Data editing in a clinical trial.

In clinical research and, particularly, in multicenter clinical trials, data are collected onto study forms designed to minimize errors in the recorded data. Nevertheless, errors of various kinds must be anticipated, and it is necessary to devise means to detect, review, and correct erroneous values. General requirements of a data editing system and methods for accomplishing these tasks are presented. A general computer program for error detection and a method for updating the data base that leaves an "audit trail" of the original data and subsequent corrections are described.