Friday, August 24, 2018

"That Next Drink May Kill You" And Other Stories


You had a drink and now you're dead.
Unless you've been living under a rock, you've seen the headlines: "No amount of alcohol is good for your overall health, global study says," "Alcohol was responsible for nearly 3 million deaths in 2016, study says," "Health risks of alcohol outweigh benefits, study says."

The study was published in The Lancet. You've heard of The Lancet, right? Lots of medical stories start out there.

Read down a bit and you discover that the study was funded by the Bill and Melinda Gates Foundation. The mission of the Gates foundation is to "help all people lead healthy, productive lives." There's nothing wrong with that.

When I see stories like this, especially when alcohol is involved, I like to skip the lurid conclusions and jump to things like who paid for the study and, most of all, what was their methodology?

The study used data from the 2016 Global Burden of Disease report, which captured information on premature death and disability from over 300 diseases by sex and age in 195 countries or territories between 1990 and 2016. This study's researchers used that report to analyze the impact of alcohol on 23 health conditions and alcohol-related risks on people between the ages of 15 and approximately 95 for the year 2016.

This is where it starts to get hinky. The study's very design suggests its objective was to see how many ways they could find to blame alcohol for premature death and disability.

To arrive at their conclusions, the study's authors use something called the Disability-Adjusted Life Year (DALY), a measure of overall disease burden expressed as the number of years lost due to ill-health, disability or early death. It was developed in the 1990s as a way to compare the overall health and life expectancy of different countries. Whether or not that metric is suitable for generating the headline conclusions of this study is debatable. This is how good science often goes awry.

In the DALY, each disability has a different weight, with 1 meaning 'dead' and 0 meaning 'completely healthy.' Urinary incontinence carries a disability weight of 0.139. Migraines? 0.441. Ebola? That’s merely 0.133. The weights are somewhat arbitrary, which may be fine when used as intended, but here they are being applied for a very different purpose.

The study's headline-grabbing claim is that the deaths of 2.8 million people around the world every year can be attributed to alcohol. Is that a lot? The population of the world is 7.5 billion people.

Now let's drill down on that 2.8-million number. We find it includes people who die in car crashes, through unintentional injuries, and violence wherever there is some evidence alcohol was involved. Were those people really killed by alcohol? Or was alcohol merely the weapon of choice in some cases, and an innocent bystander in others?

As universal as alcohol use is, it should surprise no one that many people who die for many reasons have consumed alcohol along the way. People who never leave the house probably won't die in a car crash. Is that a persuasive argument for never leaving the house?

So what? It's just another study. The headlines will be gone tomorrow. Then we'll get another story about coffee or chocolate or, heaven help us, kale.

The problem is, this study's authors also make recommendations, like advocating higher alcohol taxes and other government policies to suppress or, dare I say it? prohibit alcohol sale and consumption.

The busybodies who want to improve us without our consent will always be with us.

There don't seem to be any neutral observers on this subject. I don't claim to be one. Experts in the employ of the alcohol industry suggested some of the pushback above. The more nonsensical analysis is entirely my own.

If you want to dive into this issue even more deeply, check out this commentary from Christopher Snowdon. He explains why it is probably inevitable that policymakers and advocates will finally settle on the "no safe level of drinking" conclusion. Hint: it's political.

When it is hard to get a straight answer from experts, ask yourself "what makes sense?" Most of human experience tells us that most people who use intoxicants (alcohol and others) consider their lives enhanced by that use, not diminished. Are we wrong? I think not, and this study failed to change my mind.

13 comments:

Anonymous said...

Reminds me of a study a few years back with the headline " Diet Soda Causes Weight Gain" even though if logic is applied directly, diet soda with zero calories cannot cause weight gain. But, as always, it's all about the headline.

Unknown said...

Great read Chuck. As a scientist, who performs human studies, I hate it when a group over interprets their data, and I really hate it when the media runs with a story they don't understand, which, when it comes to science, is often. I agree with your interpretation of the paper. They start with the wrong hypothesis. They really start with "Alcohol may reduce quality of life", based on their benefactors. The correct hypothesis would be "Quality of life may be influenced by alcohol consumption". This is not biased, it does not prejudge whether consumption is good or bad, just that it may influence quality.

Richard Turner said...

The "Nannies" are at it again. ...Telling what to do, what not to do, when and when not, and how to decide. Thank goodness we have 'em looking our for us, eh? Otherwise, we may never know for sure how to live our lives.

Chuck Cowdery said...

Michael Williams, my sister is a public health professor and researcher, and I get in trouble if I criticize research the wrong way.

Harry in WashDC said...

Seems to me we entered into a grand experiment once upon a time to use taxes and prohibitions to curtail and extinguish the evils of demon rum (and its relatives). I don't recall how that turned out. Anybody?

I try to adhere to the "All things in moderation" approach. Bourbon, good brisket, and chocolate ice cream notwithstanding.

Sam Komlenic said...

Nicely reported and interpreted, sir!

Brian (AKA The Dean) said...

Michael, well put. I'm a life-long research guy and spent 20+ years of it in TV/media work. I can assure you, most of the writers/producers/etc have no idea what is good research. The only thing that really matters is if they have some reputable source to which to attribute the results/interpretation. The boom in Internet research ("polls" on website forums, for example) makes thing worse, as it get treated with the same respect as a series of double-blind studies published in a respected research journal.

I treat all news concerning research on anything with a skeptical eye. Funny thing though. Whenever I read of a study that shows moderate alcohol consumption has some health benefit, I tend to think "That makes sense!" Call me a hypocrite.

Donald Sink said...

I saw that story when it came out and I thought: "Uh-oh, can't wait to see Chuck's take on this 'study'. Looks like a clear case of confirmation bias to me (the study, not Chuck).

tms said...

I seem to recall, maybe somewhat inaccurately, that the study also found that being alive itself was a primary risk factor for mortality. Wow, how can that be? Duh.
Its all about the sensational headline,just to make people read it (and absorb whatever advertising goes with it- apologies Chuck).

Anonymous said...

I'm not a scientist but I am educated and understand basic scientific method and logic. As you observe,the authors of this study draw conclusions that are in no way supported by the data. Here is another excellent piece of reporting on this "study" https://www.vox.com/science-and-health/2018/8/29/17790118/alcohol-lancet-health-study

Anonymous said...

Figures lie and liars figure.

Jamie W Dunn said...

This reminds me of the great Terence McKenna quote:
If the word, 'life, liberty and the pursuit of happiness' don't include the right to experiment with your own consciousness, then the Declaration of Independence isn't worth the hemp it was written on.

I couldn't agree more.

Anonymous said...

acknowledging your own bias in writing this would go a long way in making this rebuttal even better (it's strong as is 0 but acknowledging your potential bias as your income is somewhat tied to alcohol would be prudent) IMO.