Child deaths since 2000: who cut them fastest, and where they are still highest

How many children die before their fifth birthday is one of the few health measures estimated for every country, every year, on one method, since 1990. This piece reads the under-five mortality rate published by the UN Inter-agency Group for Child Mortality Estimation (UN IGME) in its 2025 release: deaths before age five per 1,000 live births, for 200 countries. Every year from 1952 to 2024 on this site comes from that one release, so a difference between two years is a change in the estimates, not a revision of them. The figures are as they stood on 13 November 2026.

The middle country, then and now

In 1990 the middle country of the 200 had a rate of 44.830 deaths before age five per 1,000 live births. By 2000 the middle reading was 30.432; by 2024 it was 14.539, less than half the 2000 figure. These are medians — the country in the middle of the list, not an average weighted by births, which this site does not calculate.

The fall was broad. Of the 200 countries, 112 at least halved their rate between 2000 and 2024, and 44 cut it by two-thirds or more. The typical country fell by 52%. At the low end, 78 countries now lose fewer than 10 children in 1,000 and 40 fewer than 5; 31 of those 40 are in Europe. San Marino has the lowest estimate, 1.260, followed by Estonia at 2.016 and Luxembourg at 2.182.

Who cut it fastest

Measured as a share of where each country started, four of the five largest falls since 2000 are in Asia. The Maldives went from 39.482 to 5.439, a fall of 86%. China fell from 36.639 to 5.724 (84%), São Tomé and Príncipe from 82.167 to 13.554 (84%), North Korea from 100.235 to 16.666 (83%) and Cambodia from 105.179 to 18.410 (82%).

Measured in children, the picture changes, because the largest falls were possible only where the rate began highest. All ten of the largest absolute falls since 2000 are in Africa. Rwanda's estimate went from 184.158 to 37.737 — about 146 fewer deaths in every 1,000 births, a fall of 80%. Angola went from 184.988 to 49.000, Malawi from 170.984 to 48.783, and Sierra Leone from 222.837 to 90.500. Sierra Leone's 2024 figure is still among the highest in the world; its fall is among the largest. Both are true at once, and a list sorted by only one of them hides the other.

Where the rate is still highest

Three countries have estimates above 100 in 2024, and 27 have estimates of 50 or more. Twenty-three of the 27 are in Africa; the other four are Pakistan, Afghanistan, Kiribati and Haiti. Nigeria's estimate is the highest, at 115.649.

IGME publishes every estimate with a 90% uncertainty range — the span within which it judges the true figure most likely lies — and for the highest five, those ranges matter as much as the estimates:

Country	2024 estimate	IGME's 90% range
Nigeria	115.649	98.848 – 136.189
Niger	110.720	82.856 – 147.564
Somalia	101.066	46.406 – 222.820
Chad	97.258	63.276 – 147.578
South Sudan	96.711	24.936 – 283.181

The five ranges overlap. Nigeria's central estimate is the highest, but IGME's own figures do not separate it from Niger, Chad or Somalia with any confidence, and South Sudan's range runs from 24.936 to 283.181. Put plainly: the data says these five countries lose a very large share of their children, and does not say which loses the most.

Pakistan and its neighbours

Pakistan's estimate was 140.393 in 1990, 107.923 in 2000 and 55.959 in 2024, with a range of 44.018 to 70.360. That is a fall of 48% since 2000 — slightly less than the typical country's 52% — and 23 countries now have higher estimates than Pakistan's.

Its neighbours started from similar places and moved further. India fell from 91.782 to 26.580 (71%), Bangladesh from 85.514 to 30.498 and Nepal from 78.736 to 25.117. In 2000, Afghanistan's estimate of 131.707 was well above Pakistan's. In 2024 Afghanistan's is 53.032, with a range of 39.679 to 71.364 — so although Pakistan's central estimate is now the higher of the two, their ranges overlap almost entirely and the estimates do not say which is higher. At the other end of the same sub-region are the Maldives, at 5.439, and Sri Lanka, at 5.881.

The first month

IGME also estimates neonatal mortality: deaths in the first 28 days of life, per 1,000 live births. Set against the under-five rate, it shows where in childhood the remaining deaths fall — and here Pakistan stands apart.

Pakistan's neonatal rate in 2024 is 36.09850, the third highest of the 200 after South Sudan's 39.46577 and Nigeria's 39.04572. Read against its under-five rate of 55.959, that means close to two-thirds of the children Pakistan loses before age five die in their first month. In Nigeria the share is about a third: its neonatal rate is a little higher than Pakistan's, but its under-five rate is twice as high, because so many more children die after the first month.

Across the world, the first month's share tends to be larger where the overall rate is lower. Among the 27 countries with under-five rates of 50 or more, the middle country loses 38% of those children in the first month; among the 78 below 10, it is 55%. Across all 200, the middle share rose from 49% in 2000 to 53% in 2024. Pakistan breaks the pattern: its rate is in the highest group, and its first-month share, 65%, is above even the low-rate countries'. India (63%) and Bangladesh (59%) show the same concentration at lower rates. In South Asia, what remains of child mortality is concentrated around birth.

Where it rose, and where it stalled

Six countries have higher estimates in 2024 than in 2000: Dominica (19.371 to 35.658), Palestine (30.318 to 37.854), Fiji (22.631 to 29.123), Venezuela (21.520 to 24.200), Grenada (15.513 to 17.960) and Seychelles (13.776 to 13.959). Several are small countries, where a handful of deaths moves the rate.

Over the last nine years the list is longer. Between 2015 and 2024, 23 countries' estimates rose, among them some of the lowest in the world: Cyprus from 2.504 to 4.979, Croatia from 4.913 to 5.646, the United Kingdom from 4.458 to 4.680 and France from 4.167 to 4.343. Elsewhere it rose further from a higher base: Lebanon from 12.930 to 17.601, Zimbabwe from 59.237 to 64.661.

Even where the rate kept falling, it fell more slowly. Split the years since 2000 into two halves of twelve and the typical country cut its rate by 36% in the first and by 28% in the second; 149 of the 200 fell more slowly in the second half. The decline is real and nearly universal, and it has lost pace.

How sure the estimates are

An under-five rate is a modelled estimate, fitted to whatever censuses, household surveys and death registration a country has. Where a country registers every birth and death, the range is tight: Japan's 2.434 has a range of 2.356 to 2.515. Where the data behind an estimate is thinner, the range widens. Botswana's 2024 estimate is 33.311, with a range from 9.591 to 110.954 — wide enough to hold a rate near Europe's and one near the Sahel's. In the middle country, the range is about half the size of the estimate itself.

That is the reason this site prints IGME's ranges beside every estimate on the statistic page and in the 2024 table: two countries a few points apart are often not distinguishable at all, and an estimate whose upper bound is more than ten times its lower bound is a statement about the data as much as about the children.

What the measure does not show

The rate counts deaths. It says nothing about the health of the children who survive — how many are stunted, or how many are vaccinated — which the site publishes as separate statistics under Health & Wellbeing. It is also a rate, not a count: a small country and a very large one with the same rate lose very different numbers of children. Each country's series, back to 1952 where IGME has one, is on the statistic page with its range year by year, and that page — not this one — is updated when IGME publishes its next release.