There was a lot of hedging in these stories, especially considering how the media often treats these studies like fact, including phrases like “the first study to link the two” and “doesn’t prove” and “appears to create concern”. So I found the study. And read it.
First, a quick review of correlation vs. causation. Correlation means that we can tell that two things happen at the same time. Causation means that we can tell that one thing causes the other.
So if every August there are more murders and more ice cream eaten we cannot say that eating ice cream causes murders. If there is a rash of murders we cannot say that there is an ice cream epidemic, and we cannot conclude that taking ice cream off the shelves will cut the murder rate. They could both be caused by a third factor (maybe heat makes people cranky and they either eat ice cream or commit murder) or they could be completely unrelated and the correlation could be a coincidence. (More thorough explanation is here.)
The problems begin at the beginning – the study is called “Maternal Metabolic Conditions and Risk for Autism and Other Neurodevelopmental Disorders” It’s not called “The Effect of Obesity on Austism.”
What they actually studied was whether metabolic conditions during pregnancy are associated with autism spectrum disorder, developmental delays, or impairments in specific domains of development in the offspring.
The “metabolic conditions” including diabetes, high blood pressure, and obesity. There is an issue right there since diabetes and blood pressures are conditions with very specific metabolic indicators, but obesity is simply a ratio of weight and height – obese people have extremely varied metabolic health. There is a danger anytime we make body size a diagnosis.
The study included 1004 children and mothers (517 with Autism spectrum disorder, 172 with developmental delays) and 315 controls). That is a small sample size and it seems to me that having a few kids with developmental delays besides autism further muddies the waters. All of the kids are from California which means that if California has state laws about neo-natal care that affect autism then all of the kids in the study would have been exposed to them. Much of the data is self-reported (which is a notoriously poor way to collect data.) For example, women were asked “At any time before you became pregnant with [index child], were you ever told by a doctor that you had [diabetes, high blood pressure]?” The accuracy of memory in these situations is not great.
Their conclusion was “Maternal MCs may be broadly associated with neurodevelopmental
problems in children. With obesity rising steadily, these results appear to raise serious public health concerns.”
The first part of their conclusion is a little problematic, and depends on the meaning of the word “may”. If they mean it might be broadly associated then that’s a fair conclusion. If they mean it can be broadly associated then that’s a stretch considering the issues above. The second half is just weird. They hedge again by saying “appear to raise” instead of just “raise” which is the right thing since they are ignoring the actual metabolic conditions that they studied and making this conclusion about obesity.
One doctor has already pointed out that women with metabolic conditions tend to be given more ultrasounds and there is a correlation between ultrasounds and autism. There are a lot of other variables here. We know that doctors treat people of size with less respect and give them less time so perhaps that accounts for the difference.
Because they are only looking at correlation, they don’t have to propose an explanation for why a height/weight ratio (which can have extreme variances in amount of adipose tissue vs. muscle, in height and weight, and in metabolic health markers) might be responsible for a fetal health outcome. And that’s probably for the best because that would be extremely difficult to do. You would have to make some extremely broad assumptions about all obese women in terms of habits, body composition etc. For example, if you looked strictly at body fat percentage, some women in the control group would likely have the same, or higher, body fat percentage as the obese moms. If you looked at diet, some women in the control group would likely have the same diet as those in the obese group. Same with amount of physical activity etc. Body size is not indicative of specific behaviors or specific health markers.
I kind of makes me wonder if the researchers threw obesity into the study, and into the conclusions, to take advantage of a media that is ravenous for the opportunity to engage in fat shaming and perpetuate the myth of an “obesity crisis”.
I think that the way the media handled this is a crime. They don’t have any proof that being fat leads to autism there is no method of weight loss that works, and fat moms of autistic kids have enough to worry about with the way that society handles autism without every news outlet blaming them for their kids autism when they really have no clue what’s true.
Like the blog? Here’s more of my stuff:
Interviews with Amazing Activists!! Help Activists tell our movement’s history in their own words. Support In Our Own Words: A Fat Activist History Project!
Become a member: For just ten bucks a month you can keep this blog ad-free, support the activism work I do, and get deals from cool businesses Click here for details
The Book: Fat: The Owner’s Manual The E-Book is Name Your Own Price! Click here for details
Dance Classes: Buy the Dance Class DVDs or download individual classes – Every Body Dance Now! Click here for details