for Risk and Evidence Communication, Cambridge.