One reason why so little is known about the epidemiology of atopic dermatitis (AD) is lack of suitable diagnostic criteria. A simple list of diagnostic criteria for AD for use in epidemiological studies has recently been developed by a U.K. working party. These have performed well in hospital validation studies of subjects with skin diseases. This study sought to validate the newly proposed criteria for AD in a population setting by conducting a cross-sectional survey of 695 schoolchildren aged 3-11 years in three randomly selected primary schools in West Lambeth, London. As a point prevalence measure, the U.K. criteria had a sensitivity of 70%, a specificity of 93%, and a positive predictive value of 47% when compared with a dermatologist's examination findings. Subsequent analysis suggested that most children classified as false positives had suffered from AD in the last year, but were inactive at the time of examination. When adjusted for these cases, the sensitivity and specificity increased to 80 and 97%, respectively, corresponding to positive and negative predictive values of 80 and 97%, respectively. The U.K. diagnostic criteria for AD appear to work well as a 1-year period prevalence measure in London schoolchildren. Further validation in adults and other countries are needed.