LONDON (AP) -- Child deaths worldwide seem to have fallen faster than officials thought, as a new study estimates far fewer children are dying every year than previously guessed by the United Nations.

Using more data and an improved modeling technique, scientists predicted about 7.7 million children under 5 would die this year, down from nearly 12 million in 1990. The study was published online Monday in the British medical journal, Lancet.

The new estimate is substantially lower than UNICEF's last estimate of child deaths from 2008. Then, the agency said about 8.7 million children were dying every year of preventable causes such as diarrhea, pneumonia and malaria.

"We're quite a bit farther ahead than we thought," said Christopher Murray, one of the paper's authors and director of the Institute for Health Metrics and Evaluation at the University of Washington.

Murray and colleagues assessed information from 187 countries from 1970 to 2009. They found child deaths dropped by about 2 percent every year, lower than the 4.4 percent needed to reach the U.N.'s target of reducing child deaths by two-thirds by 2015.

Murray said death rates were falling surprisingly fast in countries including Liberia and Niger, but that progress had stalled in rich countries like Britain and the United States.

Where information was limited, researchers used modeling projections to estimate the number of deaths. The research was paid for by the Bill & Melinda Gates Foundation.

"We're very excited because this study reinforces our belief that the scale-up of interventions such as (malaria) bed nets, vaccines and vitamin A (pills) are starting to show an impact," said Mickey Chopra, UNICEF's director of health. UNICEF was not linked to the study.

He was particularly impressed with progress in countries like Niger and Malawi, where there hasn't been much economic growth. "Even while we wait for poverty levels to reduce, there is still a lot we can do to improve health," he said.

But given the fluctuating figures common in global health, not everyone was swayed by the new research. There was also no evidence to show U.N. programs are responsible for the drop in child deaths. Murray said the reduction could be because the AIDS epidemic peaked several years ago and is now hitting fewer kids.

William Aldis, a former senior World Health Organization expert who has worked in Africa and Asia, said the numbers were not entirely reliable, given the uncertainty in the data from poor countries with patchy surveillance.

Mit Philips, a health analyst at the charity Medecins Sans Frontieres, said she was encouraged fewer children were dying, but much more still needed to be done. "If you start off with a very high mortality rate, like some places in sub-Saharan Africa, even if you reduce it, it's still very high," she said.

Philips said improving health in poor countries required comprehensive health services, not just one-off campaigns handing out malaria bed nets or administering vaccines.

"Whatever the modeling shows, the picture on the ground can be quite different," she said. "In some places in southern Africa, there is a lot less progress than you'd expect."

Copyright 2010 The Associated Press. All rights reserved. This material may not be published, broadcast, rewritten or redistributed.