Many of us stuck at home have some time on our hands. We spectate history unfolding.
For someone whose only achievement in science is a C at Biology O level it’s hard unpicking the opinion from the fact, what’s proven from what’s not, what’s click bait from what’s sharp analysis. Everyone seems to have an opinion.
May be legal skills may help?
Lawyers test the evidence to see whether the evidence comes up to proof. One skill I do have which may be relevant is sifting the evidence, seeing what’s reliable, discarding what’s unreliable, and drawing some conclusions from reliable .
I posed myself a few questions. I looked at the evidence to see whether it would yield an answer.
The first question on most people’s lips is:
How deadly is Covid 19?
To answer that question you need to know the Infection Fatality Rate or IFR. The IFR is the numbers with the virus divided by the numbers who die with the virus.
The IFR is different from the Case Fatality Rate [CFR]. The CFR counts those who test positive for Covid 19 who subsequently die.
The CFR is a higher percentage than the IFR. Some people infected with Covid 19 don’t suffer any symptoms and don’t become a medical case.
We have two problems with the evidence. The evidence for cause of death is on a death certificate which itself may be inaccurate:
1. Tests are in short supply. People dying from Covid 19 may not have their death recorded as such on their death certificate, particularly if they die at home or at a care home and their corpses are not tested for the virus.
2. Covid 19 is a notifiable disease. That means that anyone who dies with the virus has to have the virus on the death certificate and is treated as a Covid 19 death even though they may not die from the virus. In Italy one estimate is that only 12% of deaths recorded as Coronavirus deaths were deaths from the virus.
The evidential data is not 100% accurate as it is reliant on a death certificate which may or may not accurately record the cause of death.
One crude way of measuring fatality is to look at excess mortality. UK has an average mortality of 600,000 deaths a year. At year end we will be able to see how many people have died in 2020. If it’s 630,000 then the excess mortality is 30,000.
There is a European Website which tracks excess mortality. Most countries in the EU are currently showing excess mortality.
There is not enough reliable data around to be able to know with any accuracy what the IFR for Covid 19 is.
Therefore let’s call on the experts to see whether they can extrapolate from the unreliable data an IFR for Covid 19.
My expert of choice is Carl Heneghan at Oxford University’s Centre for Evidence Based Medicine who puts the likely IFR at between 0.1% and 0.36% in this piece. The relevant extract is as follows:
Taking account of historical experience, trends in the data, increased number of infections in the population at largest, and potential impact of misclassification of deaths gives a presumed estimate for the COVID-19 IFR somewhere between 0.1% and 0.36%.*
So an infected person has between 1 in a thousand and 3.6 in a thousand risk of dying. However that risk is not distributed equally, the risk overwhelmingly falls on the older part of the population as well as those with underlying medical conditions or co-morbidities.
What are the chances of catching the virus and not having any symptoms of illness?
How long between infection and showing any symptoms?
How long between symptoms showing and death?
The vast majority of those infected recover or do not show any symptoms at all. For those who die of the disease the average number of days between showing symptoms and death is 17.8 days.
Are we over the worst in the UK?
According to CEBM April 8 saw the most number of covid 19 related deaths.
Whether we are over the worst or we will have other phases of infection is the big question.
In terms of phase one of covid 19 in the UK — 8 April 2020 was the day of peak death.
What does the replication rate mean and how significant is it?
The replication rate or RO is the number of new infections spread by an infected person. If we get RO below one the virus should eventually die as there will be insufficient hosts on which it can replicate. Getting the RO below 1 and flattening the curve is a key policy response of all governments.
Some governments have enforced a lockdown, others a soft lockdown, while one government, Sweden, has not enforced any lockdown merely trusted its citizens to observe public health messages such as social distancing, hand washing, protecting the vulnerable and no gatherings above 50.
What is herd immunity?
Other than a very emotive term, it’s the only way the virus will die out when the herd, the virus’ potential hosts, are either immune via vaccine or naturally via previous infection. The R0 will fall below one with herd immunity as there are not enough non-immune hosts to allow the virus to spread.
I found this a useful explanation of herd immunity.
Did we lockdown at the optimal time?
To answer that we need to know at what day was the replication rate, or R0, or the infection rate, at its highest and when did R0 start to decline.
To arrive at R0’s highest dateor the day of peak infection we track back 23 days from the peak death day.
The peak death day was 8 April 2020. 23 days is arrived at by allowing an average of 5 days average incubation before showing symptoms and 17.8 days from showing symptoms to death.
Tracking back 23 days from 8 April 2020 we arrive at 16 March 2020. Based on the evidence that looks like the day of peak infection in the UK and after that day the numbers infected reduced as people started to take social distancing more seriously, the pubs and cinemas began to empty, hands were being washed more frequently, coughs were being caught etc.
From the evidence it looks like the public health messages of social distancing, self-isolating, hand washing were pulling down the infection rate on or before 16 March 2020, a full week before the lockdown started.
That tallies with my own experience where people I know began taking social distancing, hand washing, self-isolation a lot more seriously in mid-March onwards, a full week before lockdown.
So it looks like from the data we locked down a week after peak infection. Locking down two or three weeks earlier in teh UK would have prevented more infections and would have prevented more deaths.
Locking down when we did on 23 March 2020 does not appear to have moved the R0 needle as that infection rate was declining downwards from 16 March 2020 onwards as evidenced by the death rate peaking on 8 April 2020.
Was locking down too late on 23 March 2020 worse than not deciding not to lock down on 23 March 2020?
Arguably so. As we have seen the numbers of deaths and hence infections were starting to fall a full week before lockdown. Lockdown has not resulted in the numbers of infections falling, they were falling anyway from 16 March 2020.
Sweden who did not have a lockdown, despite the country’s different demographics, size and shape, may be a useful counter-factual to the UK.
It looks like Sweden’s death rate peaked on 8 April 2020 as well.
What does the future hold?
Who knows? But the tentative evidence shows that:
1. The IFR may be lower than thought at between 0.1 and 0.37% of those who are infected die.
2. More of us may be asymptomatic than previously thought at around 50%.
3. Social distancing, hand washing, catching cough droplets and fewer crowded gatherings appears to have worked in Sweden and was eventually starting to work in the UK if peak infection date in the UK was 16 March 2020.
Feel free to bash holes in my evidence base or in my reasoning. I come to the issue without any specialist knowledge or skills.
My only relevant skill is poking the evidence to see whether it stands up or not.