Coronavirus: The disease is clearly spreading - but can we trust the figures?

Before we get to the presentation by Sir Patrick Vallance and Professor Chris Whitty this morning, do you mind if I begin with a couple of recollections?

The first of them dates back to the summer of 2016 and the midst of the EU Referendum campaign.

I vividly recall being ushered into the Treasury to be told about the severe consequences of Britain voting to leave the European Union.

Please use Chrome browser for a more accessible video player

0:55

Now, I am usually quite receptive to pointy-headed reports from economists, but even so there was something a little off about the whole thing.

Later on that day I reported on Sky News that it all smelt a little fishy.

You see, the main gist of their analysis was that there would be a recession if Britain left the EU, but when you read the small print it was pretty clear that they had wrangled their models to generate the scary R-word headline that George Osborne had clearly demanded.

What was strange about the moment wasn't so much the conclusion, which was pretty uncontroversial. Most economists assumed that leaving would be bad for Britain's gross domestic product.

It was the extent to which they were willing to use their PR machine to create scary headlines. It was, I felt, as if they were protesting too much.

The second recollection is from March of this year.

I recall in those very early stages of the coronavirus pandemic that many within government were quite sceptical that Britain would end up following in the footsteps of countries like Italy, Spain and France who had seen their caseload of COVID-19 rise exponentially. We were, they said, just different.

I found this a little odd. I am no epidemiologist but I can read a spreadsheet and even though in March few people had died from the disease in Britain, it was clear from the data that if we carried on following those trajectories, within a few weeks the toll would be well into five figures. And so it proved.

The reason I mention those two vignettes is that I felt a strange sense of double deja vu as I listened to the chief scientific adviser and chief medical officer laying out the state of the virus this morning.

One chart in particular leapt out - a graph showing the number of positive COVID-19 tests in recent weeks and an illustrative extrapolation if those cases doubled every seven days. Within a few weeks, the cases multiply up to 49,000 a day.

Now, Sir Patrick was at pains to assure us that this was merely an illustration of how the disease could spread, but when you consider the chart for a moment, a couple of questions arise.

First: why "doubling every seven days"?

Only three days earlier the government had published its latest estimate for how fast the disease is spreading and it was between 2% and 7%. Doubling every seven days implies growth of over 10%.

Has the disease really worsened so quickly? If so, why are we only hearing about that tangentially in an illustrative scenario rather than as part of a formal announcement? It is, to put it lightly, odd.

Anyway, it might sound like a small difference but if the disease is spreading at the rate they said it was on Friday, then that would put us on track for around 20,000 cases by mid October, less than half the figure Sir Patrick told us about.

That, after all, is the power of compound growth: only a small percentage divergence in a disease's growth rate can make an enormous difference in a small space of time.

Anyway, the second question mark over the data is why Sir Patrick's illustration assumed British COVID-19 cases would diverge so dramatically from the rest of Europe. It's a bit like March all over again, except in reverse.

For once again, UK positive COVID-19 cases are following almost lockstep behind France and Spain's.

But while in March many in government refused to countenance that our eventual virus toll could be as high as our European neighbours, this time around they seem unwilling to countenance that it could be as low as our European neighbours.

For the reality is after a short period when their COVID cases did indeed seem to be doubling every week or so, the growth in the number of cases in France and Spain has slowed quite dramatically.

They are still high, and this should give no one comfort. But their trajectory is nothing like the one laid out by Sir Patrick.

In short, if Britain carried on following in line with France and Spain, as it has done so far, by mid-October the number of cases will be around 10,000, rather than 49,000.

Now, on the one hand, this is simply pedantry. The disease is clearly spreading - the only question is how fast it's spreading.

On the other hand, these assumptions about growth rates can make some enormous differences: the difference between a gradual spread, such as we saw in Germany in March and April, and a rapid spread, such as we saw in this country in March and April.

So far at least, the data suggests the UK is more likely to experience the former than the latter - though of course these things can change rapidly.

But if, as happened last time around, the UK follows in the footsteps of France and Spain, the numbers produced by the government's scientific brains will rapidly look very wrong indeed.

That worries me, because government credibility is one of the foundations in the effort for fighting this disease.

Sir Patrick was, as I say, emphatic that these were simply illustrative numbers - not a forecast.

On the other hand, I can't help but recall that Brexit briefing in 2016, where the government protested rather too much in order to generate scary headlines. I'm sure this is different. But it's hard to escape the echoes all the same.