Design Analysis Indicates Potential Overestimation of Treatment Effects in Randomized Controlled Trials Supporting FDA Cancer Drug Approvals