adding a factor in the course of a design space characterization

an early decision on the design space of a pharmaceutical could imply that some process parameter is to be held fixed during production. therefore, the effect of such a parameter is not studied in the exploration of the design space with statistically designed experiments. later decisions may necessitate a changed setting of the parameter.

purpose and methods

in this paper we discuss a statistical design strategy that permit inclusion of a new parameter in the course of a design space characterization and the risks one could run in adopting the strategy. the main tool used is computer simulation based on an initial design of 48 runs and 12 factors, which may or may not have been completed before introduction of a 13d factor.

results and conclusions

if the original statistical design has a particular structure, new settings of the parameter can be introduced halfway the originally intended experimentation without having to start a completely new exploration of the design space.