(Cochran and Cox 1992; Lewis et al. 1998; Lewis 2002; Singh et al. 2005b;

Durakovic 2017). For screening as well as optimizing the factors, an array of

experimental runs, i.e. design matrix, is generated, and experimentation is done

accordingly. Coding of these factor levels is accomplished by designating them as

low (
1), intermediate (0) or high (+1). The nanoformulations are accordingly

fabricated as per the design matrix of the chosen experimental design and analysed

as per the standardized conditions determined for the formulations prepared earlier,

termed commonly as experimental runs (Cochran and Cox 1992; Singh et al. 2005b,

2011a). The entire process of relating CQAs with the factors, i.e. CMAs and/or

CPPs, for optimization is referred to as response surface methodology (RSM). Search

for an optimal solution is accomplished using mathematical (desirability function)

and/or graphical optimum (overlay plot) (Singh et al. 2005b, 2011a; Durakovic

2017).

18.4.4 Step IV: DoE Validation and Design Space Demarcation

Modelization using data fitting into linear, quadratic and/or cubic models is impera-

tive to obtain 3-D and 2-D plots to establish relationship(s) between the CQAs and

CMAs/CPPs (Singh et al. 2011b; Beg et al. 2017b). Like other studies in pharma-

ceutical technology, validation of FbD methodology is also necessary to ascertain

the applicability and prognostic capability of the model used.

Following this modelization and optimum search, a design space is demarcated as

a multidimensional amalgamation of the relationship(s) between various factors

(i.e. CMAs or CPPs) and the resultant response (i.e. CQA) (Araujo and Brereton

Fig. 18.7 An archetypal representation of 3-D response surface plot (left) and the corresponding

2-D contour plot (right) for any one response variable and two factors

324

B. Singh et al.