Introduction: For physicians and other medical professionals, it is important to find relevant information on their field of interest. Various search filters, especially for identifying clinical trials and systematic reviews, have been developed. The aim of this study was to develop a PubMed search filter for identifying studies including children for the Cochrane Childhood Cancer Group (CCG), and to validate this search filter. Materials and Methods: The CCG search filter has been developed by identifying keywords and MeSH headings from relevant publications and by discussions with experts in the field. We validated the search filter by calculating the sensitivity and precision, and compared these results with six other known search filters for children (PubMed Limit All Child, three by Kastner et al, Child Health Field, and BestBets). Sensitivity was calculated by dividing the number of retrieved relevant randomized controlled trials (RCTs) and controlled clinically trials (CCTs) by a Gold Standard set of relevant studies. This new Gold Standard was developed by collecting all RCTs and CCTs reported in Cochrane intervention reviews that included children (1951 2008). The precision of the 7 filters was calculated as Number Needed to Read (NNR) in a set of records focussed on children obtained from PubMed in September 2008. All records were assessed by two authors individually to determine whether a study included children or not. In case of diagreement a third author was consulted. Results: The five child search filters from CCG, Child Health Field, BestBets, Kastner (filter for best sensitivity ) and PubMed showed a high sensitivity to detect studies including children (98.2, 98.1, 98.2, 97.6 and 97.3 respectively). Two of the three child search filters from Kastner et al performed less well in PubMed; their sensitivity was 46,4 and 54,2 respectively. These filters were regarded as the filter for best specificity and the filter for best optimization of sensitivity and specificity by the authors. The results of the differences in precision will be discussed at the Colloquium. Conclusion: All search filters for identifying studies including children performed rather similar, except for two search filters developed by Kastner et al, which performed less well. Based on the results of this study recommendations can be made fot the improvement of the CCG child filter.