Infant deaths down to ten million – or don't we really know?

Contradictory accounts of the last few weeks paint a puzzling picture of infant mortality trends. Few outcomes are as clear-cut and none is more tragic than the death of a child; concerns about the reliability of the science that drives the statistics can only add to the agonies inside international health organizations.

UNICEF reports encouragingly that the number of children dying before the age of five has dropped below ten million for first time since records began. It says the benefits of immunization campaigns, breast feeding, mosquito nets, Vitamin A supplements and other measures have combined to drive down deaths in 50 countries from 13m in 1990 to 9.7million in 2005-2006.

The Lancet [1] is less optimistic. An article by Christopher Murray and colleagues suggests that the international community is not doing any better a job of reducing child mortality than it was 30 years ago. It seems unlikely, according to this analysis, that the Millennium Development Goal for 2015 – to reduce by two thirds the mortality rate among children under five will be met.

This and other articles in The Lancet also cast doubt on data collection techniques and consequently on the reliability of the information in UNICEF surveys. The current editorial is about the quality of the science inside the World Health Organization and UNICEF; its subtitle "the corrosion of trust" gives an idea of the flavor.

UNICEF continually point out that most of the 21 deaths a minute among children under five is preventable. The main causes are AIDS, birth defects, diarrhea, malaria, measles, malnutrition, pneumonia and premature births. For example, immunization has reduced measles deaths in Africa by 75%.

An article in The Lancet by Christian Lengeler and Don deSavigny illustrates what can be done. Their work shows how treating bednets with insecticide reduces child mortality by 44%. The prevention technique is enhanced by social marketing and free distribution of impregnated material.

All commentators agree there is much to do. The chasm between rich and poor countries remains huge. Infant mortality rates run at around six per 1,000 births in industrialized countries. In Afghanistan it is six in every 30. In sub-Saharan Africa it is over 150 per 1,000. And despite significant improvements in India, in the midst of frenzied change the rate there is still 76 per 1,000.

The conclusion seems to be agreement that the trend is in right direction, but that there is much more to do – and that the quality of science has to improve markedly.

• references and links

The Lancet audio series [2] includes a podcast looking at Millenium Development Goals and the benefits of insecticide treated bednets.

Christopher Murray, Julio Frenk and Timothy Evans, "The Global Campaign for the Health MDGs: challenges, opportunities, and the imperative of shared learning", The Lancet, 370, 9592, pp 1018-1020.

Christian Lengeler and Don deSavigny, "Programme diversity is key to the success of insecticide-treated bednets" The Lancet, 370, 9592, pp 1009-1010.