DIRECT study results much poorer than presented?

[ianf0ster]

I have been seeing tweets about the NHS Shakes and that the DIRECT 2yr study results don't justify the Very Low Calorie Diet push.
Indeed, if the underlying data is as tweeter @AmandaZZ100 says, I would consider it scandalous that they have been presented as such a success!

Relevant Twitter thread:

https://twitter.com/x/status/1302903680950300672

 

[Eddy Edson]

Amanda needs to take a breath and go back and read the paper more carefully.

 

[Deleted member 21371]

I think it's quite laughable how some people can't accept that there is a now a recognised treatment that works for diabetics, and are desperate to blow the dog whistle for their followers.
Let them get on with it, it won't affect the ones who do achieve reversal of their diabetes.

The correct analysis of the results are in the report for anyone to read.

Direct Trial

https://www.directclinicaltrial.org.uk/Documents/Patient%20Info%20Website%20Feb%202018.pdf

 

[ianf0ster]

Amanda needs to take a breath and go back and read the paper more carefully.

You have access to the raw data do you? That is what she claims her analysis was done on.
So where is she wrong? - Specifics please.

 

[Bruce Stephens]

I think it's quite laughable how some people can't accept that there is a now a recognised treatment that works for diabetics, and are desperate to blow the dog whistle for their followers.

For some people with (Type 2) diabetes. Nobody's claiming it always works.

 

[Eddy Edson]

You have access to the raw data do you? That is what she claims her analysis was done on.
So where is she wrong? - Specifics please.

She refers to the paper and suggests that it says only 4 people out of 52, all from the control group, achieved < 42 mmol/mol.

The paper doesn't say that.

 

[Bruce Stephens]

She refers to the paper and suggests that it says only 4 people out of 52, all from the control group, achieved < 42 mmol/mol.

The paper doesn't say that.

No, though it does report that 4/119 of the control group had HbA1c <42 mmol/mol after 24 months. It doesn't seem to say anything about who else achieved that (though I'd expect them to have mentioned if anyone else had done so) since they're defining remission as <48 mmol/mol.

(Doesn't seem that helpful to look at who got <42 given that that doesn't seem to be what they were looking for.)

 

[Eddy Edson]

No, though it does report that 4/119 of the control group had HbA1c <42 mmol/mol after 24 months. It doesn't seem to say anything about who else achieved that (though I'd expect them to have mentioned if anyone else had done so) since they're defining remission as <48 mmol/mol.

(Doesn't seem that helpful to look at who got <42 given that that doesn't seem to be what they were looking for.)

That part is referenced just to "those on anti-diabetic medication". That excludes everybody in the intervention group who were in remission at month 24, so you can't infer anything about how many in the intervention group got to HbA1c < 42.

But it is a pity that the paper doesn't include more granularity on absolute HbA1c outcomes, rather than just changes.

EDIT: I guess the info might be in a SI for the paper but I can't find access to it.

However, I can find the SI for the earlier 12 month results paper: https://www.ncl.ac.uk/media/wwwncla...files/DiRECT1yrPaperSupplementaryAppendix.pdf

This gives more detailed HbA1c outcomes - see p4. At 12 months, 29% of the intervention group had HbA1c < 42, versus 4.7% for the control group.