Why you probably don’t need to worry about recent headlines linking pregnancy stress to personality disorders in children
There are endless opinions about the experience of pregnancy, from delightful to horrendous, and everything in between. But one thing many people do agree on is that, at some point in the process, it can get a bit stressful. It makes sense — you are making a human life and bringing it into the world, and if nothing else is true about humanity it’s that we are all very stress-inducing animals.
And for anyone who is currently pregnant or is planning on being so in the future, this week must’ve been an extra-stressful experience. Headlines from around the world have been screaming that not only is stress during pregnancy a pretty uncomfortable experience at the best of times, it’s also causing your child to have mental health problems later in life.
According to the media, any stress during pregnancy makes your child “three times” more likely to develop personality disorders — a broad term that encompasses depression, anxiety disorders, and some types of psychosis — rising to nearly “ten times” more likely if you have a lot of stress.
Fortunately for pregnant people (and their kids), the reality is much less terrifying.
You don’t have to stress too much about stress. It’s probably not making your children mentally ill.
The study that has caused the uproar is a fairly basic piece of epidemiological research known as a retrospective cohort study. The scientists took a group of mothers who had been surveyed about their mental health during their pregnancies in the 1970s, and followed up the children after 40 years. They grouped the pregnancies into three groups: no stress, some stress, lots of stress, and compared these groups in terms of the children’s’ mental health, specifically looking at personality disorders.
The results were fairly alarming — children who were born to women who reported being stressed during their pregnancy in the 70s were three times as likely to have a diagnosed personality disorder four decades later. For some conditions this figure rose even higher — for example, stress during pregnancy was associated with thirteen times the risk of anxiety disorders!
Cue the scary headlines: “Don’t get stressed during pregnancy or your child is sure to suffer”
Except, that’s not right at all. The science wasn’t nearly as scary as the media reported.
The first thing to note was that this was observational research. Now, I talk a lot about observational research, and usually my main point is that it can be hard to infer causation from what are very complex studies. Basically, even after for controlling for everything you can think about, there’s still a good chance that society is behind the effect seen in the research.
In this case, I don’t think that’s even necessary.
You see, in epidemiological research you usually control for dozens of variables, because there are so many things that can effect people’s health. In this case, income, ethnicity, marriage, overall health, and hundreds of other things may influence whether someone has a personality disorder.
But this study didn’t really do any of that at all. The researchers controlled for a very small range of things, which means that there are innumerable other factors that may have caused the result. The most obvious example is that we know that income influences who gets mental health problems, especially personality disorders, but this study made no attempt at all to control for income or socio-economics at all. It’s very likely that a better explanation of the results is that rich women have less stressful pregnancies and healthier children because of access to medical services, but we simply don’t know if that’s the case because the study just didn’t control for this.
Why didn’t they use these sorts of controls? Well, it might have something to do with the sample size. While the total sample was reasonably big — 3,626 mother/child pairs — the number of personality disorders was very small at only 40 cases. That makes statistical modelling extremely difficult, especially the statistical model used in this research known as a logistic regression model.
Without going too deeply into the stats here, logistic regression models have known issues with this type of analysis. When you use a logistic model with only 40 outcomes (in this case personality disorders) and up to 5 covariates, the answer you come up with is more likely to be an artifact of statistical chance than the truth. This gets even worse when you start including lots of factors like socio-economics and ethnic background, which means that it’s possible the scientists didn’t control for these simply because they couldn’t.
Another issue with the reporting of the study was the very small absolute risk increase. Everyone reported the odds ratios from the study’s logistic model — the “triple risk” I talked about earlier — but the absolute risk difference was only about 1% between mothers who were and were not stressed. To put it another way, if 100 mothers had a stressful pregnancy, you’d expect to see 1 extra case of personality disorder in their children compared to 100 mothers who were stress-free.
That’s not nearly as scary as the headlines made the study seem, is it?
Ultimately, there’s not much you can take away from this study. For doctors, who see pregnant women all the time, it might be important to consider whether someone is stressed at their check-up — in fact, that’s what the authors say their study should be used for — but realistically most doctors are probably already checking that anyway.
For anyone who is/will be/has been pregnant, the results probably don’t mean very much. There might be a small risk increase associated with stress during pregnancy, but given the statistical issues it’s actually quite hard to know if that’s the case. Even if it is, the study didn’t demonstrate causation, so it’s more than likely that actually something like how much money mothers have is influencing the mental health of their children, rather than the stress itself.
Stress isn’t ideal during pregnancy, but firstly, it’s often unavoidable, and secondly, we already knew that anyway. No one wants to be stressed while pregnant, but the reality is that sometimes it happens. Terrifying pregnant people about how the stress is harming their baby isn’t just unhelpful, it’s also a bit absurd.
If you’re worried about your health or the health of your baby, go and see a doctor — if you’re pregnant, probably an obstetrician.
Just don’t stress too much about getting stressed.
Chances are it isn’t hurting your baby after all.
If you enjoyed, follow me on Peerdiy, Twitter or Facebook!
You can now listen to Gid on the Sensationalist Science podcast for your weekly dose of scientific shenanigans and media muddling: