Todays (20th May) news on the risk of Dying from Coronovirus with Diabetes

[Crispycrystal]

Todays news says that those people with type 1 are 3.5 times more at risk of dying if they become so ill they need to be hospitalised but those with type 2 are only twice as likely.

NHSE statistics on coronavirus deaths in people with diabetes

A study from NHS England revealed that one in three people (33.2%) who have died in hospital in England following a diagnosis of coronavirus also had diabetes. We know this statistic raised a lot more questions than it did answers, and it may have left you feeling increasingly worried.

www.diabetes.org.uk

Figures from elsewhere (Gov.uk website) show that in 2019 approx. 8.1% of the UK population have Type 2 diabetes and about 0.9% have Type 1.
So out of 23804 deaths in hospital, you would expect 8.1% (1928) would have Type 2 and 0.9% (214) would have type 1 if diabetes was not a factor (ie same ratio as the whole of the population)

The new data shows us that of the 23,804 deaths recorded in the study, 7,466 (31.3%) who died from coronavirus had type 2 diabetes, 7466 is 3.85 times bigger than the expected 1928 So Type 2 are 3.8 times more likely to die than those without diabetes.

365 (1.5%) had type 1 diabetes, 365 is 1.7 times bigger than the expected 214 So Type 1 are 1.7 times more likely to die than those without diabetes.

Has the article got it the wrong way around, or am I calculating it incorrectly?

 

[Browser]

See the thread below this about discussion on this topic. I was confused like you and posted my concern/ confusion. According to a reply to my post, it seems the figures given relate to Type 1 and Type 2 in relation to per head of the population. As I understand it, the chances of death for Type 1 are greater on average because there are fewer Type 1s than Type 2s in the demographic. Like you, I think the information is confusing and badly presented.

 

[Docb]

The problem is that the politicians and media want to put out simple summaries of complex information and all you get is confusion because none of them really understand it. I have a a bit of a look at the base data being used for all these announcements and am struggling to follow their trail.

 

[Bruce Stephens]

Figures from elsewhere (Gov.uk website) show that in 2019 approx. 8.1% of the UK population have Type 2 diabetes and about 0.9% have Type 1.

Those aren't the figures used in the paper, https://www.england.nhs.uk/wp-content/uploads/2020/05/valabhji-COVID-19-and-Diabetes-Paper-1.pdf

Of the 61,414,470 individuals registered, 263,830 (0∙4%) had a recorded diagnosis of Type 1 and 2,864,670 (4∙7%) of Type 2 diabetes.​

There's probably a bit of difference in precise definitions. This paper's using what GPs (and specialist services where they have those figures) have. They're also adjusting the risk by age (since it's known that age is very significant). (It's also England only, though the proportions are presumably about the same.)

 

[Crispycrystal]

See the thread below this about discussion on this topic. I was confused like you and posted my concern/ confusion. According to a reply to my post, it seems the figures given relate to Type 1 and Type 2 in relation to per head of the population. As I understand it, the chances of death for Type 1 are greater on average because there are fewer Type 1s than Type 2s in the demographic. Like you, I think the information is confusing and badly presented.

Thanks did you put a link in to the other thread. I looked up the per head figures and used those to calculate. The article still seems to have got it wrong but the sample is probably not statistically significant anyway.

 

[Bruce Stephens]

but the sample is probably not statistically significant anyway.

The authors seem careful enough, and give the usual 95% confidence intervals (some of which are rather wide, admittedly). I'm sure it was done in a hurry (and hasn't been through peer review) so maybe it'll change. They also mention a few things which don't reach statistical significance (at the usual p<0.05 level).

 

[everydayupsanddowns]

This was discussed on this thread @Crispycrystal

Latest Covid 19 info

Unfortunately the data (which is a breakdown of hospital deaths in England) does not combine multiple data sets, so does not include pre-existing conditions by age. However it reveals the following: Of those aged 60–79: 9,346 died (38.7% of all deaths) of which 8,723 had pre-existing...

forum.diabetes.org.uk

 

[Crispycrystal]

Thankyou