Covid: the modellers never sleep

By Richard North - December 26, 2021

Most normal humans – those who have had the opportunity – will have been taking it easy for the last 24 hours. While struggling to qualify as “normal”, I even took the chance to complete an actual Airfix model (pictured). Originally produced in 1973, it represents a SAM-2 “Guideline” missile and its Zil-157 transporter, and is now so rare that individual kits trade for three-figure sums.

Meanwhile, an altogether more sinister group of modellers have been plying their own trade, juggling figures and data rather than finely-crafted plastic parts, to an altogether different purpose – their intent, come what may, to see us locked into further Covid restrictions before the year end.

The name of the game is to convince Johnson that the threat posed by the omicron variant is far more serious than recent studies would suggest. While conceding that the variant is less virulent than its predecessors, in particular the delta variant.

The trick is to show that we are about to be hit by a huge wave of cases so that, even though the proportion of sufferers needing hospital treatment will be reduced, absolute numbers will be far higher, thus risking swamping the NHS, with the spectre of patients going untreated (more so than at present).

Thus, we find the Mail (and others) doing the “shroud waving” on behalf of SAGE and the rest of the Covid Mafia, calling for a decision as early as Monday on whether we are to be dragged into another round of curbs.

Already, though, the Telegraph is recording continued push-back, with UK Health Security Agency chief executive, Jenny Harries, being accused of disseminating “dodgy data” which prolongs the uncertainty about hospitalisation rates.

Harries is said to have been behind a claim by health secretary Sajid Javid that there is typically a 17-day lag between patients becoming infected and requiring hospitalisation, thereby keeping over the threat of a spike in hospitalisation over the festivities.

Data from the Office for National Statistics (ONS), however, suggest that the average lag is nine to 10 days, indicating that we should already be seeing a substantial uptick in admissions, which is so far refusing to materialise except in limited area of London.

Harries, herself, has been reluctant to tone down the scare rhetoric, having described the emergence of omicron as “probably the most significant threat we’ve had since the start of the pandemic”, describing optimistic findings on reduced virulence as merely a “glimmer of hope“.

Now, prof “lockdown” Ferguson has re-joined the fray, with another of his reports, this one published late on Christmas Eve, thereby avoiding coverage in the media.

Entitled, “Hospitalisation risk for Omicron cases in England”, it does its best to play down the significance of the South African analyses, and stuffs the pages with impenetrable jargon. But, where it lacks sufficient data, notes that there were “a very low number of recorded hospitalisations for later cases at the time of analysis”, it relies on the Harries ploy, stating that this is “a result of lags in routine hospital data flows”.

One can imagine Johnson’s eyes glazing over at the sight of some of the formulae on offer (see page 4), unaware that these are being used to produce estimates of hospitalisation rates based on a number of assumptions for which the evidential bases are extremely slender, and which must – perforce – rely on extremely small numbers.

Only reading the small print does one see the caveat that the analysis is only an “initial step” towards assessing the clinical severity and risk posed by the Omicron variant. Given the limited data, Ferguson and his team have had to rely on hospital attendance figures only, “the lowest level of severity associated with tertiary care”, with no information on resort to critical care or dwell times. And then we come to this gem:

There are several limitations of our analysis. The conditional Poisson regression models we have used implicitly assume that the timescale of progression to more severe infection outcomes does not vary by variant. We are analysing data in real-time, with shorter follow-up periods than is usual for studies of hospitalisation outcomes of disease; while our analysis controls for right-censoring, our estimates may be biased by any differences in rapidity of reporting or in admission practices between the largely inner-city hospitals responding to Omicron cases in the study period and the more varied range of hospitals responding to Delta cases. We did not have access to or make use of data on reported reasons for hospital attendance, so some proportion of the hospital events recorded may be unrelated to SARS-CoV-2 infection.

The report then goes on to say:

Over the study period, the demographic distribution of Omicron and Delta cases still differed substantially in England, with Omicron more commonly affecting younger adults of non-White ethnicities living in major cities than Delta. While the stratified regression (akin to a case control study) we adopted aims to control for such differences, residual bias may be inevitable. Last, there was limited power to examine more severe outcomes than hospital attendance; this analysis will be updated to include a wider range of outcomes as data accumulate.

In short, given the limitations of the data, the shortness of the study period and the restricted geographical scope – where other areas of England are likely to be different from the study area – there are no reliable conclusions which can be drawn from the work.

However, if Johnson does attempt to buy into this “initial step”, he may be in trouble. The Observer noting that Cabinet ministers vying to succeed Boris Johnson have been warned by Tory colleagues that they will damage their chances unless they stridently oppose further Covid measures, as MPs called for New Year’s Eve restrictions to be ruled out.

Tory MPs, we are told, are hardening their attitudes against further restrictions and now want cabinet ministers to be vocal in their opposition. The paper cites one member of the Covid Recovery Group of Tory MPs, saying: “If No 10 proposes tighter restrictions straight after Christmas, those cabinet ministers with freedom-loving instincts – who gave us all so much hope last week – must speak out”.

This member reinforces his view with that warning that, “In any future leadership contest, we will all remember how they acted this week. We need real, gutsy, freedom-loving Conservatives to rescue us from this madness”.

Needless to say, the Observer gives the last word to the Covid Mafia, aka “scientists”, having professor Adam Finn of Bristol University saying: “The wiser course is clearly to do everything we can to minimise transmission while the size and nature of the impact of Omicron clarifies itself and while everyone who is willing gets a booster and gets time to mount an immune response to it”.

Then, in a classic statement which typifies the risk averse stance which is dominating much of the discourse, he says: “Better to take the flak for overreacting than for failing to act when there was still time”.

Of course, this neglects the very considerable downside of imposing controls, not forgetting that the backlog of non-Covid treatment has its own penalty. It may well turn out that more people die through lack of such treatment than were killed by Covid.

Whether the Tory MPs opposing further restrictions are entirely aware of this remains to be seen, but – as it stands – the case has not been made for taking any more action. Johnson will now find himself severely tested with the potential that, if he does succumb to the scare dynamic, he will have to face another mauling in parliament.

He really should insist that the modelling is left to Airfix.