This week, I will be posting a series of threads on the risks of a second wave - and of our response to it. I hope this will help us overcome the twin dangers of fear and complacency, prevent a second wave - and a second lockdown. This is no. 3: How deadly/dangerous is COVID-19?
There has been a lot of disagreement about how deadly Covid-19 is – with many people hugely overestimating the risk and others insisting it is no worse than flu. We now have much better evidence with actual data (as opposed to modeling) to answer this question.
Firstly, you need to look at the infection fatality rate (IFR) which is the percentage of those INFECTED who die (not just those who are diagnosed as positive cases - which is a huge underestimate of the actual number infected -eg in the UK, we have recorded roughly 400,000 cases
and 40,000 deaths but the true number of infections is at least 4 million based on antibody testing) - which gives an IFR of about 1% - but it is reducing due to better treatment - and not all infections are picked up by antibody testing, and so it could be as low as 0.5% now.
Most estimates internationally are also between 0.5 to 1.0% – whereas flu is less than 0.1%. Further evidence for this comes from New York, where about 20,000 died and so an IFR of 0.1% would mean that 20 million had been infected but its population is only 8 million.
The IFR also varies hugely by age - in children and adults under 45 it is almost zero (less than flu) whereas in over 65s, it is above 5% (higher than flu) - and by comorbidities (other diseases) especially severe obesity and diabetes.
Also, we now have good evidence from death certificates that Covid-19 was the underlying cause of death in about 50,000 people – it’s not that they died just ‘with Covid-19’ – which tallies with my own experience of the patients I saw in the first wave.
We also need to take into account the illness/disability that COVID causes - not just death. We have a lot more evidence now about the longer-term adverse health effects that occur in 10-15 % of symptomatic cases - including in the young and those with mild symptoms.
So we should neither overestimate the risks of COVID- which leads to unnecessary fear- nor underestimate it, which leads to unjustified complacency. Our response should be proportionate to the actual harms- but equally importantly, also balanced against the harms of the response.
With restrictions having ended across all home nations, it’s time to have a look at the results of the 'natural experiment' that we've had in the UK since July 19 when policies diverged with England having fewer restrictions than Wales, Scotland & Northern Ireland since then.
1/6
While some described England’s approach as a "dangerous & unethical experiment", @CMO_England, I & others thought that ongoing restrictions in a highly immune population would simply postpone infections - not prevent them.
So what has happened with infection & death rates?
2/6
The graphs below show cumulative infections & deaths across the home nations from July 19 to April 9 (last date data available) & also from December 17 to April 9 when Omicron became dominant and show no significant differences in infection or death rates over that time. 3/6
Studies like these (based on models) ignore what is now well-established i.e. key role of voluntary behaviour change in controlling Covid - with behaviour changing more in response to the level/fear of Covid ('voluntary lockdown') than govt. mandates 1/7 telegraph.co.uk/news/2022/04/1…
e.g. household visiting & mobility fell more in the week before lockdown than after lockdown came in. There is limited evidence that the govt. mandate produced a bigger fall than that which occurred anyway & we know that compliance with mandates is associated with level of risk
And we also saw what happened with the Omicron wave in December when, contrary to what was widely predicted based on models, voluntary behaviour changes led to an outcome similar to what was projected to happen with lockdown equivalent to step 2 of roadmap (only schools open.)
With hospital admissions in England approaching the 1st Omicron peak (& hospitals being busier overall due to more non-Covid patients) & ONS likely to show a record no. infections this week, some are again calling for a return of mandatory measures assuming this will help but
1/6
evidence for this is lacking.
In the absence of better evidence, we look at the real world data from the natural experiment we've had since July 19 when the home nations diverged & we don't see any indications from the case, death or ONS survey data that England has fared worse
from have less mandates & relying more on public health guidance & people taking personal responsibility.
This is also true for the most recent Omicron wave where Scotland has had the highest infection rates & higher hospitalisations than first peak despite having more mandates.
V. important paper published in the Lancet - the first peer-reviewed global estimates of excess deaths (the most reliable way to compare Covid deaths) over first 2 years of pandemic with findings that will surprise many & correct five widespread misconceptions / assumptions:
1/9
1. Far from the UK having the worst death rate in Europe (or even Western Europe) as many still think, it is actually 29th in Europe & 9th in Western Europe - below the Western European average & at the same level as France & Germany (no statistically significant difference)
A note on excess deaths last wk: 1. They are 8.6% lower when comparing to the 5 yr avg which includes 2021 - unsurprising as it was peak of 2nd wave & no-one fully vaccinated - but 2. More significantly, they are also 3.3% lower than the 5 yr avg. for 2015-19 (pre-pandemic) 1/4
This is a huge contrast to the same week last year when there were 18,676 deaths (about 1.5 times higher than now) and deaths were 41.3% above the five-year average again illustrating the massive impact of vaccines and Omicron's decreased severity (& very low deaths from flu)
2/4
It is also consistent with point I raised in December re. '28 day positive test' measure over-estimating deaths due to Covid with Omicron (as shown below from @AlistairHaimes).
The proportion of 'due to' (72.9%) vs. 'involving' Covid is also lower than previous week (77.4%). 3/4
There is understandable concern about increasing numbers of children being admitted into hospital with a positive Covid test recently. This needs urgent investigation but from what we know so far the data suggests the headline figures may be hiding a more complicated reality:
Firstly, the situation with children is in many ways similar to that with adults in that we need to distinguish between those who are admitted for a Covid-like syndrome versus those who are admitted with another condition (eg. abdominal pain) but have an incidental positive test
(There are also those who are admitted with a condition which Covid may have contributed to - and for the in-hospital figures those who were admitted before and now have a hospital-acquired infection.)