UK: Almost 10 per cent of coronavirus deaths were not related to Covid-19

[elSquid]

https://www.yahoo.com/news/almost-10-per-cent-coronavirus-185438040.html

The Department of Health has been urgently reviewing the way in which it records deaths after Oxford University noticed in July that former coronavirus patients were being included in mortality figures even if they had recovered and then died of something else.

On Wednesday, the true extent of the problem emerged when PHE published a report showing that 3,664 people who had been included in 40,160 English coronavirus deaths did not have Covid-19 on their death certificate.

It is also now clear that England's death rate has been diminishing far faster than official figures showed. Since the middle of June, at least half the reported deaths have not been due to coronavirus and have now been excluded from official figures.

(Snip)

Keith Neal, emeritus professor of the epidemiology of infectious diseases at the University of Nottingham, said that, if the data error had not been spotted, the number of deaths of Covid-19 survivors would have continued to increase and made tracking the epidemic impossible.

"The previous measure of always being a Covid-19 death even if recovered was unscientific," he said. "As Covid-19 deaths fall, the number of recovered patients, particularly the very old and those with severe underlying conditions, are now dying from these conditions and not Covid-19.

"These non Covid-19 deaths in survivors would become an ever increasing percentage of the England Covid-19 deaths being reported. It had become essentially useless for epidemiological monitoring."

( more at link )

That is a fairly large mistake to be made. Of particular note is the last quoted paragraph; if the error continued, no doubt it would be driving public policy in a factually incorrect manner.

— Michael

[theLBC]

probably worse and that is all they discovered so far, and the same thing is happening here with people that die from leukemia, acute alcohol poisoning or a motorcycle accident are listed as covid-19 deaths.

[sd_shooter]

Shocked, but not shocked. The errors are likely even worse here.

[balgor]

https://www.theguardian.com/world/2020/jul/21/analysis-why-englands-covid-19-death-toll-is-wrong-but-not-by-much

Are official figures overstating England's Covid-19 death toll?

https://fullfact.org/health/coronavirus-death-count-data/

Changes to government Covid-19 death data won’t make a big difference to the overall death count

The Public Health England (PHE) numbers were just wonky. They recorded deaths by tracking anyone (I assume through their EMR system) who tested positive and periodically checking whether they were still alive. It's a bad system b/c using only positives will undercount and not setting a time limit for counts will overcount. The much more reliable source for death data is from death certs, which we use, and the Office of National Statistics (ONS) uses:

https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/articles/coronaviruscovid19roundupdeathsandhealth/2020-06-26

ONS covid death toll is 51,710 as of July 1st, higher than the revised PHE of 41,329 or unrevised 46,706 numbers(August 11th) anyhow. The ONS number is just like our National Vital Statistics System (NVSS) number, so it's lagged (no idea how long the UK data is lagged, not familiar with their ONS), but it strictly uses death certs.

[SAN compnerd]

Why I'm shocked!



The Brits love to cook the books on health and crime stats, just look at how they count murders and crimes, they don't go on the books until the criminal is convicted. This results in odd spikes in stats like when a serial killer was caught a couple of decades ago and his years of murders went on the books all at once.

Seems kinda weird to not count murders as soon as they are committed, but that's the Brits for ya.