Design Space Approach for Pharmaceutical Tablet Development

Abstract

Methodologies to determine the Design Space of a pharmaceutical product, within which continuous improvement can be implemented and postapproval changes in material attributes and process parameters can be introduced without prior approval, are presented. The type of methodology used depends on the type of experimental data obtained for the purpose of determining the Design Space. However, when one is unsure about the data collected to determine the Design Space, one can determine the quality of the data from successive application of the various methods and evaluation of the desirability measure for each method. The Design Spaces for a two-and three-process tablet manufacturing are determined by Response Surface Method (RSM), Bayesian Post Predictive Approach (BPPA), and Artificial Neural Networks (ANN), based on local or global specification limits of the response variables provided by multiresponse optimization and overlapping responses, respectively. The Response Surface Method is the most effective of these methods in determining the Design Spaces for the aforementioned data sets, confirming that the particular data are complete and lack uncertainty or structure, the specific features that the Bayesian Post Predictive Approach and Artificial Neural Networks methods are suited to address, respectively.

Definition of Design Spaces Using Mechanistic Models and Geometric Projections of Probability Maps

“…These constraints can be product related or process related and are either one or two sided inequality constraints that are easily represented in mathematical terms. A pivotal moment in the advancement of QbD was reached when practitioners recognized that it was possible to mathematically defined a design space as a subspace of the knowledge space, where the estimated probability for a system to fulfill all given constraints is greater than a minimum acceptable risk. − This can be written as follows: where DS, design space (a subset of x ); f , the model of the system; g , the set of constraints; h , the probability function for each element of x to fulfill the constraints ( g ); x , all possible combinations of process parameters in the n dimensional knowledge space (K n ); y , model predictions (presumably quality attributes of interest); π, minimum acceptable risk; σ , expected common cause variability; Θ , set of model parameters; and Σ Θ , variance covariance of model parameters…”

Narrow Operating Space Based on the Inversion of Latent Structures Model for Glycosylation Process

“…Rathore and Winkle used DOE method to identify the parameters for process characterization [15]. Chatzizacharia and Hatziavramidis did a comparative experiment in which several different methods were used to find the operating conditions for various data features [16]. Schoberer assess systematically the controllability of different types of specific glycoforms and glycans in glycosylation process through a statistical design of experiments scheme and Analysis of Variance (ANOVA) [17].…”

Optimization of a chromatographic process for the purification of saponins in Panax notoginseng extract using a design space approach

“…Recently, Quality by Design (QbD) and design space concept are gained increasing popularity in pharmaceutical manufacturing process, such as ethanol precipitation [17,18], powder blending [19], tablet manufacturing [20], pharmaceutical co-precipitation [21] and chromatographic process [22][23][24][25]. The application of http://dx.doi.org/10.1016/j.seppur.2015.09.035 1383-5866/Ó 2015 Elsevier B.V. All rights reserved.…”