I love most of your writing, but unfortunately I don't know what to do with the extent of statistical fallacy that you're conveying in this message.

What's the point of conveying stats in your posts if you actually believe that "doesn't mean anything to the woman who dies" is a relevant argument?

Airplanes are safer than cars per mile, but that doesn't mean anything to the person who dies in an airplane crash.

Seatbelts are safer than none, but that doesn't mean anything to the person trapped by their seatbelt in a sinking vehicle.

Getting a vaccine is safer than getting Covid but that doesn't mean anything to the person who dies to severe anaphylaxis.

It's just rubbish.

Regarding comparing getting pregnant and not, you don't end up with two children by not getting pregnant. So I'm not sure what you're getting at there. It's all apples.

Confounding variables? Possibly--they controlled for a fair few, though there are always more you could try. (Having more emotional support wouldn't be a confounding variable, though...that would be part of the mechanism if it were true--you don't randomly get extra emotional support because nothing! You get it (if you get it) specifically because you have children!) Still, it's not like this is the first study to get a broadly similar result (citations 2-4 and 9-11 all report roughly the same thing). And the effect size isn't tiny; it's on the order of 20% lower death rates at a time when a substantial fraction of the cohort is dying (~10%). Compared to maternal mortality in Japan (0.005%), this effect is huge.

And yes, as we have both said multiple times this paper only covers death (including, to a limited extent, cause of death--with children, cancer down, cardiovascular events up), not other aspects of health.

I am very sorry that you personally experienced a nearly-fatal event. And I do not at all support withholding information about potential negative health consequences of childbirth (including a realistic assessment of the chance of death). And I don't think we do nearly enough follow-up post-birth to even be able to tell people what to expect. But it is intellectually dishonest to either misrepresent the rates of risk, or to fail to include the positives along with the negatives. People deserve the whole story, as best as we know it, along with help developing the cognitive tools needed to understand the story, if necessary.

Additionally, there is one pretty substantial side-effect of having children that is kind of getting lost here: after having children, you then have...children. If you want them, that's a pretty big plus to weigh against the negatives. If you don't, that's a pretty big minus to weigh against any positives. The idea that a 0.02% chance of death is a huge deal compared to that is just...weird. If you decide to stay at home to take care of them and that stops you from driving to work, the effect on mortality will be vastly larger. (If you were walking to work or working from home but now you have to drive the kids to activities all over the place, the effect on mortality will also be vastly larger in the opposite direction.) This is the kind of poor reasoning I was arguing against.