Patient subsets and variation in therapeutic efficacy.

摘要: 1 The determination of what treatment is best for kinds patients a general objective clinical research. We consider here the extent to which this can be accomplished reliably in single trial. 2 reliability subset analyses often poor due problems multiplicity and limitations numbers studied. Implications design trials are presented. 3 Statistical approaches analysis reviewed manner. In order obtain degree usually demanded therapeutic evaluations, "statistically significant' interactions between relative efficacy subsets should demonstrated. Exploratory differences important but reported as hypotheses tested separate studies.