Neuroblastoma is one of the most common solid tumours of childhood and is unique amongst paediatric cancers in that it results in the urinary excretion of catecholamine metabolites which are easily measured in spot urine samples and so is a condition for which screening may be considered. The continuing poor prognosis associated with late stage disease has stimulated great interest in this proposition. The Japanese have been undertaking pioneering studies of such screening since 1974 and since 1985, all 6 months old babies have been offered screening. Preliminary data would appear to suggest that screening is effective in greatly improving the survival of children with neuroblastoma. However there are difficulties associated with the interpretation of survival data since screening undoubtedly results in the detection of cases which would otherwise have remained 'silent', and the well known problems of lead-time and length-time bias complicate matters still further. The time is not yet ready for universal implementation of screening and further investigation is required.