Research into women who consume diet fizzy drinks during pregnancy has revealed they’re more likely to have overweight children by the time they reach their first birthday. The study was conducted by a university in Canada with 3,000 women. But Professor Kevin McConway, Emeritus Professor of Applied Statistics at the OU, says casual conclusions can’t be drawn from this study and more research is needed…
Should pregnant women stop drinking diet fizzy drinks now?
Given the data that they had to work with, this study is statistically competent. But it’s important to be clear that their main conclusion is that ‘further research is warranted’, and certainly not that prospective mothers must stop drinking artificially sweetened drinks now. That’s because of the limitations that are acknowledged clearly in the study report. What the mothers ate and drank was self-reported, so maybe there are inaccuracies. But crucially, this is an observational study, so causal conclusions can’t be drawn.
I’ll spell that out. A fairly small group of these Canadian mothers – about five per cent of them – reported drinking artificially sweetened beverages at least once a day. (That includes diet soft drinks, and also tea or coffee with artificial sweetener added.)
In other words, drinking these drinks once a day or more was a fairly unusual thing for Canadian mothers to do. You might expect that these unusual mothers would be different from the average in some ways, and indeed information in the study report shows that they were. For example, they were more likely to be overweight, to have smoked during pregnancy or to have had diabetes. On average they stopped breastfeeding earlier.
“The study found that this group of mothers were also more likely to have babies who were overweight at one year old. Most of them – nearly 9 out of 10 – did not have overweight babies, but the chance of overweight was about double the chance for mothers who drank very few artificially sweetened drinks or none at all.
They can’t know for sure
Perhaps that increased risk of overweight was indeed caused by the artificially sweetened drinks, or perhaps it was caused by something else that was unusual about the five per cent of mothers. (In the jargon, perhaps the increased risk is due to confounders.)
The researchers made statistical corrections to allow for differences in things they knew about, that may have made this group unusual apart from their sweetened drink consumption, such as maternal smoking and diabetes, a measure of the overall quality of their diet, the duration of breastfeeding, and several more. But they can’t know for sure that they accounted for all these things completely, and there may be other aspects of these women’s unusualness that could not be taken into account at all because they weren’t recorded.
So a study of this nature simply can’t establish what causes the increased risk of overweight one-year-olds. The research report makes all this explicitly clear, and that’s really why the conclusion can’t go further than asking for more research.