Covid and the difficulties of following the science

By Ross Clark Spectator June 15, 2021

Did anyone fancy being in Boris Johnson’s shoes before he made the decision to delay the full lifting of Covid restrictions? Keir Starmer, perhaps. But even Starmer might have preferred opposition if he had read the latest paper by the Scientific Pandemic Influenza Group on Modelling (SPI-M) committee, which would have informed the Prime Minister’s decision. It reinforces just how difficult it is for any government to ‘follow the science’. If you can sum the paper up in one sentence it would be ‘sorry, but we really don’t have much of a clue as to what will happen’.

Here are just a couple of highlights: ‘The scale of this resurgence is highly uncertain, and it could be either considerably smaller or larger than previous waves’ and ‘R is estimated to be 40-80 per cent higher for Delta [Indian] than for Alpha [Kent], although a figure higher or lower than this cannot be ruled out’.

When there is such a wide range of uncertainties in the input data, there is inevitably going to be an even greater range of possibilities spewed out by the models. Look through the accompanying graphs and they can be used to support any assertion, with hospitalisation estimates ranging from 10,000 a day by early August to 100 a day by the same date (187 people with Covid were admitted to hospital across the UK on 8 June, the latest data published, and this reached a peak of 4,160 on 19 January).

The government has said that its change of heart is down to the emergence of the Indian variant, and in particular the claim that it is more likely to land you in hospital than previous variants. But is this true? As you can see in Figure 1 in the SPI-M report, the number of hospitalisations per 1,000 confirmed cases of Covid has, in fact, been on a downward trend all year – and has continued a little downwards even as the Indian variant has come to dominate. The claim that the Indian variant may be more likely to land you in hospital stems from a much smaller analysis involving emergency admissions only. Trouble is, we don’t have the full data on this because not all samples from infected patients are analysed to see which variant they are.

We have been here before: in January the government called a press conference to deliver what it thought was the grim news that the Kent variant, as it was then called, was around 30 per cent more likely to kill you than the previous Wuhan variant. That claim, too, was based on incomplete data. Two studies published in the Lancet in April contradicted the claim – suggesting that the Kent variant did not, after all, have a higher mortality rate.

There are plenty of frightening graphs in the latest SPI-M report to support the case for delaying the full reopening – either to 19 July or indefinitely. But if you react to worst-case scenarios based on incomplete data and huge uncertainties in the models, then you will never reopen society. All we can say for certain is that ‘the science’ – in the shape of SPI-M modelling – is not providing the government with very useful guidance.