Playing safe and preserving integrity: making the FDA model work.

The proposed FDA model for Data Monitoring Committees (DMC) includes the recommendation to separate the reporting statistician from the co-ordinating centre statistician. The former writes reports for the DMC; the latter is the statistician responsible for the statistical design of the trial, its ongoing day-to-day operation, and the final analysis of the data. This model, while protecting the scientific integrity of the trial, risks producing inadequate reports to the DMC unless the reporting statistician understands the trial and the database. This brief paper describes methods by which the reporting statistician can gain the knowledge necessary to produce informative reports. It also emphasizes that reports produced for the purpose of making decisions during the course of a trial differ in purpose, and should therefore differ in form, from reports summarizing the data and results at the end of the trial. Copyright 2004 John Wiley & Sons, Ltd.