No such thing as a safe level of drinking

[999tigger]

The study is not quite what it seems.

 

Firstly they only looked at people who were at least 50 years of age, so the results are not necessarily relevant to people under this age

 

Secondly because they only included lifetime teetotallers (excluded people who used to drink but stopped), some of the sample sizes are tiny (16 people in one group, 19 in another) leading to huge margins of error making most of the results meaningless

 

The unadjusted results showed "protective effects [of drinking] were identified across a broad range of alcohol usage in all age-sex groups" - which I'm sure you would agree is certainly not what the headlines are suggesting.

 

The only group in the study which had a sufficiently large size for meaningful results were women 65 or over, and for this group the study found that there WAS a protective effect. For other groups no conclusions can be drawn because of the stupidly high margins of error.

 

The report then goes on to suggest a number of reasons why the observed protective effects may not actually exist, although they don't offer any evidence for these theories.

 

The conclusion of the report is that moderate alcohol intake is beneficial to women 65+, but there is no evidence of beneficial effects in any other group.

 

It would be just as valid to conclude that there is no evidence of any harm in any other group.

 

The report:

http://www.bmj.com/content/350/bmj.h384

 

A very interesting analysis of the data in the report:

http://understandinguncertainty.org/misleading-conclusions-alcohol-protection-study

 

EDIT: Just to clarify, pregnant women were not considered in this post

 

Are you claiming thats the study the CMO used for the guidelines?

[andysm]

This is the study that was the basis of headlines in some newspapers.

[999tigger]

This is the study that was the basis of headlines in some newspapers.

 

So its just a single study and when you refer to it as the study thats actually misleading in the context of this thread because the CMO and experts looked at 44 studies. You also have no idea whether these experts were able to take into account the flaws you highlight in your study, which would form the basis of any critical analysis.

 

---------- Post added 11-01-2016 at 14:42 ----------

 

I forgot to add thumbs ip for A reading a report and B making points of analysis based on it. A rarity.

[andysm]

I wasn't trying to mislead anyone, others in this thread have mentioned newspapers and stories on websites, which I would guess is where most people get their news these days.

 

The Alcohol Guidelines Review report is here, but I don't have enough time to read and digest it all:

https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/489797/CMO_Alcohol_Report.pdf

[999tigger]

I wasn't trying to mislead anyone, others in this thread have mentioned newspapers and stories on websites, which I would guess is where most people get their news these days.

 

The Alcohol Guidelines Review report is here, but I don't have enough time to read and digest it all:

https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/489797/CMO_Alcohol_Report.pdf

 

The stories are about the CMO guidelines which looked at 44 studies though and not about your single study. Already linked the guidelines. There are even more detailed reports as to how three of each working committees comprised of academics and clinicians rebiewed the data and reached their recommendations. I'm just pointing it out because several teams of people seem to have spent a significant amount of time doing a comprehensive review of what all the latest studies tell us and not as some suggest just made some arbitrary decision based on zilch.

[andysm]

I was referring to headlines such as "Drinking is only good for you if you are a woman over 65" and "Alcohol has no health benefits after all", these headlines were based on the report I mentioned.

 

The new guidelines report itself shows some health benefits to drinking 7 units per week (tables 10 and 11), which doesn't quite fit in with the 'no safe level' message we are being given.

 

I notice that Sheffield University were involved in parts of the new guidelines report.

[PeteMorris]

Is it my imagination, or has 'drinking' only just (in this advice) been linked to cancer? I don't recall it being linked previously...But of course I could be wrong.

[RonJeremy]

Is it my imagination, or has 'drinking' only just (in this advice) been linked to cancer? I don't recall it being linked previously...But of course I could be wrong.

 

I know. They'll be saying smoking is bad for you next.

 

Don't believe a word of it.

 

 

... and bacon.

[999tigger]

Is it my imagination, or has 'drinking' only just (in this advice) been linked to cancer? I don't recall it being linked previously...But of course I could be wrong.

 

Seems its been linked for a long time.

 

The International Agency for Research into Cancer (IARC; part of the World Health Organisation) has classified alcohol as a Group 1 carcinogen since 1988 [1]. IARC's rulings are the gold standard in terms of determining if something causes cancer, and Group 1 is their highest risk category. It means that there is convincing evidence that alcohol causes cancer in humans. More recent reviews by IARC and other agencies have also concluded that drinking alcohol causes cancer [2-6].

 

A study published in 2011 found that alcohol is responsible for around 4% of UK cancers, about 12,800 cases per year [7, 8]. The proportion of cases down to alcohol was highest for mouth and throat cancers (around 30%), but bowel cancers accounted for the greatest overall number of cases linked to alcohol (around 4,800 cases a year).

[milquetoast1]

I've been following the blog of medical statistician Adam Jacobs for some time, and here's his interesting thoughts on the new guidlines:

 

http://www.statsguy.co.uk/new-alcohol-guidelines/