America's lowest child death rate was 2022. Satoria's central forecast reaches 6 per 1,000 by 2034.

Satoria's forecast

Satoria forecast for United States: children who die before their fifth birthday, per 1,000 born 5.98 per 1,000 births in 2034, down from 6.48 in 2024 - could go either way. 80% range 4.24 to 6.93.

The low point was 2022, and our path does not beat it until 2027.

The lowest measured reading is 6.46 deaths before age five per 1,000 births in 2022. The two years since were both higher, and our central path stays above the low point through 2026, at 6.48. Only then does the line start falling again, reaching 5.98 in 2034.

It is a risk per 1,000 births, not a count of deaths.

It asks how many out of 1,000 babies born this year would die before turning five at this year's death rates. It is not a percentage. For a parent it is the plainest measure of how safe the first five years are where you live, and in rich countries almost all of that risk sits around birth.

The world is still winning this, just far more slowly.

An estimated 4.9 million children died before their fifth birthday in 2024, and deaths before age five have more than halved worldwide since 2000. But since 2015 the pace of that fall has slowed by more than 60 per cent. America's own line shows the same thing in miniature.

Where does America land in 2034?

America fell from 7.6 deaths before age five per 1,000 births in 2009 to 6.54 in 2019, every single year. Then it almost stopped: 6.48 in 2024. Move the slider to where you think our model puts it ten years on.

Our own range allows ten more years without progress.

The published range for 2034 runs from 4.24 to 6.93 deaths before age five per 1,000 births. Its top is above the 6.48 measured in 2024. In rich countries nearly half of these deaths are newborn deaths, so care around birth is what would move the line either way.

One reading would settle whether 2022 was the floor.

America has been between 6.46 and 6.48 deaths before age five per 1,000 births since 2021. A year clearly below 6.46 would be the first real progress since then, and our central path does not expect one before 2027. We will record either answer.

What the world counted, and what the measure means

The joint release of March 2026 puts the world figure at 4.9 million children dying before their fifth birthday in 2024, including 2.3 million newborns, and says deaths before age five have fallen by more than half since 2000. Since 2015 the pace of that fall has slowed by more than 60 per cent.

The measure itself is a probability, not a count. The World Bank defines it as the chance per 1,000 live births that a baby born this year dies before turning five, worked out from this year's death rates at each age. It is not a percentage, and writing it as one would make it look forty times smaller than it is.

Where the deaths happen has changed the arithmetic for rich countries. Newborn deaths are nearly half of all deaths before age five, led by complications from preterm birth at 36 per cent and complications during labour and delivery at 21 per cent. In Europe and Northern America the leading infectious diseases cause only 9 per cent of these deaths, against 54 per cent in sub-Saharan Africa.

America fell for ten years, then stopped

In 2009 America recorded 7.5984 deaths before age five per 1,000 births. By 2019 it was 6.5395, and it had fallen in every one of those years.

Then the line went flat. The lowest measured reading is 6.463 in 2022. The series gives 6.4701 in 2023 and 6.479 in 2024, so both years since the low point are slightly higher than it.

The whole move from 2019 to 2024 is from 6.5395 to 6.479, which is smaller than what used to happen in a single year. These are estimates and the differences are small, so the honest way to read them is as a stall rather than as a rise.

What our model says next

Our central path does not undo the stall quickly. It has America at 6.4808 deaths before age five per 1,000 births in 2026, which is still above the 2022 low point of 6.463, and only then starts falling again.

By 2034 the central path reaches 5.9793. That is half a point of progress across ten years, where the 2010s delivered more than a whole point in the same time.

The published range for 2034 runs from 4.2448 to 6.9254. The ceiling is above where America stands today, so our own range allows a decade with no progress at all, and the floor allows a return to the pace of the 2010s. The model is well tested on this measure: on data held out from fitting and model choice, across 49 country series, its errors were 71% smaller than assuming no change. That is a good score for ordering countries, and it is still a wide range for one of them.