Invisible Loss

Most miscarriages happen because of a mistake in cell division. The egg and sperm meet, the chromosomes line up, and something goes wrong in the first few copies. An extra chromosome, a missing piece. The embryo grows for a few days or weeks, and then it stops. Chromosomal errors explain somewhere between 50 and 65 percent of first-trimester losses, and most of them are random. Not inherited. Just bad luck at the cellular level.

This happens in about 15 percent of recognized pregnancies. A 2021 Lancet series put the global number at 23 million miscarriages a year, which is 44 every minute. Probably more, because a lot of very early losses happen before a woman knows she is pregnant. A period comes a few days late. Nothing gets recorded. Nothing gets counted.

That last part is the part I keep thinking about.

I read Pachinko a few months ago. It follows a Korean family across the twentieth century, and one thing it does quietly is show women losing pregnancies. Not as tragedy exactly. More like texture. Something that happened, was absorbed, and life continued. The pregnancies that survived were the ones you counted.

My grandmother's generation wasn't far off. Different country, mid-century India, same shape. My maternal grandmother had nine miscarriages. I only know because my mother told me. There is no clinic note, no registry entry, no number in any dataset. Nine pregnancies, and the only place they exist is in something a mother said to her daughter.

Why Did It Change?

So what changed? The obvious answer is hospitals and prenatal care, and that isn't wrong. But when I went looking for data to say exactly what changed and by how much, I found that large parts of the record simply don't exist.

Here is what we do know. The biology hasn't changed. Chromosomal errors are as common now as they were then, maybe more so since people have children later. Global miscarriage cases did fall between 1990 and 2019, from about 49.6 million to 42.4 million a year. But deaths from miscarriage and its complications fell much faster, from around 59,000 to under 20,000. The pregnancies are still being lost. Women are just surviving them.

So the change wasn't really about stopping the biology. It was about everything around it. Iodine deficiency raises miscarriage risk. Poor nutrition makes every other risk worse. Infections may cause up to 15 percent of early losses and up to 66 percent of late ones. Untreated syphilis leads to bad pregnancy outcomes in over 40 percent of cases. Malaria triples or quadruples the risk of first-trimester loss. The twentieth century's war on infectious disease was also, quietly, a war for surviving pregnancies. And when something did go wrong, ectopic pregnancy, hemorrhage, preeclampsia, hospitals caught the cases where failure didn't have to mean death.

The Data Gap

Miscarriage is one of the most common things that happens in human reproduction and one of the least tracked. The same Lancet series that produced the 23 million figure asked for standardized global data collection, because none exists. Countries don't report miscarriage the way they report maternal deaths or stillbirths. Early loss happens at home, before anyone enters a hospital, in places with no way to record it. Research on low- and middle-income countries found underreporting in household surveys is everywhere, driven by stigma, memory, and the plain fact that no system was ever built to catch it.

So when you try to map global miscarriage data, you don't get a picture of where miscarriage happens. You get a picture of where it gets counted.

The map of the data is not a map of the phenomenon. It's a map of who was watching.

What the Map Shows

Look at it for two seconds and you'll see the point. A small cluster of countries is dark: Western Europe, North America, parts of East Asia. Everywhere else fades out. Sub-Saharan Africa, South Asia, Central Asia, most of the Middle East. Those aren't real numbers. They're estimates built from models and regional averages, standing in for pregnancies nobody wrote down.

The color is a map of attention, not suffering. The dark parts are just the places where someone was watching closely enough to keep a record.

My grandmother's nine miscarriages are somewhere in the pale part. This is what that looks like at scale.

We know a lot about the causes of loss that medicine learned to fix. We know much less about the baseline, the real total across all people and all of history, because nobody collected it. The record we have is mostly a record of places that were already well resourced. Women who weren't near a clinic, weren't in a country with a registry, weren't in an era where anyone thought to ask, mostly didn't show up at all.

Researchers responding to the Lancet series said the true burden is much higher than what gets reported, and asked for studies from low-income countries to fill the gap. That was 2021. The gap is still open.

This isn't only a miscarriage problem. Women's health research has been underfunded and underreported for a long time. The holes in the record aren't accidents. They're decisions about what was worth counting.

The biology of miscarriage hasn't changed. What changed, in the places where anything changed, was the world around the pregnancy. What a woman could eat. What infections she carried. What happened when something went wrong, and whether anyone was there to help. Those aren't small things. They're the difference between a generation of women who expected to lose pregnancies and a generation that mostly doesn't.

But that's only half the story, because we only have that story for part of the world. For the rest, there's the pale wash on the map. The pregnancies nobody counted. The loss that went unrecorded because no one built a way to record it.

My grandmother doesn't appear in any dataset. Neither do most of the women alive when they were. The gaps in the map aren't blank. They're full of people.