Design of experiments (DoE) in pharmaceutical development

Abstract

At the beginning of the twentieth century, Sir Ronald Fisher introduced the concept of applying statistical analysis during the planning stages of research rather than at the end of experimentation. When statistical thinking is applied from the design phase, it enables to build quality into the product, by adopting Deming's profound knowledge approach, comprising system thinking, variation understanding, theory of knowledge, and psychology. The pharmaceutical industry was late in adopting these paradigms, compared to other sectors. It heavily focused on blockbuster drugs, while formulation development was mainly performed by One Factor At a Time (OFAT) studies, rather than implementing Quality by Design (QbD) and modern engineering-based manufacturing methodologies. Among various mathematical modeling approaches, Design of Experiments (DoE) is extensively used for the implementation of QbD in both research and industrial settings. In QbD, product and process understanding is the key enabler of assuring quality in the final product. Knowledge is achieved by establishing models correlating the inputs with the outputs of the process. The mathematical relationships of the Critical Process Parameters (CPPs) and Material Attributes (CMAs) with the Critical Quality Attributes (CQAs) define the design space. Consequently, process understanding is well assured and rationally leads to a final product meeting the Quality Target Product Profile (QTPP). This review illustrates the principles of quality theory through the work of major contributors, the evolution of the QbD approach and the statistical toolset for its implementation. As such, DoE is presented in detail since it represents the first choice for rational pharmaceutical development.

Application of Multiple Linear Regression and Artificial Neural Networks for the Prediction of the Packing and Capsule Filling Performance of Coated and Plain Pellets Differing in Density and Size

“…Design of experiments (DoE), supplemented with polynomial model fitting via multiple linear regression (MLR), is gaining acceptance as a prediction tool in pharmaceutical formulation work due to its simplicity, software availability, and the physical interpretation of the effects and interactions [8,9]. However, there are cases where high precision levels in conjunction with generalizing ability are required and MLR may not adequately satisfy these requirements [10].…”

Green Formulation of Triglyceride/Phospholipid‐Based Nanocarriers as a Novel Vehicle for Oral Coenzyme Q10 Delivery

“…Experimental factorial design has been recognized as an effective method for the estimation of correlations between parameters and performances as well as recognition of interactions among them. It will provide a strong support for rational selection of these parameters during study and shorten the time required for the preparation of efficient formulation (Dejaegher & Heyden, 2011;Politis, Colombo, Colombo, & Rekkas, 2017).…”

Experimental Planning Factorial: A brief Review

“…A factorial experimental design defines the imposition to make changes in the data production of synthesis in order to establish control of processes such as monitoring and control of processes with the purpose of detecting possible upgrades in the development and processes of research tests. Additionally, it assists in taken preventive decisions, becoming thus intended action to eliminate the possible causes of an impending shift of an unwanted situation in order to prevent these causes can come actually to occur (ANDERSON; WHITCOMB, 2010; FREUND; WILSON, MOHR, 2010aMOHR, , 2010b But the design of experiments (DOE) provides a conceptual interface through research that disrupt a phenomenon in order to understand their behavior andIt is a way to understand the process through the establishment of mathematical relationships from beginning to end of the process (POLITIS et al, 2017; According Kenett and Steinberg (2014), for an experiment to be considered statistically designed, some issues require determination (Table 1). The experimental design separates into two distinct phases: 1) screening phase (screening phase) and 2) modelingphase (phasemodeling) (BOXet al, 2005).. During the screening process, a group of variables receives a consideration stochastic based on their power levels shown from one or more response (MONTGOMERY, 2006).…”