Major-Elechi BT, Sloan JA, Nov
Context: In planning oncology phase II and phase III clinical trials, the size of the expected effect for endpoints such as tumor response and overall survival are key parameters driving the sample size. We applied the empirical rule effect size approach, also known as the ½ standard deviation (SD) method, to define clinically significant effect sizes for overall survival and tumor response endpoints in a series of clinical trials. Methods: The observed effect size was calculated for 12 phase II and 27 phase III completed cancer clinical trials identified by experts as being notable. Results: The effect sizes of the phase II and phase III clinical trials ranged from -0.32 to 0.84 and 0.01 to 0.44 SDs respectively. Effect sizes for all but four of the phase II trials were less than a ½ SD. For phase III studies, the effect sizes for all but one study were below 0.4 SD and roughly 67% of them had an effect size smaller than 0.2 SDs. There were no differences across disease sites, although colorectal and breast trials did have slightly larger effect sizes. Conclusions: Even highly noteworthy existing phase II and phase III oncology clinical trials rarely achieve the ½ SD level of clinical significance. This method allows for more ready interpretation of the clinical significance of overall survival and tumor response endpoints. It allows for cross-study comparison across different endpoints. The method also facilitates study design as it directly builds clinical significance into the study.