The method for dose-response modeling that is widely called MCPMod allows a sponsor to measure the likelihood that particular dose-response curves are the right mathematical model for a given set of data. Since several different dose-response curves might be able to fit the data, how can you determine which is the best curve for your purposes?

The best curve, in this case, is the one that is the best approximate for the unknown, actual curve. The importance of getting the best curve can have important ramifications for your trial.

The benefits of choosing the best curve are obvious: Not only can sponsors move forward having faith in the doses that they have chosen, but they also know something more about how their new therapy works. The shape of the curve, particularly as it relates to untested doses, tells you something more about the chemistry of the new product. A mathematical model conveys information that fills in the gaps provided by the data.

However, this also highlights the consequences of choosing the wrong curve. The wrong curve, in this case, means that the calculated curve does not properly approximate the unknown, actual curve. Therefore, the information such a curve provides will not approximate the actual state of affairs. Sponsors will have a mistaken idea about the safety and efficacy of the new product, meaning that they will go forward into Phase 3 with mistaken dose-selection.

So, a lot resides on calculating the best possible dose-response curves. Given that there are many options, however, how does one choose?

One option is the MCPMod method. This method effectively takes several different possible curves, and calculates how likely none of them are to be the right fit. Once this null hypothesis is rejected using multiple comparison procedures (i.e. MCP) the method then uses modeling techniques to make prescriptions about which curve to use moving forward. (Incidentally, rejecting this null hypothesis also achieves the same objectives to a proof-of-concept trial. Can you see why?)

This intuitive method is one that requires complex calculation and specialized, rigorous programming (although there is one validated software tool that accomplishes this.) Usually completed between Phases 2 and 3 of a trial, the method can be used in conjunction with other dose-finding methods to ensure that you choose the right dose. It has the added benefit that if the choice of a dose needs to be adjusted at all at some point during a Phase 3 trial, the dose-response curve can help calculate what this dose should be, without any further dose-response studies.

Interested in learning more?