Seriously

[eve11]

You'd better not have a C-section because it will cause your baby to have weight problems later in life!

Are you KIDDING me? Apparently a doctor wrote this article. And tries to convince us that "a baby needs to be exposed to bacteria that they might not get if delivered by C-section" as the strongest reason why this association is causal. Scientific News Reporting, can you please stop this nonsensical bullshit and, you know, report on the merits and implications of the actual study instead of this drivel that panders to the lowest common denominator? What the hell kind of editor reads this story and says, "Great! Pass it along!"

I deplore the state of science education, science reporting and general analytical thought in our region if this is its best representative.

Comments

[flowsoffire.livejournal.com]

Hahaha, oh god. *shakes head* What kind of idiocy is that? If there's a problem with too many women getting c-sections when it isn't necessary in every case, how about giving some actual information instead of "oh look, you really shouldn't, your baby might turn out fat?"

(I was born by C-section and I say eff the statistics, btw ;))

[eve11]

Yeah, it's not the statistics that are bad; it's that the conclusions that the writers reach don't seem to be supported by the statistics in the study. Or at least that they made no attempt to explain why or how they are. So frustrating.

[flowsoffire.livejournal.com]

Yeah -_- It just sounds really rubbish…

[doctorpancakes.livejournal.com]

Way to mom-shame women who NEED to have c-sections in order to give birth safely, too. Ugh. I mean, it's shitty research to begin with, and then I guess I bristle at anything that combines shitty research with making mothers feel unnecessarily guilty about ruining their kids.

My brother was born by c-section, and apart from a bit of a tummy (which we can plainly blame on, like, fried chicken an potato wedges, ferchrissakes) he's an averagely thin dude in his mid-thirties. So yeah.

[a-phoenixdragon.livejournal.com]

Ummm...seriously?! My youngest was delivered via C-section - and he is anything but fat or too thin. Actually, both of my boyos are healthy weight and the doctor's office goes crazy over it because everything nowadays is scaled to obesity. So this article? Total horseshit...

[skurtchasor.livejournal.com]

I've you've not done so, you should read some of the other "articles" by the same author, who's touted as a leader in medical journalism. Just make sure that your rage level is low before you start.

[aoife-hime.livejournal.com]

Ugh. UGH UGH UGH.

Not all physicians are created equal. They really, truly aren't. I feel pretty lucky to work with excellent physicians who are intelligent and do good research, but there are so many out there who are the exact opposite.

[auntiemeesh.livejournal.com]

Other factors may also play a role in C-section babies gaining weight later on, including maternal weight, age and lifestyle.

I love how all of these other factors that may weigh in to both the need for a C-section and the later life weight/health of the children is relegated to a single sentence at the very end of the article, written as dismissively as possible.

[clocketpatch]

Bwuhhh? This "fact" apparently didn't stick to me on the C-train out.

[giallarhorn.livejournal.com]

I thought I recognized a username in the comments :P

Part of the problem is how media and science interact in a sensationalized manner, but yeah. This is pretty bad for publishing- I suspect they picked something they felt that could make a splashy headline?

Also, here's the full paper. You might be able to make more sense of the methodology than I could. As much I could tell, they looked at other studies and determined a correlation?

[eve11]

Wow. That paper ... they had sample sizes of like 40,000 and then found "statistically significant" effects of an odds ratio of 1.16 or 1.25!

And oh yes, let's also include the fact that they are using BMI as a measure of obesity and have made their entire focus on the difference between means in the groups regardless of the population structure. They outright say that the difference in means in the populations is 0.44 BMI higher on average for ceasarean vs. vaginal delivery. Which of course because of their massive sample size, is going to be "statistically significant".

Also, right at the top of the study it says "There is a strong association between CS and increased offspring BMI, overweight and obesity in adulthood. Given the rising CS rate worldwide there is a need to determine whether this is causal, or reflective of confounding influences."

At any rate, what they did was mine a bunch of medical articles to try and find a sampling frame of data suitable for their analysis, following up with a bunch of authors of different papers to get the authors to report them the mean BMI for the people in their studies broken down by their variables of method-of-birth. They figured the "effect" would be different across different studies but they seem to have made highly minimal attempts to control for other factors. They mention that they broke down by elective vs. medical (maybe?)... I don't think that is going to be enough. One thing they don't have for example, is any notion of the BMI of the parents that made it into their studies. I'd love to see if their 1.16 odds ratio or whatever is commensurate with the BMI of parents.

What complete bunk. Jeez.

[keypike.livejournal.com]

C-sections are often pushed onto women whose babies are measuring big in utero, so that probably plays a part, too. Big baby, mother gets c-section, "OH LOOK, THE BABY IS BIG, must be because of the c-section!!!!" Ridic.

I really wish that every article that references a scientific study would be required to include an explanation about correlation and causation.

[eve11]

Yeah, and oddly enough, if I read the paper correctly, when they controlled for eg emergency vs. elected, the effect went away.

I can't believe that the authors basically looked at the only study in their meta-analysis where mother's BMI was recorded, saw it had a sample size of 6000 and that oh hey, the differences were no longer statistically significant in that study, and then had the gall to suggest it was because that study "had lower power" while plowing on through to the conclusions they wanted to see.

[giallarhorn.livejournal.com]

Is the odds ratio to the sample size ridiculous because of how large the sample size is, or is have to do with how large their sample is?

Yeah, I'm no expert in physiology but BMI has been pretty much debunked as a useful measuring tool beyond providing a very loose prediction in growth, but the difference in their means doesn't seem to be that remarkable?

I suspect that factoring in the BMI of the parents would probably throw off their correlation by a bit, since while obesity isn't genetic, certain environments do lend themselves to it.

So basically, not nearly as bad as the vaccines cause autism paper but not nearly so rigorous in the testing to actually be able to state ceasarean vs. vaginal has an impact on obesity?

[eve11]

Well, it's more like, do you think that an odds ratio of 1.25 really means something, when an odds ratio of 1 indicates no effect. They say, "25% more likely to have weight problems" because it sounds more dramatic than "0.25 times more likely to have weight problems". The larger the sample size, the smaller sized effects you will be able to detect as "statistically significant" even if they don't really mean anything practical. But the bigger problem is definitely that they didn't really control for any kind of underlying factors, so even that effect size that they are reporting is likely also a product of confounding factors in the populations.

They found and measured a correlation with statistical significance, but plausibly, it is much more likely due to non-causal factors. They would have to do a different kind of study to try and isolate a causal effect. But most concerning for their hypothesis is that the one study that did control for obvious confounding factors, found no residual effect. Instead of them saying, "Well, this is evidence against our claim", they said that since the study had smaller sample size, it had lower power than their study and so wouldn't be able to detect as subtle of an effect as found in their study. But who the heck knows what the magnitude of a real causal effect (were there to be one) would be. The study that controlled for confounders can at least provide a better guide for an upper bound.

[lordshiva.livejournal.com]

Also if people could just stop terrifying expectant mothers who actually have no options but to deliver via C-section that would great, You say it better, because all I can ever think is fear-mongering assholes.