Leveraging an Experimental Database to Inform Design of Experiments Planning: Application to OMARS Design Selection

ABSTRACT Design of Experiments (DoE) studies generate valuable process knowledge during pharmaceutical development, and aggregating scientifically related studies can make this knowledge more reusable for future experimental planning. However, prior study results are often difficult to reuse because data, metadata, and statistical model outputs are stored in inconsistent formats across studies. This paper describes the development of a centralized experimental database and statistical decision‐support workflow, implemented in a data science workbench, for standardizing and compiling historical process characterization DoE studies in alignment with FAIR principles. The database captures raw data, study metadata, process parameter information, response information, and selected model artifacts, enabling cross‐study comparisons and dashboard‐based review. We demonstrate the value of this framework through a case study in which historical patterns of effect significance and co‐significance were used to inform OMARS design selection and factor‐to‐column assignment for a process characterization study. This work presents a practical statistical knowledge‐management framework that moves beyond ad hoc data reuse to support more transparent, reliable, evidence‐based experimental planning.