Almost 3% of pregnant women take antidepressants. Is that a risk? Do antidepressants cause bad outcomes for kids? Short thread🧵 Firstly, they do not cause mental disorders. Using siblings as controls, there's no significant effect of antidepressant usage during pregnancy.

Meta-analytically, cohort studies suggest - No significant elevation of miscarriage risk - Significantly elevated preterm risk - Likely no elevation of risk of low birth weight But these studies are non-causal. What do causal studies say?

In large-scale population register data, antidepressant usage is, indeed, related to miscarriage risk. But when you compare people exposed during to those exposed before pregnancy, risk evaporates—that means confounding! Comparing the same mothers over time, also no effect.

Antidepressants have also been related to birth defects, and specifically, to cardiac birth defects. This, too, is driven by confounding. Look at people with depression diagnoses? Effect disappears. Use siblings as controls? Effect disappears.

Selection effects dominate when it comes to the harms of antidepressants. Individuals on them are depressed, and depression is not random. But using differentially-exposed siblings, using people as their own controls, and by being theoretically informed, we can avoid confusion.

It's important to think causally about this issue: Antidepressants prevent suicides and keep people's lives in check. False alarms from fake safety signals have led to drug withdrawals before, and we have to be careful or they might lead to them for antidepressants, too.