TY - JOUR T1 - Breast feeding JF - Archives of Disease in Childhood JO - Arch Dis Child SP - 91 LP - 92 DO - 10.1136/adc.87.2.91 VL - 87 IS - 2 AU - A Nicoll AU - A Williams Y1 - 2002/08/01 UR - http://adc.bmj.com/content/87/2/91.abstract N2 - An effective but under applied intervention in Europe Over 95% of the world's children are initially breast fed and as the result of promotion programmes, prevalence has increased throughout the 1990s in many developing countries.1 Although breast feeding could improve the health of mothers and children in all parts of the world, ironically its potential has still to be realised in many of the world's wealthiest countries where prevalence remains low. United Kingdom data collected during the last quinquennial national survey of infant feeding (September 2000) have recently been released.2 They show only a minimal increase in England and Wales (70% of babies were breast fed at birth, as opposed to 68% in 1995), largely explicable by the confounding effects of increased maternal age and educational attainment.2 Significant increases were, however, seen in Scotland and Northern Ireland, traditionally areas of low breast feeding uptake, where there have been vigorous promotional campaigns.2,3 The UK picture is typical of much of Europe which has the lowest breast feeding rates of any global region.4,5 The reasons which underlie Europe's failure to capitalise on breast feeding as a child health promotion strategy need to be explored.Despite a wealth of evidence, some professional scepticism endures about the link between infant feeding and health outcomes. For some the randomised controlled trial (RCT) has become a sine qua non. However, the ethical problems associated with randomising mothers' feeding choices have usually meant that case–control and cohort study experimental designs have been the only practical methodological option. Among the principal research problems encountered have been inadequate definitions of “breast feeding” (for example, failure to distinguish “exclusive” from partial breast feeding), ascertainment bias, and the possibility that maternal choice was biased by disease risk (such as a family history of allergic … ER -