December 29, 2021 - 8:31am
Two weeks ago, when panic about the Omicron variant of Covid was at its peak and it felt imminent that the UK would be going back into lockdown, I gave an interview to UnHerd’s Freddie Sayers. I am a PhD in engineering and a research associate at the IFK University of Johannesburg — I am not an epidemiologist — but I had been closely following the data of the Omicron wave that started in Gauteng province in my home country of South Africa.
On 4th of December I put together some basic projections for all the main Omicron variables — cases, hospital beds, ICU beds, ventilated beds, deaths — using simple growth rate extrapolations. Each variable goes through an exponential phase during which daily growth peaks, followed by a period with a consistent decline in growth rate. This consistent decline makes it possible to predict the date of the peak with reasonable accuracy. Once a likely date range for a variable has been determined, it becomes easier to gauge the likely peak levels of all variables. It was clear already by then that the Omicron wave was going to be extremely mild compared to the Delta wave, which my forecasts confirmed.
What has been remarkable in the intervening weeks — despite the very clear data trends and statements of practitioners like Dr Angelique Coetzee, head of the South African Medical Association, that the variant was mild in all groups — is how slow other countries have been to accept this as fact. Sir Patrick Vallance and Chris Whitty, the UK’s chief scientific and medical officers, seem to be the last to admit what has been evident to most observers for almost a month: namely, that Omicron is much milder than Delta. With numbers in UK hospitals rising, a degree of caution is of course understandable, but the refusal to acknowledge the evidence from South Africa has been odd.
It is true that South Africa is a very different country to the UK, with a younger population and a different profile of prior immunity. But the population also has a major obesity problem, poor metabolic health, widespread HIV and had been hit relatively badly by previous waves of the virus. So the idea that ‘just because Omicron has not done much damage in South Africa it could still be devastating in the UK’ was always hard to believe.
In fact, the Omicron wave in Gauteng turned out to be even milder than my forecasts suggested; since the interview, every single one of the variables came in below projected levels. My estimate that deaths would be 25 times lower than the Delta wave still holds, and I have revised my estimate for the infection fatality rate down from 0.053% (11th December) to 0.036%.
As all variables seem to have peaked, we can also make estimates of the peak numbers per 100,000 — which should be useful for planning purposes in other regions and nations.
Join the discussion
Join like minded readers that support our journalism by becoming a paid subscriber
To join the discussion in the comments, become a paid subscriber.
Join like minded readers that support our journalism, read unlimited articles and enjoy other subscriber-only benefits.
SubscribeNot bad, but the projection for cases was still pessimistic by nearly 2x. From this two of my pre-existing beliefs are reinforced:
This projection didn’t come alone at the time – it came with the information that the symptoms were mild and hospitalisations and deaths were down. Real information, as opposed to just a projection.
Because this variant is so mild in comparison, even more cases are a good thing.
Everything I’ve seen so far with regards to epidemiological modelling suggests it is snake oil.
The one model we’ve seen the source code for is the IC model and I have a couple of major issues with it.
Firstly, the IC model clearly demonstrates that the developers behind it are complete devoid of the requisite software engineering skills. It is littered with errors that should not be present (at least in that magnitude) in any professional piece of software. As John Carmack pointed out, it was likely ported from Fortran to C++ but the IC team do not understand C++, and that’s an understatement. A useful analogy here would be if having work on the electrics or gas in your house the tradesmen requires some actual certifications, apparenlty modellers are an unregulated industry with the power to destroy not just a house but a whole economy.
Secondly, I’m unconvinced by the logic employed. The model simply takes the abstract IFRs and breaks them down into concrete interactions. I was a little surprised at how simplistic it was because at first I thought it might start from modelling how virons interact with the immune system, attack cells and propagate but extrapolated to a whole population. It’s the exact opposite, which I suspect will always lead it to a predictable result – honestly you can do a back of a beer mat calculation and get similar numbers as the IC model – because you’re using IFR/CFR that is already calculated from the existing data.
A final point, one that I’ve made a few times before, is that I think models are engineering and not science. Engineering is the process of utilising science to produce some form of artifact. Calling it science is misleading because it creates the impression we’re observing the natural world, we’re not, we’re observing a human artifact we built from a rudimentary understanding of epidemiology and a new virus.
I saw the source code of the model used by Fergusson and, as you said, utter junk!
Good assessment, software engineer for over 40 years 100% agrees. But there is another vital issue.
Consider fluid dynamics. Computer modelling is very good, to the point where, for example, aircraft aerodynamic design can be done almost entirely in a computer. This is because it was possible to experimentally test models against the real world; to perform experiments.
With epidemiologic modelling, this is not possible. So, there is no way to be able to assign confidence limits to the models, rendering them essentially useless.
Agreed.
It’s not my speciality, but the company I work for develop material and stress analysis software and it is highly useful in a range of engineering applications. As you say, the ability to feed results back into the system from testing is how they become more accurate.
The IC model was apparently designed from the Spanish flu (which is questionable in itself due to the lack of genuinely good data from that period) and just given a few different variables for the viruses it’s been used to model in recent years. My suspicion is that viruses have a lot more variability between strains than say different types of metals.
What I can’t understand is that what Neil Ferguson says STILL appears regularly in the news. He started with a bad track record, and so it has continued (while in between, he ‘resigned’ for an indiscretion). Is it because he’s predicting what some people wish to hear?
You mean like Fauci on this side of the pond?
Thank you, very informative analysis from someone qualified to give it. I’ve become increasingly sceptical over the extent to which computer models – whether for covid or climate change, for example – are being used to make confident-sounding predictions, steer policy and in some cases make decisions which have a deep impact on many aspects of our lives. It’s about time the limitations of such models were more widely discussed and appreciated. No doubt they have their proper role, but they should not be taken as gospel truth on their own.
Thank you for the article . Is there any way of calculating what the equivalent death rate in the UK would be ? I believe that covid deaths refer to deaths in all settings in this country .
It is nice to have some comparatively good news
“….but the refusal to acknowledge the evidence from South Africa has been odd.”
The takeaway from this piece is that the sky is not falling, the virus is mutating–that’s what viruses do– but the variations are becoming less virulent, and humans are learning to live with it. That there is a “refusal to acknowledge the evidence” is not odd, but de rigeur. Whatever happened to follow the science?
Because of the measured tone backed up by (seemingly) reliable data, this piece should be banned as the wacko propaganda of the extreme right. It might people less afraid, and the boffins cannot have that. FaceBook–please take this down!
Taken to its logical conclusion, it would seem that somewhat normal life can be resumed, which is not what the boffins–Neal Ferguson, Jacinda Ardern, Fauci, Sleepy Joe Biden and their ilk–want to hear. If there is no emergency, then it follows that there can be no emergency measures. Freedom, not an inherent right but something granted by the leave of our betters, might be returned to the people. This is something that the nanny state will fight to their last breathe….
This piece shows what an absolute clown show the “scientific” advisors have run, and the BBC is one of the worst offenders. Their “scientific” experts have touted “vaccine equity” at every turn, virtually demanding that we in the West not have second shots, not have booster shots until the corrupt and feckless Third World attains “equity.” These Churlish Unhinged Negative Tw@#s still push this argument to an extent, but since it has failed to gain much traction on the ground, speak wistfully that this has not happened. Simply disgusting!
The “boffins” are morons of course. But dangerous ones! From humble beginnings covid has escalated into the mother of all dramas. And the boffins have become intoxicated on the power afforded them in this, their holy mission to protect us all. In 2 years they have destroyed much and accomplished little. And now they are trapped. Going back is not an option: admitting that the danger covid presented was hopelessly overestimated would mean not only huge loss of face but explanations & repercussions which they dare not contemplate. So they blunder on, inflicting ever more nonsensical and absurd restrictions upon us all. Liberty is curtailed. Be aware, the next “wave” can be switched on at any time. New fetters appear every month or two. Where will it end? How did we, the public, elect such idiots in the first place? And why do we continue to put up with it all? It is time, surely, for elections, before they, too, are put on hold indefinitely.
Cheers, Art! I just re-read this, and it is an excellent comment. You’ve hit the nail on the head! The boffins are intoxicated with power and nothing will deter them from their “holy mission!”
Well said!
We have been hearing good news since the identification of the ‘super mutant variant’ Omicron in November in South Africa. Unfortunately governments, organizations, certain big businesses and many citizens are either unable or unwilling to welcome good news and the possible end to the pandemic.
Some citizens have become addicted to fear and panic (by far, mostly from the middle classes who have money or an income) – that much is evident. They have shown an inability to think for themselves throughout and seem unable to start embracing good news and adjusting behaviour.
More nefarious motives also exist – many governments are locked into worst case scenarios and some are in bed with the organizations and large businesses who have accrued staggering amounts of money that is being made from the pandemic. Some of this sector have also become addicted to the huge amount of power they have amassed.
Enough.
I can’t comment on what has been going on on corporate media, as I very seldom watch it any longer, but I am betting they are not dialing right down.
Hear hear Lesley! I always appreciate your well informed comments.
Corporate media is not dialing it down… I am staying with a friend for a few days and she watches the news a lot. What I hear is that “Omicron cases are surging” and “people are concerned about Omicron.”
This appears to be the primary messaging. Now, if you look at what I wrote it’s quite interesting. Notice how the news is not saying people are dying, hospitals are filling up, the death rate is rising, there are new horrible symptoms, or anything of that sort. No. They are making vague reports that are in one sense true, but from a broader picture very inaccurate. They are like the horror movie which presents the scary music and points to the shaded corner in allusion to the boogieman being there, when he’s not.
That’s the state of affairs in the mass media.
It’s interesting. I always took a dim view of epidemiologists previously due to their dire history of pronouncements in the field of nutritional epidemiology- otherwise known as data dredging food frequency questionnaires for dubious correlations between food substance X and cancer. But I’d given their infectious disease counterparts a pass because of John Snow etc., but no more, they’re as useless as each other.
I wasn’t looking forward to getting Covid. I’m 71 and have high blood pressure (though well controlled) but I knew it was inevitable. Well here I am, Day 4 and I’ve only had a blocked nose for 36 hours. I feel back to normal now; Omicron is like a very mild, short cold for the majority of people.
Pleased to hear that Joy.
“….but the refusal to acknowledge the evidence from South Africa has been odd.”
The takeaway from this piece is that the sky is not falling, the virus is mutating–that’s what viruses do– but the variations are becoming less virulent, and humans are learning to live with it. That there is a “refusal to acknowledge the evidence” is not odd, but de rigeur. Whatever happened to follow the science?
Because of the measured tone backed up by (seemingly) reliable data, this piece should be banned as the wacko propaganda of the extreme right. It might people less afraid, and the boffins cannot have that. FaceBook–please take this down!
Taken to its logical conclusion, it would seem that somewhat normal life can be resumed, which is not what the boffins–Neal Ferguson, Jacinda Ardern, Fauci, Sleepy Joe Biden and their ilk–want to hear. If there is no emergency, then it follows that there can be no emergency measures. Freedom, not an inherent right but something granted by the leave of our betters, might be returned to the people. This is something that the nanny state will fight to their last breathe….
This piece shows what an absolute clown show the “scientific” advisors have run, and the BBC is one of the worst offenders. Their “scientific” experts have touted “vaccine equity” at every turn, virtually demanding that we in the West not have second shots, not have booster shots until the corrupt and feckless Third World attains “equity.” THE “EXPERTS” still push this argument to an extent, but since it has failed to gain much traction on the ground, speak wistfully that this has not happened. Simply disgusting!
The issue with the IC model as well as all the other models is not how well or badly the code is written, or whether the code is spaghetti code, but rather an understanding of what modeling is about under circumstances where there are many unknowns (both known and unknown). Ferguson et al. have fooled the politicians and public health authorities by making the simplest model of infectious disease spread as complicated as possible, thereby giving the appearance of increasing sophistication and accuracy (just as the climate modelers have done). Unfortunately nothing could be further than the truth.
You don’t need a model to know that if you isolate everybody from one another, the number of cases will go down, but as long as immunity within the community is not established, then once the mitigation procedures are removed things will start up again just as they did from the get go.
Take masks, for example. It was argued that these are a very effective method of source control based on idealized lab experiments. Let’s assume this is correct and let us assume that masks are close to 100% effective in this regard. Then clearly after 6 weeks or so where the infected at the time of mask introduction will have either survived or died, and one might therefore argue that the respiratory tract infection, in this case COVID, would all but disappear. But this argument has a major and fatal flaw because the uninfected population would still be naive to the virus, so that the moment strict mask mandates were lightened up, all it takes is just one infected person to start the whole thing back up again. And that’s precisely why masks, social distancing, and other mitigation procedures don’t work. All they can do is flatten the curve while leaving the area under the curve unchanged. And the only escape is either through natural immunity, sterilizing (or close to sterilizing) vaccination (which we clearly don’t have as the current crop of vaccines have clearly failed), or effective early at-home treatments.
In other words, fighting VOCID is a bit like fighting a war. If one is going to get into a war, one has to have an end-game (and as recent history has shown the US has failed to have an end game in their middle east adventurers, and hence each one has ended up in disaster and left the place worse off than when they came). The end-game with COVID that Fauci et al envisaged was vaccination. Clearly if one can make use of mitigation measures to hold things down for a sufficient length of time to enable an effective vaccine to be developed, the mitigation measures would have been worthwhile. But this requires an effective vaccine to be developed, not one that lasts at best for 6 months after 2 shots, and only provides limited protection for 10 weeks after a 3rd shot. Further, it is not evident that the mandated mitigation procedures introduced, for example, in the US and UK, were any more effective than the completely voluntary changes in behavior that took place in Sweden.
Fortunately for us, it would seem that the virus will have done the work of the public health authorities for them by mutating to a less virulent, albeit more transmissible (supposedly) variant. That’s what happened with the Spanish Flu, and it’s clear that Omicron represents just such a development with COVID. And if that is indeed the case, as it certainly looks as if it is from not just the South African data but also the Danish experience, then we should all be rejoicing the advent of Omicron as a great Christmas and New Year’s present to all.
The problem with the IC model was that it was a fraud. The projection on mortality rate was based on the number of confirmed cases and calculated as a percentage.
The IC model had only one purpose and it served it well. It forced the governments to abandon the existing pandemic response plans, created after SARS and lock down the healthy. It’s purpose was to start the death spiral of big reset.
Omicron may be our salvation. The health officials in almost every country, except South Africa, are desperate. The big pharma criminal corruption that resulted in vaccination mandates, forcefully vaccinating children, including unborn children, is being exposed. That means that the Pfizer bonuses may stop coming. In Israel the practicing doctors are warning against using the booster against Omicron, no evidence they say, the panel advising the Israeli government is pushing desperately for the boosters.
The thing I find most frustrating about UK media (and I’m sure they’re not alone in this) is the way they reel off numbers so haphazardly that the viewer/listener is always confused by the lack of context even if not panicked.
For example:
They almost never compare cases with test numbers when comparing countries.
So you would never know unless you looked up WorldOMeter that UK test numbers are 5,781,814/1m pop; 5 times that of Germany and 16 times South Africa’s at 349,893/1m pop.
Excellent article, with some predictions that make sense. Maybe Engineers should replace that bunch of epidemiologists modellers…
“Sir Patrick Vallance and Chris Whitty, the UK’s chief scientific and medical officers, seem to be the last to admit what has been evident to most observers for almost a month: namely, that Omicron is much milder than Delta.” SAGE’s minutes in December are illuminating. The last mention of South Africa is in minute 99 when it referred to rapidly increasing hospitalisations at a time when cases were starting to plateau. Minutes 100 and 101 have no mention of South Africa at all so no good news was communicated to the Government
SAGE have already been exposed as an essentially corrupt organisation… witness the recent Twitter exchange between Fraser Nelson and the chairman of SAGE.
Bad as they are, the ‘independent’ SAGE are even worse.
The UK has been using omicron to scare people into getting the booster. That’s why there’s been no acknowledgment of omicron’s mildness. Could be worse I suppose.
The abuse of mathematical models for political purposes during covid undermines the population’s confidence in modelling in general, which could have far reaching consequences, e.g. climate change predictions.
Politicians may well have done more long term damage than they realise.
As someone with considerable experience with engineering modelling I find the repeated excesses/wild predictions of the IC model to be difficult to understand. Repeated refining of model predictions based on actual data normally achieves an asymptotic approach to high confidence predictions. This appears to be completely absent from the IC approach.