STATISTICAL VS CLINICAL-SIGNIFICANCE IN PRODUCT TESTING - CAN THEY BE DESIGNED TO SATISFY EQUIVALENCE

Statistical issues associated with demonstrating significance between treatment groups (efficacy or superiority) and nonsignificance (equivalence) are presented and discussed. Methodologies for demonstrating efficacy of a product are proposed and contrasted, incorporating clinical and statistical criteria, with emphasis on situations in which placebo groups are precluded from the study design. Distinctions are drawn between study designs for demonstrating superiority and those for equivalence, including the determination of sample sizes needed for the different approaches. The "at least as good as" criterion is proposed as a reasonable alternative to that of equivalence in active control equivalence studies for demonstrating that dental product modifications or new products are efficacious.