Empirical research and observation predates written history. Aristotle and Archimedes extolled the virtues of deductive reasoning. Early scientists like Hippocrates and others used empirical methods in their research and writing.

Scientists are taught to frame their experiments in a hypothesis, which should be stated as a fact, so research can poke holes in it. The hypothesis is formulated before it is proved or disproved by experiments.

The second framework for experimental design is to build a model, which provides an explanation for the data set. A model is different from a hypothesis, because it is created after data is collected. A hypothesis is created to be proved or disproved, while a model is created to be verified. Flawed hypotheses are often scrapped completely, while flawed models are often refined into another iteration.

The top- down deductive reasoning stemming from a hypothesis and the bottom-up inductive reasoning of a model have worked for eons, but do they work for pharmaceutical research in the 20th century? There is a better way – Quality by Design.

Pharmaceutical research has used an empirical approach, which involves continued experimentation and product testing to determine efficacy and quality. It also causes lost time and lost batches in manufacturing. Quality by Design was outlined by Joseph M. Juran, who posited that quality could be planned and incorporated into the R&D process from the beginning. It is now part of the FDA imperative that includes Risk Based Monitoring.

Quality by Design looks at the end product and the compatibility of all of the components that will go into the end product. It sets critical formulation attributes and process parameters at the beginning of the process. The most important aspect of Quality by Design is that it saves time and money and builds business flexibility into the development, testing and manufacturing process.

We still need to do research and conduct experiments. With Quality by Design, we build into that research an awareness of the end goals and a reliance on the deep domain expertise of the researchers.