Key points are not available for this paper at this time.
A RECURRING decision that has to be made in planning a controlled clinical trial of a new treatment is the size of the patient sample necessary. In this decision, a major determinant is ethical, for the investigator must keep at a minimum the number of patients exposed and the duration of the exposure to the inferior treatment. To achieve this ethical goal, the investigator may be tempted to analyze his results repeatedly as data accumulate. Just such a procedure was apparently used in a study recently reported in the Journal.1 , 2 Yet, as my communication will show, the use of conventionally . . .
Klim McPherson (Thu,) studied this question.