EMA Adopts MCP-Mod Dose-Finding Studies Under Model Uncertainty
The European Medicines Agency (EMA) has adopted use of the “multiple comparison procedure — modeling,” or MCP-Mod, approach to dose response testing and estimation.
“The analysis of dose finding studies can be classified into two major strategies: multiple comparison procedures and modelling techniques but none of these alone represent a comprehensive approach,” the EMA says. The agency’s Committee for Medicinal Products for Human Use recognizes that MCP-Mod may present benefits that may also come from other methodologies and thus, should be used alongside them.
Model uncertainty will remain after Phase II and users of MCP-Mod must still decide how to incorporate information, the EMA said. Additionally, “the model describing dose response may be updated as further information comes to light,” it added.
MCP-Mod works by first allowing researchers to design candidate models in the trial design stage and then test doses using trend tests in the trial analysis stage. Once researchers establish a dose response signal, they select the best of the candidate models. MCP-Mod also allows for investigators to draw conclusions with control of false-positive error rate, which is mandated in the confirmatory developmental phase.
The opinion reflects an industry shift toward evaluating the full dose response relationship and estimating the minimum effective dose. Last year, the FDA issued an addendum to International Conference on Harmonisation guidance it adopted that encourages sponsors to provide rationale for dose selection and to account for characteristics of the dose-response relationship. It also suggests adding pharmacokinetic-pharmacodynamic approaches may help.
The opinion does not apply to vaccines, biologics, gene or cellular therapies. Read the EMA’s opinion at www.fdanews.com/ext/resources/files/02/02-11-14-MCP-Mod.pdf. — Lena Freund
Subscribe to Drug Industry Daily for complete coverage of the pharmaceutical industry. Click here for more information.