To the Editor: The study by Drs Psaty and Kronmal1 and the accompanying Editorial by Drs DeAngelis and Fontanarosa2 regarding rofecoxib illustrate the recurring problem of discretionary or data-driven analysis. Given the possibility of selective analysis based on observed data and the risk of positive results due to chance alone, it is critical to know in detail which analyses were prespecified and when the prespecification occurred. This can only be done via a full protocol repository, as we suggested in 2002.3 Our proposal would establish a secure real-time repository for protocols, amendments, and statistical analysis plans. The material would be kept confidential for intellectual property reasons during the clinical trial and its analysis and then made public at the time of publication. The proposal has yet to be implemented, although the Food and Drug Administration (FDA) Act of 2007 (US HR3580), which provides for a Web site for results of all phase 2, 3, and 4 trials,4 also requires posting the protocol.