Child mortality has fallen mainly because far fewer young children die
of infections: vaccines, malaria nets, better treatment and
better-educated mothers did much of the work. The death rate is down 60%
since 1990, but the fall has slowed since 2015, newborn deaths have
fallen more slowly than older children's, and the 2025 aid cuts put the trend at risk.
Far fewer children now die of infections
Child mortality has declined mainly because far fewer young children die of infections. Vaccines, bed nets against malaria and better care for sick children are the best-documented reasons, and mothers who spent longer in school probably helped too, though how much each did comes from models rather than measurement146.
Since 1990 the share of children who die before their fifth birthday has fallen by 60%, and the number of those deaths from 13.0 million a year to 4.9 million in 20241. The fall has slowed since 2015, though, and newborns now make up almost half the deaths. After the aid cuts of 2025, one leading model projects the first rise in the number of child deaths this century11.
That's how many fewer children died before their fifth birthday in 2024 than in 1990, worked out from UN estimates
From one child in 11 to one in 27
In 1990, 93.5 of every 1,000 babies born alive died before turning five: about one in 11. By 2024 it was 37.4, about one in 271. The number of deaths fell a little further than the rate, by 63%, from 12.97 million to 4.86 million a year1. Had every country’s rate stayed at its 2000 level, 48 million more children under five would have died between 2000 and 2015 alone3.
Even now, 4.9 million deaths a year is one child every six seconds1. Between 2000 and 2024, 174 million children died before the age of five, close to the whole population of Bangladesh1.
The world figure also hides where the deaths now are. Eastern Asia cut its child deaths by 96%. Sub-Saharan Africa cut them by 26%, from 3.8 million to 2.8 million, and its share of the world’s child deaths rose from 29% to 58%1, partly because more babies are born there each year, so a falling rate still leaves a large number. Southern Asia accounts for another quarter1.
Infections fell fastest, deaths around birth slowest
The deaths that fell were mostly deaths from infection. Between 2000 and 2024 the under-five death rate from measles fell 86%, from diarrhoea 73%, from pneumonia 58% and from malaria 39%1. UN IGME, the UN group that produces these estimates, credits more immunisation, nutritional support, and a standard way of treating sick newborns and children1.
Models put numbers on some of the levers. Measles vaccine accounts for nearly 94 million of the 154 million lives, of all ages, that vaccines have saved since 197454. In Africa, insecticide-treated bed nets account for 68% of the 663 million malaria cases prevented between 2000 and 20157. And mothers learned more: women of child-bearing age in developing countries went from 2.2 to 7.2 years of schooling between 1970 and 2009, which one study credits with about half the fall in child deaths over those years6. (Schooling is still far from universal; see children out of school.)
Why newborn deaths fall more slowly
Newborn deaths are falling more slowly because most come from being born too early or from complications at birth, which UN IGME calls “multifactorial and more complex to prevent and treat” than the infections vaccines and nets beat. Deaths in the first month of life fell 45% after 2000, against 57% for children aged one month to five years1.
The death rate from prematurity fell only 32%, and newborns’ share of child deaths rose from 41% in 2000 to 47% in 20241. That helps explain why the overall pace slowed: with the infections beaten back, the deaths left were the harder ones. The death rate fell by 3.9% a year from 2000 to 2015, and by 1.5% a year from 2015 to 2024, a 61% slowdown1.
Vaccines alone saved 146 million young children, on one model
Vaccines have saved about 154 million lives since 1974, 146 million of them children under five, on the best-known model, which compares today’s world with one that never vaccinated at all4. It credits vaccination with 40% of the fall in infant deaths worldwide, and 52% in Africa. On the same model, a child under ten in 2024 was 40% more likely to live to their next birthday than without vaccines4.
WHO, one of the programme’s founders, funded the study4, so the estimate comes from inside the programme it credits, which is worth bearing in mind.
Six countries that beat their neighbours, and the programmes behind them
Some countries did far more than others with similar resources. Bangladesh, Ethiopia, Nepal, Peru, Rwanda and Senegal cut child deaths faster between 2000 and 2015 than their neighbours or countries with similar economic growth. A study of the six found common habits: officials used data to decide, primary health care was built out, communities were involved and donors were coordinated12.
The international programmes behind them are old. The World Health Assembly set up the Expanded Programme on Immunization in 1974, and Gavi, the vaccine alliance, was created in 2000. By its own account Gavi has helped immunise more than a billion children and halve child mortality in 78 lower-income countries5.
Much of the daily work is done by community health workers. Josephine Barasa, in Kenya, lost her paid post when funding was cut and kept visiting the women she served as a volunteer11.
How much credit each cause gets is still a model
Nobody has measured how much of the decline each cause produced. The shares come from separate models: vaccines are credited with 40% of the fall in infant deaths4, women’s schooling with 51% of the fall in under-five deaths between 1970 and 20096, and US aid through USAID is associated with 30 million under-five deaths prevented between 2001 and 20219. They cover different years and ages and measure different things, so they cannot simply be added, and no study has split the whole decline between them.
The USAID figure is the weakest of the three. It compares countries that received more aid with those that received less, which shows an association, not a cause, and two letters criticising the method were published in The Lancet9. Growth alone doesn’t explain the decline either: the six countries above beat others with similar economic growth12.
The numbers are also thinnest where most children die. The 20 countries with the highest child death rates account for 44% of the deaths but 11% of the data collected since 1990, and in fragile and conflict-affected states the most recent good survey is on average 8.7 years old1.
Sixty countries are off track for 2030
The world will not meet the UN’s target on current trends. The Sustainable Development Goal (SDG) for child deaths is 25 or fewer per 1,000 births in every country by 2030. By 2024, 134 of 200 countries had met it and six more were on track1. The other 60, home to about 42% of the world’s children under five, are off track, and 50 of them would need to at least double their pace1.
Fragile and conflict-affected countries have a quarter of the world’s births and half its child deaths1. If death rates simply stayed at their 2024 levels, rising births in sub-Saharan Africa would push the yearly toll back up to 5.0 million by 20301.
The 2025 aid cuts could bring the first rise in child deaths this century
The number of child deaths may be rising again, though no measurement shows it yet. The UN’s death rate did tick up once recently, from 39.2 in 2020 to 40.0 in 2022, while the number of deaths kept falling (from about 5.26 million to 5.22 million); the report doesn’t say why, and the rate fell again after2.
The new risk is money. Development assistance for health, the aid that pays for much of this work, fell by a fifth to a quarter in 2025. The Institute for Health Metrics and Evaluation (IHME), a university research centre, first estimated a 21% fall, from $49.6 billion to $39.1 billion, with US funding down 67%10; its December update put the fall at 26.9%11.
IHME’s model projects under-five deaths rising from 4.6 million in 2024 to 4.8 million in 2025, the first rise this century11. IHME counts fewer deaths than the UN to begin with (4.6 million against 4.9 million for 2024), so the rise only holds in its own figures. The USAID study projects 4.5 million more under-five deaths by 2030 if the US cuts last, assuming aid keeps working as it did from 2001 to 20219. Bill Gates, whose foundation published the projection, put it this way11.
Malaria shows how little slack there is. Deaths worldwide rose from 578,000 in 2015 to 610,000 in 2024, which WHO puts down to population growth: the death rate kept falling, from 14.9 to 13.8 per 100,000 people at risk. In Africa three in four malaria deaths are children under five8. Separately, resistance is spreading to the insecticides on the bed nets that did most of the work in Africa, and to the main drugs87. The cuts also hit the household survey programme that supplies much of the data behind these estimates1.
The first measured count for 2025 is due in 2027
The 35-year decline is a measurement; the 2025 rise is so far a projection. UN IGME’s next round of estimates will be the first to measure 2025. If it shows more children dying than the year before, it will be the first such year in the UN’s series this century2. If cuts of a fifth persist, IHME’s modelling puts the cost at 12 million more child deaths by 204511.
Our reading is that the next few years will be the hardest to see clearly: the cuts that could raise the toll also defunded the surveys that would show it. UN IGME’s own assessment is that what holds progress back is money, uneven delivery, inequality and political commitment, not a lack of treatments that work1.