There follows a guest post by our in-house doctor, who notes the NHS has revealed that almost a quarter of ICU beds in English hospitals are unoccupied in the middle of January, which is unprecedented in recent history – normal ICU bed occupancy at this time of year is well over 90%. Why then the continued narrative of fear and doom in some quarters?
On Wednesday January 19th, Boris Johnson announced in the House of Commons that the majority of ‘Plan B’ restrictions would end in England on January 26th. Any objective analysis of the U.K. data confirms the information from South Africa in early December in relation to the mild nature of Omicron was correct. Finally, the Government has been persuaded that our South African colleagues did know what they were talking about.
After a brief incursion onto the territory of Libertaria, the forces of Hysteria have been repelled and British citizens partially regain their sovereignty. Time to hang out the bunting and open the champagne? I think not.
Reaction by lockdown zealots to removal of societal restrictions has been tediously predictable. In an echo of July 2021, a range of commentators are again using compliant mainstream media outlets to predict imminent catastrophe if societal restrictions are lifted and to condemn the Government for putting public safety ‘at risk’. Before looking at warnings of doom in more detail, I will briefly examine the data which have led to the volte face by the Government. I will not dwell too long on these numbers as most readers will already be aware that the “nailed on tsunami of admissions” has simply failed to show up. Again.
Graph 1 shows the most important information this week from the Primary Diagnosis spreadsheet. Readers will recall this shows the total numbers of patients in hospital on vertical orange bars, versus the number of people in hospital with acute Covid on blue bars. The ratio between the two is depicted by the grey line. Notice the blue bars (people ill with Covid) are falling on the right-hand side of the graph, as is the ratio. So, when the BBC reports that Covid inpatient numbers are about 15,000 and static, the true picture is there are fewer than 8,000 patients and the numbers have been falling for the last 10 days.
Graph 2 shows the true numbers for London, which front runs the rest of the U.K. True numbers of Covid patients are 40% of the daily reported figure and numbers have been falling steadily since January 4th.
The pattern is repeated in all datasets I have examined. This week the ICNARC report revealed that almost half the 552 patients testing positive for Covid in English ICUs did not have Covid as the primary reason for admission. Before Christmas, when the main variant was Delta, 95% of patients reported as being in ICU with Covid, actually were ill with it. The demand for ICU care for Covid patients has fallen off a cliff since Christmas. In mid-December, 9% of patients admitted to hospital for Covid needed ICU care. Now that percentage is around 2%.
Readers can access the report here. The important graphs are Figures 29 and 30 on pages 53 and 54.
Even more astonishing is the critical care capacity information from the weekly NHS hospitals report. On January 18th, there were 2,602 patients in English ICUs being treated for non-COVID conditions and 645 patients being treated with positive Covid tests (though about 200 of these were not unwell with Covid). On that day, the NHS reports that there were 976 available critical care beds in English Hospitals.
The excel spreadsheet can be accessed here under the link “weekly admissions and beds January 20th 2022”.
Readers may not appreciate the significance of the snippet. In its own data, the NHS has revealed that 23% of ICU beds in English hospitals are unoccupied in the middle of January. This is unprecedented in recent history. Normal ICU bed occupancy at this time of year is well over 90%.
My first thought on considering this figure was that the denominator must have changed – i.e., the total number of critical care beds must be greater than usual due to expansion during the pandemic. However, the total of the numbers presented is 4,200 critical care beds, which is about normal capacity in pre-pandemic terms. So the reduction in percentage utilisation appears correct. Absence of influenza for the second year in a row may account for some of the slack and the postponement of urgent surgery ‘because of Covid’ may also play a part, but by any standards this is a remarkable observation.
The 103rd meeting of SAGE on January 13th notes that: “The increase in hospitalisations… anticipated following the observed increase in cases in older age groups, has not been seen so far.” Readers may recall a huge increase in admissions of elderly people was confidently predicted after social gatherings around Christmas. Failed to show.
Another huge increase in admissions was predicted when the school term started. Failed to show.
In fact, all the predictions confidently made in December have failed to show.
Graph 3 shows just how wrong the projections were. It comes from the paper presented on December 30th by a group frequently advising SAGE and shows the mid-range prediction. Readers should note that even the lowest range prediction (10% severity) still over-called the number of actual admissions. Naturally, the worst case model (100% severity of Omicron compared to Delta) was the one reported in the press. This over-called actual admissions by more than an order of magnitude – predicting 26,980 peak admissions per day on January 14th against an observed number of 1,692.
The same pattern was repeated in all published projections in relation to total inpatients and Covid deaths. By any objective analysis, this represents another catastrophic failure by expert advisers to Government. Yet standard excuses are again trotted out by the authors of these documents to evade accountability for their repeated mistakes. It is not sufficient for professional modellers and public health doctors to abrogate responsibility for the consequences of their actions by hiding behind such excuses. If there are no consequences for repeated errors, those errors will continue to happen (so called absence of Moral Hazard). Steve Baker MP has highlighted how faulty modelling was used to coerce the government into imposing an unjustifiable lockdown in October 2020. I suggest the same tactics were used before Christmas 2021 by the same people with the same result.
I will now look at recent warnings of imminent cataclysm. Readers will recall in past contributions I have drawn attention to various statistical tricks used in data presentation. Alteration of axis scale, cropping of timescales to exaggerate or diminish trends, use of percentage variation rather than absolute numbers to exaggerate rates of change and many more.
This week has seen literary equivalents being used to confuse and mislead the public. Techniques such as conflation, elision and obfuscation have been used – I will discuss examples of each.
Conflation merges two different sets of information into one. In this article published in the Financial Times, the author conflates Covid with malaria. I have seen similar erroneous comparisons to TB. The central theme relates to the impending endemicity of Covid – that the virus is now with us forever and will resurge from time to time in a manageable way. The author asserts malaria is the same thing – endemic in many parts of the world, but still kills many people every year. She forgets to mention the central point about malaria – 80% of malaria deaths occur in healthy children under five because of lack of acquired immunity. With Covid, the majority of deaths have been at the opposite end of the age spectrum and in people with pre-existing medical conditions or obesity. There is still no effective vaccine against malaria, whereas Covid vaccines are helpful in reducing severe disease and death, as are newly approved monoclonal antibody medicines designed to reduce disease severity in vulnerable people. For indigenous adults in malarial regions, malaria is a mild seasonal infection, coinciding with the end of the rainy season when mosquitoes start breeding – this manifestation of naturally acquired immunity is a more accurate comparison between the two diseases. The implication that Covid will remain a mortal threat to humanity just does not stack up.
In this press release from the British Medical Association, we see the use of elision – the process of merging different things into one.
In relation to the Prime Minister’s announcement of the revocation of Plan B, Dr. Nagpaul states: “This decision clearly is not guided by the data. When Plan B was introduced in December, there were 7,373 patients in hospital in the U.K. – the latest data this week shows there are 18,979.”
When Plan B was announced on December 8th, the vast majority of Covid inpatients in the U.K. were infected with the Delta variant. Today, almost all the patients have Omicron (now known to be far less severe). From information provided by the Primary Diagnosis spreadsheet, we know that on December 8th, about 75% of the headline figure for Covid inpatients were admitted with acute Covid, but now that number is around 50% and falling. Finally, in early December, admission rates were trending upwards, now they are firmly trending downwards. The BMA is not comparing like with like, so the comment that “this decision clearly is not guided by the data” is, in itself, clearly not guided by the data.
Obfuscation (being evasive, unclear or obscure in the telling of facts) can be seen in this excerpt , again from the Financial Times.
Christina Pagel, a Professor and member of the Independent SAGE group of science advisers, said that the Prime Minister was dropping restrictions to “appease” the right wing of his party “at the cost of the population’s health”.
Professor Pagel asserts that the removal of Plan B is a political act rather than a rational response to the data. Her comment reveals more about Professor Pagel’s personal political preferences than an objective analysis of the numbers and is calculated to obscure the reality that societal restrictions are neither necessary nor desirable.
The deliberate use of emotive language or graphic images to influence behaviour through fear has been a common thread since March 2020. If we are to end the confected ‘perma-crisis’ and restore normality, emotion needs to defer to rationality and a clear-headed quantitative assessment of harms versus benefits in relation to non-pharmaceutical interventions must be undertaken. Failure to curb malign exploitation of fear for political ends by parties with specific agendas and hidden incentives will perpetuate the cycle of poor decision-making and worsen the eventual outcome.
As Judge Dredd remarked: “Emotion? There ought to be a law against it.”
I’m certainly not going to argue with him.
This post has been corrected. An earlier version stated that 30% or nearly a third of ICU beds are unoccupied instead of 23% or nearly a quarter.