Objective: To estimate the difference between the number of subjects actually included in pediatric sequential trials and the sample size that would have been included with a fixed-sample design.
Study design and setting: A systematic review of pediatric sequential trials was performed. Methodological quality was assessed using a criteria list based on the CONSORT (Consolidated Standards of Reporting Trials) statement, and data were extracted by two reviewers independently. Where possible, fixed sample size calculations were performed using the same assumptions as those of the sequential design, and compared with the reported number of included patients.
Results: Twenty-four sequential trials, published between 1963 and 2005, were found. In nine studies, the information about the assumptions was sufficient to calculate a fixed sample size. The median reduction in included sample size in these trials compared with the fixed sample size calculation was 52 subjects (range: -22 to 229), a reduction of 35% (range: -42% to 90%) of the fixed sample size. The median sample size reduction when considering the number of subjects included in the analysis until crossing of the boundaries was 77% (range: 15-90%).
Conclusion: Sequential design is a useful method for optimizing the sample size in pediatric clinical trials and may lead to substantial sample size reductions.