I Love the Smell of Confirmation Bias in the Morning

"Excess deaths" are all over Contrariworld at the moment. Some want to use them to undermine vaccines, others (with overlap to the first group) to undermine lockdowns. 

The Daily Telegraph has had the latter in a told-you-so frenzy with a truly awful headline (find the piece yourself!): 

"Lockdown feared to be killing more people than Covid"

The Twitter version appends 'effects' to 'lockdown', while the piece itself moves rapidly away from lockdown itself to issues related to the effects of the pandemic on healthcare. These would happen, probably more so, in the absence of mitigations, so attributing them to lockdowns is dubious. An inversion of reality, one might say. 

The core evidence of a concern is a snippet of the graph of excess deaths - the deaths over and above those expected based upon an average of equivalent recent periods.

In this period, the deaths directly attributed to Covid don't account for the whole of the excess. So what is it? Oh, let's just say ... lockdowns! Why? Don't like 'em. Riiiight... 

Let's pull back. Here's a chart I made in Excel from ONS data that can be downloaded from the 'data download' at the left of this page.

The Telegraph chart starts almost exactly where the yellow bars start to overtop the blue at the far right, which is misleading at best. As luck would have it, you can get a good idea when our mitigations were in place by noting the points the excess turns negative! These roughly coincide with restrictions easing. Now, I'm not making any causal argument of my own from this pattern, but it is at odds with the idea that lockdowns caused these recent excess deaths. If they do, the effect is subject to a considerable lag, and they may even be initially protective!

What are we to make of the extended period between December 2021 and April 2022 where Covid deaths were significant but the excess was nonetheless negative? If you subtract Covid, the excess is even more negative than depicted. Did we get any Telegraph articles wondering what the hell was going on there?

To reach the boldly stated conclusion requires elimination of other causes. One possibility is simply statistical artefact. Data is somewhat sensitive to the method used to define the 'expected' deaths. A more likely cause (though no cause should be considered exclusive) is a delayed pandemic-related parameter, such as

- a reluctance to access healthcare for chronic conditions

- healthcare itself being unable to service demand

- Covid sequelae

The latter are likely to be significant. Contrarians never cease to insist, misleadingly, that a death within 28 days of a positive Covid test is counted as Covid. A corollary of this is that a death more than 28 days later is not, even if it caused the damage that proved fatal. Of course this 28-day rule does not apply anyway when it comes to death certification, but it remains the case that increasing time will lead to fewer and fewer certificates indicating Covid, even if it set in train the events resulting in death. Therefore Covid may still be a contributor, even if not formally identified.

A friend reports having been given a diabetic eye test due to pancreatic damage from Covid. The nurse apparently said he was the 5th Covid-related case that week. Take that with a massive pinch of anecdotal salt of course - I am the first to arch an eyebrow at stories of 'my vaccine-damaged mate' - but it remains worthy of consideration when looking for causes. Others have suffered damage to other organs. We cannot simply pretend that Covid never causes long term problems, just as we should be objective about vaccine harms.

 

Finally, I'll relate a tale nicely illustrating the determination of people to leap to conclusions and stick to their statistical guns. "Look at excess deaths!" I was told. So I did, which is what prompted me to construct the chart above, showing a pretty clear relation between Covid and the excess. That wasn't the bit I was supposed to be looking at, but you can't cherry-pick the bits that support your case and put the rest down to 'bad data'! There followed an increasingly elaborate multiplicity of causes for the inconvenient patterns. Occam would be turning in his grave.

"The first spike is due to increased midazolam use in care homes". 

That is, rather than accept a significant fatality rate from 'true Covid', they felt a more plausible explanation to be that doctors were wholesale turning into Harold Shipman, descending on care homes, administering lethal doses of a sedative to people who didn't need it, and falsifying the deaths as Covid. Huge if true. 

"The second spike is due to vaccine rollout".

So even though that spike was well underway in December 2020, vaccination then caused so many additional deaths among the relatively small numbers of people vaccinated by spring as to account for the excess - all certified as Covid. Again, you'd think someone would notice. You'd also wonder why this murderous blip ceased as the rollout progressed through the summer.

I was dismissed as brainwashed by the mainstream media, a cross I'll have to bear.

Talking of mainstream media, have you seen this article in the Telegraph? 🤔