AbstractEarly identification of children with develop-mental delay is a longstanding goal in pediatric research and practice, and a large number of screening and assess-ment tools have been developed to this end. Unfortunately, these measures are problematic: Definitions ofBdelay^are not consistent, thresholds are somewhat arbitrary, some scoring systems are affected by systematic error, and the natural variability in the timing of development may limit the accuracy they can achieve. Significantly, the ability of early screening topredict longer-term functional problems has proven to be very limited. False positive results are also not harmless, and the effectiveness of early interventions is in some doubt. The benefits of early screening using existing instru-ments and interventions are therefore doubtful. Some im provements may be possible, however, by making use of more information sources, by using variable thresholds, and by making more rigorous use of cost-benefit analysis. KeywordsChild development.Screening.Surveillance. Validation