Flip Flop Flu, WHO Rule 3 & The Blip Of COVID (part 2)

A festive frolic involving a bit of time series modelling of deaths from influenza and pneumonia so that we may see what we once saw

In part 1 of this series we homed-in on estimates of the impact of WHO Rule 3 on the annual count of pneumonia & influenza death across England & Wales. In doing so we discovered a hole in the tally for 2020 and beyond that may well arise from coding errors, with pneumonia & influenza death being erroneously ascribed to COVID. I also strutted around a bit after dishing out what I considered to be a cracking time series model¹!

What I want to do today is take my modelling/cake making at face value and make some simple corrections to the database so we may see what reality may have looked like before the ONS bowed down to the global code mangle wielded by WHO. Before that is done I ought to use the same baking method to check for missing pneumonia & influenza deaths in 2021 and 2022 as well as 2020. Here’s the nubbins:

This is telling us what that decay factor of 0.8 has already been telling us; that missing deaths (likely due to dubious coding under WHO Rule 3) are not just isolated to 2020 (19,675 missing deaths, p<0.001) but spill over into 2021 (16,747 missing deaths, p<0.001) and 2022 (11,259 missing deaths, p<0.001); these are the figures I shall use. One thing we can be sure of is that such coding errors aren’t a passing blip on the radar but are going to be clinging to ONS data like dog doody to a shoe.

Yer Tis

A local colloquialism translated as ‘here it is’, I find that one may learn to speak proppa using this guide.

After correcting for what I boldly deem to be 2020, 2021 and 2022 COVID coding errors, and after throwing WHO Rule 3 back in the trash can, I arrive at this slide:

That’s proper job and make no mistake! See how neatly the two series blend when layered with a bit of jam? And see how the annual tally continues at a steady 50,000 – 60,000 deaths per year? And see how that hole centred on 2021 now becomes a blip during 2020? Yes indeed, and zounds, for it sure looks to me like another burst of bog-standard influenza did the rounds back then but got re-labelled as COVID due to terrible testing, dodgy diagnosing, corrupt coding, malevolent managing, rubbish reasoning, and all the rest. If I calculate the crude mortality for pneumonia & influenza it now looks like this:

That makes a lot of sense to me; that is to say, a steady state mortality for recent decades makes sense if we are not making strides in wonderful new drugs and treatments. And no, I’m not buying the flu vaccines are truly magical things story unless somebody can put an honestly-conducted, genuine, gold standard RCT report on my desk that isn’t flawed. Better still, a quality meta-analysis of several such studies by an independent team with impeccable credentials whose grant award didn't come from big pharma or certain ‘philanthropic’ individuals.

Gone is the flip flop flu and instead we have a nice slice of something with custard spreading along at around 100 deaths per 100k, with a tiny nutty bit at the end. I can testify that 2020 saw nut clusters ‘coz I got clobbered by a nasty flu-like bug during December 2019 that lasted several weeks – so did many folk in these parts and thereabouts (including those who worship flu vaccines).

COVID Cobbler

So what happens with the tally for COVID now that it has been trimmed to the correct size? And how does this count compare with the adjusted count for pneumonia & influenza once we abandon WHO Rule 3?

Well, this is what has come out in the wash as a result of all this ARIMA intervention modelling and pressing of marzipan into the corners:

Surprising innit? I’m glad I didn’t bet for my money was on über COVID making all the running. Instead, we are looking at a cobbler with a soggy bottom. We now see that deaths due to COVID did not dominate as some rather unscrupulous officials would have us believe. Quite how many COVID deaths should be counted under the umbrella of pneumonia we cannot be certain for there will be considerable overlap; this means I cannot add the two time series together to arrive at a grand respiratory curve. Indeed so, for if it is argued that COVID must necessarily be a subset of pneumonia (viruses may create havoc but they don’t do the killing) then the sea green line effectively becomes that grand respiratory curve under which COVID takes it's rightful place along with other pathogens and conditions.

Taking this line of reasoning and bashing it with a rolling pin we arrive at this ARIMA model:

This is another decent-fitting model but this time the table of outliers doesn’t contain an entry for 2020, 2021 or 2022, which suggests that nut cluster we see is merely seasonal respiratory business as usual. Here’s the pudding:

So there we have it – a two part series inspired by my grapple with La Grippe that reveals just what can be achieved by fiddling with coding at a strategic level and setting the global healthcare system down a path such that a (non-lethal) pandemic can be conjured numerically at will when the conditions are right.

Death Is Complex

As I have stated time and time again death is rarely a black and white affair, which is why my former cardiac surgery unit coded up to seven causal categories that were offered to surgical teams as multiple choice. Trying to squeeze a multicausal chain of events into a unicausal coding-hole is what the ONS/WHO are all about and this opens the door for nonsense, over-cooked tripe and abuse by officialdom. You can see how this comes about by the idiotic design of the UK death certificate that attributes a single underlying cause of death as well as a single direct cause of death.

A physician may enter respiratory arrest under section 1a as the direct cause of death, followed by pneumonia under section 1b with COVID-19 under section 1c. This then becomes a ‘COVID death’ and not a ‘pneumonia death’, which is particularly irritating if we learn that it was bacterial pneumonia or ventilator associated pneumonia. Then you’ve got the can of worms that is respiratory arrest brought about through intervention, drug reaction, comorbidities and shoddy patient management. These awkward corners of medical reality are not going to be declared on the certificate and will be scraped into the kitchen bin as we do with burnt toast.

And how about ‘respiratory arrest’ being accompanied by cardiac arrest, or kidney failure perhaps? Or sepsis? Or multi-organ failure? What happens to the ONS/WHO swanky unicausal MUSE coding engine when faced with real life (other than churning out utterly meaningless and rather misleading statistics)?

Death as a clinical endpoint can be seriously tricky to figure, which is why I’d sit down with a willing registrar on a regular basis and chew over casenotes in which SHOs had ticked cause of death unknown. Unknown?!!!! Yes indeed, ‘unknown’; because it’s quicker for an impatient (and super busy) member of the surgical team to tick any old box and get another sodding form out of the way.

Then we have those lengthy debates between two consultant surgeons thrashing out the likely primary cause of death when, just before agreement is reached, an anaesthetist opens their mouth. Worse still an intensivist (a tough breed of anaesthetist) who starts by saying, “well, what you’ve all failed to noticed is the liver function back at 5am the previous morning…”

N.B. A colleague tells me that anybody on their unit dying 3 - 6 weeks into admission isn’t going to be dying from any bug - deadly novel mutant coronavirus or nay - unless that bug is bacterial in nature and not responding to antibiotics.

Neither are bugs always to blame. Bizarre as though it may seem an in-patient can collapse in the toilet and that fall alone can induce pneumonia. Any decent SHO can tell you that. Whether or not that patient also tested positive for SARS-COV-2 should be irrelevant: such is the insidious and incredibly damaging game that NHS and other officials have played with people.

As if that wasn’t bad enough I am told that there are reports of families being offered COVID diagnoses for their deceased relative to avoid the lengthy process of post mortem. That’s just what we need. Words fail me. I no longer recognise the service in which I once served.

One thing we can be certain of is that COVID killed evidence-based medicine back in 2020. And now I see that we’ve got nose only masks…

Kettle On!

1. It occurs to me that ARIMA time series modelling is not unlike pressing marzipan into the corners to ensure a perfectly square cake prior to icing.