(精选)Clinical Trials教学课件.ppt

? Scott Evans, Ph.D. and Lynne Peeples, M.S. Placebo Controlled Trial Limitations: If effective standard of care (SOC) treatment exists, then randomization to a placebo may not be ethical In order to get product approved, a sponsor does not necessarily need to show that the new product is better than another Instead it is necessary to show that the new product is “at least as good as” the SOC ? Scott Evans, Ph.D. and Lynne Peeples, M.S. Absence of Evidence is Not Evidence of Absence Thus we cannot conclude equivalence (or noninferiority simply because we fail to reject H0 ? Scott Evans, Ph.D. and Lynne Peeples, M.S. Noninferiority Trial Randomize to one of two arms New product or “active-control” (e.g., a standard of care (SOC) treatment that is already approved) Blind the intervention to patients and investigators Analysis compares the two arms with respect to response Need to show new drug is “no worse” than SOC product ? Scott Evans, Ph.D. and Lynne Peeples, M.S. Investigational Treatment Active Control Randomize to Treatment SAMPLE YES NO YES NO Population of Patients TIME Response Active Control ? Scott Evans, Ph.D. and Lynne Peeples, M.S. Noninferiority Trial Decide on “maximum difference that is clinically irrelevant” (this defines “practical equivalence” or the “noninferiority margin) Analysis Get a confidence interval (CI) for the difference between arms (new product minus SOC with respect to efficacy) and note if lower bound of CI is within the noninferiority region ? Scott Evans, Ph.D. and Lynne Peeples, M.S. A B D E F Superiority: HO: p1-p2 = 0 HA: p1-p2 ? 0 Noninferiority: HO: p1-p2 - D HA: p1-p2 = - D C Noninferiority Margin p1-p2 P1 = efficacy of new treatment p2 = efficacy of control group - D 0 STATISTICALLY SUPERIOR STATISTICALLY INFERIOR CLINICALLY NONINFERIOR CLINICALLY INFERIOR Efficacy ? Scott Evans, Ph.D. and Lynne Peeples, M.S. Noninferiority Study: Example A randomized, double-blind, multicenter study comparing the efficacy and sa