Rebecca Goldin Ph.D
The New York Times (and according to Google News, some 400 other news stories) takes the cake – but no coffee, please – for offering its readers some poorly done science with a little guilt on the side. The newspaper touted a study purportedly showing a link between caffeine and pregnancy, but overlooked the serious flaws that the experts cited in the piece point to. And these flaws are serious enough to call into question even a mild correlation between an increased caffeine intake and an increased miscarriage rate.
The study asked approximately 1,000 women about their caffeine intake and their miscarriage rate. They authors then noted that those who consumed more caffeine (more than about one small cup a day) were also more likely to have a miscarriage.
But there were huge problems with the study’s design. First, of the 172 miscarriages in the study, 102 occurred before the interview – a huge opening for “recall bias.” In other words, women who have miscarried may be more inclined to remember or exaggerate their caffeine intake, perhaps out of guilt that they “lost” the pregnancy.
Second, it may well be that those women with healthier pregnancies were less inclined to drink coffee since healthier pregnancies are more likely to have associated nausea. Thus causality may be going the other way – those with healthier pregnancies were less likely to consume coffee.
These problems were noted by experts commenting on this study – and they openly dismissed its conclusion. Yet the New York Times went with the study in framing its headline, rather than suggesting that readers discredit it.
The lead author of the original study commented that “Stopping caffeine really doesn’t have any downside.” Well, maybe for him it doesn’t – ask a pregnant woman, and she might have a different point of view. With a substance so easy to use in a controlled environment, it seems absurd to accept experiments with obvious potential and significant bias as the basis for a conclusion. In fact, we reported on a study last year that showed that in a double-blind experiment, pregnant women consuming similar amounts of caffeine did not have an increased miscarriage rate in the second half of the pregnancy.