Comparison between continuous and discrete doses using Escalation With Overdose Control (1708.04792v1)

Abstract: Although there is an extensive statistical literature showing the disadvantages of discretizing continuous variables, categorization is a common practice in clinical research which results in substantial loss of information. A large collection of methods in cancer phase I clinical trial design establishes dose of a new agent as a discrete variable. A noteworthy exception is the Escalation With Overdose Control (EWOC) design, where doses can be defined either as continuous or as a grid of discrete doses. A Monte Carlo simulation study was performed to compare the operating characteristics of continuous and discrete dose EWOC designs. Four equally spaced grids with different interval lengths were considered. The loss of information was measured by several operating characteristics easier for clinicians to interpret, in addition to the usual statistical measures of bias and mean squared error. Based on the simulations, if there is not enough knowledge about the true MTD value as commonly happens in phase I clinical trials, continuous dose scheme arises as an attractive option.