Like drinking soda? Think again...

[Sjofn]

I went through a period where I was thirsty all the fucking time. I got tested for the diabeetus, but that wasn't it (I predict that is in my future, though). Apparently some people, when stressed, drink a LOT MORE than they would normally. Just another way my brain is busted, yay!

I am trying to cut back on soda but unfortunately my backup drinks are juice or 1% milk (I fucking LOATHE coffee and while I like tea, it's not something I drink to quench a thirst).

[tazelbain]

Add me to the list of people who dropped a good deal of weight cutting out sugar.  Water tastes better now.

[Teleku]

Beer is a perfectly healthy substitute for soda, right?  Because the same time I cut out soda from my life is around the same time I started drinking beer heavily...

[Johny Cee]

edit2:  I've also found when trying to switch to non-carbonated stuff that I drink it much faster and my daily intake goes way up so instead of 120oz of soda I would drink 200oz of tea/water/juice/flavored drink.

Um?  That may be a question for your doctor at your next checkup....  excessive water/fluid drinking could indicate a mental or medical illness.

I'd say mental, it is just a really bad habit that I need to have a drink at hand when coding.  Sort of a mini reward system, code a block then sit back and review what I just did while taking a few sips.

Yah, I'm not trying to be internet diagnostician guy....  just that you seem to drink much more fluids than is normal which is usually a hormone thing or a mental thing.  The only reason I know about the psychological side of it was a part-time job in college working with DD/mentally hadicapped folks where we actively had to ration how much they could drink and it was a side effect of some other issues.  Forgot the term for it.

Just saying it could be something to mention the next time you go in for a physical.  Could just be an indicator that you're more stressed by your work environment than you realize.


I pretty much cut soda out of my diet a couple years ago, especially after we got a Britta pitcher at work.  Diet sodas, besides not tasting very good, are still full of the other nasty shit that is standard in carbonized drinks.  Even diet soda does a number on your teeth.

Coffee with lots of milk (no sugar) and water are about the extent of my drink choices now.  Sometimes I have a flavoured iced tea of some sort as a treat, though I usually drink one bottle over the course of a week.  Unsweetened tea in the winter or when I have a cold is nice.

Like taz said, cutting out the soda is a surprisingly easy way to lose some weight.  Especially if you're desk bound.

[ghost]

While I don't believe that drinking lots of soda is good for you, the pancreatic cancer link is possibly a bit overblown.  Pancreatic cancer is very rare and even with their numbers it appears that you are talking about a difference of 45 people out of the 45,000 or so that appeared to be in the study.  I always raise an eyebrow when they say, "you have an X% increased chance of gettting X disease because of doing X."  Increased percentage chance of having something happen is only particularly relevant if the percent increase is very significant or if the numbers were significant in the baseline group.

[Nebu]

While I don't believe that drinking lots of soda is good for you, the pancreatic cancer link is possibly a bit overblown.  Pancreatic cancer is very rare and even with their numbers it appears that you are talking about a difference of 45 people out of the 45,000 or so that appeared to be in the study.  I always raise an eyebrow when they say, "you have an X% increased chance of gettting X disease because of doing X."  Increased percentage chance of having something happen is only particularly relevant if the percent increase is very significant or if the numbers were significant in the baseline group.

I agree completely.  Buying two powerball tickets doubles your chances of winning. 

I wasn't attempting to make a political statement here.  I just thought it was an informative study as I ran across it in my weekly journal review. 

[ghost]

I agree completely.  Buying two powerball tickets doubles your chances of winning.  

Actually, I don't think it does.  I have read somewhere that you have to buy an insane number of tickets to actually increase your chances.  Statistics are strange things.

I wasn't attempting to make a political statement here.  I just thought it was an informative study as I ran across it in my weekly journal review.  

I know.  I just know medical research and thought it would be fun to discuss the intricacies of the article.  I'm out of the political arguments.

Edit:  FYI, I am highly critical of the soda industry, HFCS and food additives in general.  I love it when people weigh 400 pounds and wonder why they're obese-  "what do you mean I get 2800 calories per day from this case of Mountain Dew that I drink?"

[Nebu]

Actually, I don't think it does.  I have read somewhere that you have to buy an insane number of tickets to actually increase your chances.  Statistics are strange things.

The odds of being a jackpot winner for like Mega Millions are 1 in 175,711,536.  Wouldn't buying two tickets make the odds 2 in 175,711,536 or double?   

I don't claim to be a math wiz. 

[Cyrrex]

Yes, those are doubled odds.  Maybe he's saying something along the lines of "even though your statistical odds have doubled, your actual chances of winning are so unlikely that it's not even worth noting the increase."  Or something.  Who cares.

As far as the on topic part of the thread goes:  Soft drinks, whether diet or otherwise, are basically just cans full of poison.  We all know they're bad for us.  If it has an ingredient you wouldn't use in your every day language, chances are it isn't good for you.  I'm currently drinking a Powerade (my wife gets them for free, otherwise I wouldn't), and the list of ingredients on the label is fucking ridiculous. 

And diabetic concerns aside, I'll never believe that all sugars are created equal.  I know that a glass of pure apple juice may have as much (or more) in terms of sugar calories as a can of Coke, but you cannot compare these things.  The one is good for you in almost every way.  The other will fucking kill you dead.

[Sky]

Beer is a perfectly healthy substitute for soda, right?  Because the same time I cut out soda from my life is around the same time I started drinking beer heavily...

Light beer is about 100 calories per can.

http://www.beer100.com/beercalories.htm

[ghost]

Actually, I don't think it does.  I have read somewhere that you have to buy an insane number of tickets to actually increase your chances.  Statistics are strange things.

The odds of being a jackpot winner for like Mega Millions are 1 in 175,711,536.  Wouldn't buying two tickets make the odds 2 in 175,711,536 or double?   

I don't claim to be a math wiz. 

Odds are not necessarily the same as your probability of winning.  Odds do not equal probability.  Regardless, the point remains valid in that the closer you get to infinitely small the less that doubling, tripling or even quadrupling that number will matter.  Drug companies, food companies, cigarette manufacturers, etc. have all exploited this discrepancy in the past and it appears that the author of the article is playing to this effect.  140 cases out of 60,000 people is very small (.23%).  Additionally, one would have to question whether or not the environmental effects of just being in Singapore could be a reason for the difference-  I don't see what control they were comparing their group against to make the "87% increase" proclamation. 

[01101010]

Confounding variables be damned! Full speed ahead! What I would like to know is who is at the end of this funding money line...

[ghost]

Confounding variables be damned! Full speed ahead! What I would like to know is who is at the end of this funding money line...

Big Corn

[Nebu]

I don't see what control they were comparing their group against to make the "87% increase" proclamation. 

If you're that interested, you can grab the article and contact the PI by email.  If you have a legitimate reason to ask, I'm sure that they would take the time to respond to your question.  I field questions about my research all the time, particularly as it relates to something that I've published or presented. 

I don't want to debate semantics... that wasn't the reason that I posted the article summary. 

[ghost]

If you're that interested, you can grab the article and contact the PI by email.  If you have a legitimate reason to ask, I'm sure that they would take the time to respond to your question.  I field questions about my research all the time, particularly as it relates to something that I've published or presented. 

I don't want to debate semantics... that wasn't the reason that I posted the article summary. 

You think having an appropriate control group is "semantics"?

Again, I know why you posted it-  soft drinks are bad for your health.  There's nothing wrong with some good natured nitpicking of the research.

[Nebu]

You think having an appropriate control group is "semantics"?

Nope. I don't. 

[rattran]

No, but you're casting aspersions on the study without finding out specifics. As Nebu said, "If you have a legitimate reason to ask, I'm sure that they would take the time to respond to your question" otherwise you're just shitting up the thread. Don't be a Broughden.

[ghost]

Why is pointing out that an 87% increase in pancreatic cancer might not really mean anything "shitting up the thread"?  Every educated person on here knows that this is true.  Why not just fucking say "hmm" and move on, then? 

[Nebu]

Go ahead and Den this.  I'm sorry I even posted it. 

[ghost]

Nah, just den my part out of it.  I thought Nebu might have enjoyed discussing some of the points of the article.  I guess I was wrong-  sorry.

[ezrast]

Odds are not necessarily the same as your probability of winning.  Odds do not equal probability.

Hate to further the shitting up of the thread, but, what?

[Lantyssa]

If it has an ingredient you wouldn't use in your every day language, chances are it isn't good for you.

Uh, my first degree is in chemistry...

[Righ]

I thought Nebu might have enjoyed discussing some of the points of the article.

I can't say whether he would or he wouldn't but that wasn't really one of the points of the article. It's reasonably clear from the précis that the scientists conducting the study were well aware of the narrow focus of the study group - it was the commentary that (daringly) suggested that it could be applied to other groups considered similar in the author's opinion. Typically speaking, the commentary will be written by an independent scientist who has reviewed and supports the paper. It doesn't mean that the researchers made claims without appropriate controls. It's some guy stating "I like this paper, more people should read it". It's the science equivalent of a hype man.

[Raging Turtle]

ITT people show they have no idea how experiments are run. 

Seriously, if this was a peer-reviewed, published study (as I'm assuming it is from the mainly from the magazine name, correct me if I'm wrong), people far smarter than the ones babbling here have conidered every single point that's been brought up so far. 

The fact that is was only 47 people (or whatever) that got cancer?  There's no reason that wouldn't be enough to be statistically significant (although I'll admit the study says it limits it's statistical power).   
87% more likely to get pancreative cancer!  Soda's going to kill you, me, and everyone we love!  Well, no, if there's 42,000 cases a year in the US (from wiki), and there's 300 million people, an 87% increase is still tiny. 
The fact that they did in in Singapore?  Well, it is pretty much a first world country, but they're brown/non-white, so it can't possibly be relevant for people in the U.S. 
An "appropriate" control group?  Yes, you can actually take into account differences in age and lifestyle and still come up with meaningly data.  Statistics is very much a hard science - sure, you can set up an experiment to show any result you want, but that's where peer review tends laugh you out on your ass. 

And this was especially fun: 

Confounding variables be damned! Full speed ahead! What I would like to know is who is at the end of this funding money line...

Big Corn

It says in the OP that the study was funded by the National Cancer Instititute.  But keep on believing that it's all a conspiracy!

Now, this is just one study, and despite what CNN or FOX says, just one study on any given subject is pretty much meaningless.  But what if there were more studies?

From the OP:

However, this study adds to the evidence that soft drink consumption plays a role in the development of pancreatic cancer, they conclude, and that "clinical studies examining biomarkers for glycemia and insulinemia and taking a mechanistic approach to the question of soft drink consumption and pancreatic cancer are warranted."

And that's why some people are shitting up the thread.  Because clearly, those damn 'educated elites' are far too stupid to think of these blindingly obvious things.

/science journalism is a joke, I know
//I like root beer
///haven't had a soda in months

[Sheepherder]

Do they make a distinction between sugar and HFCS? Or is that irrelevant in this case? I would also like to know if they made a distinction between fruit juice (100% not from concentrate) and or "Juice drinks" (Typically from concentrate, 10% juice and a ton of HFCS or sugar water)

Note for all the people drinking "Fresh, not concentrate" juice: it's not fresh in any sense of the word that you would recognize.

...even “not from concentrate” orange juice is heated, stripped of flavor, stored for up to a year, and then reflavored before it is packaged and sold.

[01101010]

I shal refrain from spouting off any quips right now, but I will say that I have been published in a scientific journal and the whole process does in fact hinge on getting enough of "your" people to agree with you to push your journal article through. Granted, "your" people are others in the same field and knowledgeable on the same topic, but once they get your article and they already "like" you, its not uncommon to have a mild review and get published. My article was my thesis and I hardly knew anyone in the field being a Master's student so it was a bitch of a process and the article was submitted to several different journals before being R & R'd then accepted.

And don't bring up the blind academic review bullshit either. They way people write an article is like signing your name. My adviser at LSU sat on several journal review committees in his long career and he could pick out an author even though it was never revealed - just by the style it was written, statistics used, and topic at hand. If he liked you, he would suggest you clean it up a bit here and there. If not, he would outright reject it, which lead to the author having to submit it elsewhere.

That whole process was an eye opener and was a definite motivator for me to move away from peer reviewed research and my PhD as a whole. I don't give a shit about conspiracies, I have seen the process first hand and was in the grinder for almost a year before my article came out.

[Abagadro]

Talking about a "control group" in this kind of study is nonsensical.  This isn't that type of study where you are applying outside stimuli and seeing how it works so you need another group who doesn't receive it to measure against.  The fact that they are all from one area/ethnicity is actually helpful because it is one less independent variable you have to control for in trying to isolate correlation/causation for the dependent variable.  There could always be intervening unaccounted for variables that would skew things, but the fact they are from the same population is somewhat the opposite of such a phenomenon.
 
As for the small number of cases, you can always run into a "small N" problem, but here it looks like there was statistical significance to the increase and the journal would have likely found any issues in the actual math in the review process.

It is an important piece of research and I hope there is more follow up.  This is what killed my dad and causation for pancreatic cancer is a bit of a mystery (my suspicion was him being downwind of the atomic testing next door in Nevada, but that is just my speculation).

[ghost]

Talking about a "control group" in this kind of study is nonsensical. 

Actually, it really isn't at all.  Controlled, randomized trials are the very epitome of good research.  If you think otherwise you are a blithering idiot that knows nothing of research.  They absolutely have to have a "control" group to compare to get their baseline figures.  I am assuming that they are comparing the "haves" to the "have nots" within the study.  This is a problem with a longitudinal observational study.

This isn't that type of study where you are applying outside stimuli and seeing how it works so you need another group who doesn't receive it to measure against. 

True.  It is an observational study, which are traditionally shit and looked upon by medicine as the worst form of research. 

The fact that they are all from one area/ethnicity is actually helpful because it is one less independent variable you have to control for in trying to isolate correlation/causation for the dependent variable.  There could always be intervening unaccounted for variables that would skew things, but the fact they are from the same population is somewhat the opposite of such a phenomenon.

Two points on this:  first, a study on Chinese-Singaporeans that compares to other Chinese-Singaporeans has little comparison value to the American or European populations, particularly if they don't have a group that is followed which is outside for comparison.  They might as well have done the study on fucking dogs, for all that it matters to the rest of the world.  Hence why these types studies are shit.
 

As for the small number of cases, you can always run into a "small N" problem, but here it looks like there was statistical significance to the increase and the journal would have likely found any issues in the actual math in the review process.

Or they could run into the exact opposite problem in which they keep adding subjects until they find something to write about-  they call that mining data and it leads to suspect results. 

It is an important piece of research and I hope there is more follow up.  This is what killed my dad and causation for pancreatic cancer is a bit of a mystery (my suspicion was him being downwind of the atomic testing next door in Nevada, but that is just my speculation).

Sorry to hear about your pops.  Pancreatic cancer is a horrible, horrible thing.  You are probably right about the nuclear testing, but how many sodas did he drink per day..........?  With this article you might have a lawsuit on your hands. 

[Goreschach]

It's funny how in these threads you can always immediately recognise the ex-fatties. They're always the ones who go on this idiotic zealous crusade against sugar, or fat, or preservatives, or plastics, or whatever the latest health fad bullshit is, like they're trying to 'make up' for all the years of being a big fatass.



Being born is statistically 100% correlated with future incidence of death. Deal with that one.

[ezrast]

It's funny how in these threads you can always immediately recognise the ex-fatties. They're always the ones who go on this idiotic zealous crusade against sugar, or fat, or preservatives, or plastics, or whatever the latest health fad bullshit is, like they're trying to 'make up' for all the years of being a big fatass.



Being born is statistically 100% correlated with future incidence of death. Deal with that one.

Heh, I almost posted something about how I was going to outlive all you fuckers because I basically never drink soda. But I didn't! I've also never been fat. Coincidence???

Ghost, have you, uh, had much in the way of formal education regarding statistics?

[Abagadro]

That is my question too.  I don't agree with your characterization of this type of research at all.

[Raging Turtle]

Ghost, have you, uh, had much in the way of formal education regarding statistics?

Rather than SirBruce his Sirbrucing, I'm going to say that, as someone who *has* had a fair amount of education in these areas (but not professional experience, I'll admit) that no, he hasn't. 

[bhodi]

Go ahead and Den this.  I'm sorry I even posted it. 

Don't be. I and others thought it was interesting :)

Besides, this is a good way to engage in a little bit of societal conditioning of our own. Be stupid, be made fun of publicly!

[Lantyssa]

Actually, it really isn't at all.  Controlled, randomized trials are the very epitome of good research.  If you think otherwise you are a blithering idiot that knows nothing of research.  They absolutely have to have a "control" group to compare to get their baseline figures.  I am assuming that they are comparing the "haves" to the "have nots" within the study.  This is a problem with a longitudinal observational study.

There are plenty of studies which take a look at a large population (60k in this case) who have been tracked for a large number of conditions and living styles over a number of years.  Then they look for various correlations between disease and lifestyle.  Not cause, but correlation.  Which is why they usually say in their conclusion that this needs to be studied in further detail to determine why these results happened.

The control is the group itself, because you split out people who did X from people who did Y and see what the differences are.  60,000 people is an excellent sample size for a long-term study.  Many disease researchers would love to have that many people.  You've got 10k who drank soda, 10k who drank juice, and 40k who drank neither.  Confirm that you have equal proportions of other factors spread amongst these groups and that those conditions show no significant statistical difference and thus you have Group A, Group B, and Control.

And it doesn't mean it is definitive, it means they have found something which can be looked at in other broad, longitudinal studies to see if similar results are found.

It's perfectly valid, and arguing it's not shows you're either ignorant or an idiot.  Fortunately one of those can be fixed if you are so inclined.

[Abagadro]

The control is the group itself, because you split out people who did X from people who did Y and see what the differences are

Indeed. In this kind of study/analysis, every group is, in effect, a "control group." If you expanded it to other racial/nationalities you wouldn't have a "control group," you would just have another independent variable that could be plugged into the model and controlled (via regression most likely, I didn't get that into this particular study) to determine its effect. Crabbing about no "control group" shows a fundamental misunderstanding of this type of statistical analysis.  Expanding to include that population might be a good idea as it would eliminate a possible intervening variable that isn't accounted for in the model, but it in no way would constitute a "control group."