Definition

Design of Experiment—DOE—is a structured, statistical approach to understanding how multiple process variables affect an outcome, tested simultaneously rather than one at a time. In pharmaceutical process development, it is used to map the relationship between input parameters and critical quality attributes, identify which factors actually matter, and define the boundaries within which a process reliably performs.

DOE in Process Development—Why It Matters More Than It Might Seem

The traditional alternative to DOE is one-factor-at-a-time (OFAT) experimentation: change one variable, observe the result, fix it, move to the next. The problem with OFAT is not just efficiency—it is that it fundamentally cannot detect how variables interact with each other. In chemical synthesis, interactions between parameters are common and often significant. A reaction temperature that gives excellent yield at one solvent ratio may perform very differently at another. OFAT will miss this entirely; a well-designed DOE will capture it.

In API process development, DOE is applied to understand how temperature, pH, solvent composition, reagent stoichiometry, reaction time, mixing intensity, and other parameters collectively affect yield, purity, impurity formation, and particle characteristics. Screening designs—fractional factorial or Plackett-Burman—narrow the field to the variables that genuinely matter. Response surface designs then map the landscape of those key variables in detail, generating the data needed to define a process design space.

The design space concept comes directly from ICH Q8 on pharmaceutical development and is central to Quality by Design submissions. DOE data supports the establishment of Proven Acceptable Ranges and Normal Operating Ranges—the specific numerical commitments that go into regulatory filings and constrain how the process can be run commercially. DOE is therefore not just a development tool; it generates regulatory-grade process understanding data.

Related Topics

Related Terms

  • API Route Scouting
  • Route of Synthesis for an API
  • CMC (Chemistry, Manufacturing and Controls)
  • cGMP (Current Good Manufacturing Practice)
  • API CDMO

FAQs

1. What is the main advantage of DOE over OFAT?

DOE evaluates multiple variables simultaneously and captures interaction effects between them. OFAT cannot detect interactions at all. The result is more comprehensive process understanding with fewer total experiments.

2. What DOE designs are used most often in API process development?

Screening designs—Plackett-Burman and fractional factorial—identify which factors matter. Response surface designs—central composite and Box-Behnken—then map those factors in detail to optimise the process and define the design space.

3. How does DOE support regulatory submissions?

DOE generates the multivariate data needed to define a pharmaceutical design space under ICH Q8, establish Proven Acceptable Ranges, and demonstrate the process understanding that regulatory agencies expect in Quality by Design submissions.

4. When in development is DOE typically applied?

Screening DOEs can be useful even in early development. Optimisation studies are most common during Phase 2/Phase 3 development. Late-stage DOE work supports process validation and establishes the commercial control strategy.

5. Is DOE a standard part of CDMO process development services?

At any CDMO with serious process development capability, yes. For NCE programmes where regulatory expectations for process understanding are high, DOE is essentially standard practice.

 

Learn how Neuland's process development team uses DOE to build robust, well-characterised API processes. Explore Custom Development