Thread by @_johnbye

The government seems to have published a shelved PHE study today just to defend Matt Hancock against accusations by Dominic Cummings about discharging patients from hospitals to care homes without testing them for covid.

And the study is flawed, due to that very lack of testing!

https://twitter.com/kallmemeg/status/1397940294524575762

It's been widely assumed that many deaths in care homes in the first wave were caused by patients being discharged from hospital without being tested. A policy that continued until April 15th.

This issue was raised by Cummings in his testimony yesterday.

dailymail.co.uk/news/article-9…

The study looks at this issue and claims it only caused 286 deaths.

It was conveniently published today, just in time for Jenny Harries to reference in a press conference.

But it's dated April 2021.

And it's clear it was actually written last autumn!

gov.uk/government/pub…

The report talks about data for the last six weeks (up to October 12th!) being undercounted due to "delays in reporting".

The conclusion even says it should be "re-assessed in November and December"!

It's not clear if this was done. Or why they sat on the report for 7 months.

Maybe because the study is flawed?

Understandably it only counts outbreaks, cases and deaths confirmed by testing.

But early on care homes struggled to get tests at all, and those that did were rationed to testing the first 5 residents with symptoms.

bbc.co.uk/news/uk-522896…

So data for the first wave, which is when most seeding from hospitals took place, is very incomplete, and will underestimate cases.

Which would be fine, except they've defined the first wave as everything up to October 12th! Well into the second wave for the North and Midlands.

This means they're mixing data from when seeding from hospitals was most likely to have taken place but there was almost no testing in care homes, and data from when there should have been very little seeding but care homes were being routinely tested.

Which will skew the data.

To give an idea of the possible scale of the problem, the study identifies 5,882 lab confirmed outbreaks up to October 12th.

But the last PHE care home outbreak report (which also includes suspected outbreaks) already showed 6,811 by July 19th.

gov.uk/government/sta…

So both the number of care home outbreaks and the number of cases and deaths associated with each one is significantly undercounted for the period when most seeding took place.

But not in later data when patients were being tested before discharge, reducing seeding.

As the study averages things out over the entire pandemic, up to the point when it was really written in October, the proportion of care home deaths and outbreaks caused by hospital discharges in the ACTUAL first wave will be FAR higher than the 2% it quotes for the "first wave".

It's hard to interpret this as anything other than dusting off an old report that's been sitting on a shelf for the last seven months and publishing it purely to provide dubious cover for Matt Hancock in today's press conference. 🙄