Statistical analysis plans for clinical trials—reply.

The protocol should contain the principal features of the intended analyses in sufficient detail so that all parties involved in making the decision to start the trial can judge whether the planned trial is ethical and scientifically and practically acceptable. Further development in the form of a SAP gives all concerned more time, permits useful reaction to accruing external information while providing others (regulators, editors, and readers) the necessary guarantees that choice of analyses were not (unfairly) driven by desired results.