selecting Key Pharmaceutical Formulation Factors by Regression Analysis

Authors: N. R. Bohidar^a; F. A. Restaino^b; J. B. Schwartz^b

Abstract The use of Selective Regression Analysis to determine which formulation factors are governing product properties is demonstrated. The techniques of using and combining the two procedures called “All Possible Regression” (APR) and “Stepwise Regression” (SWR) are presented and applied to a multivariate pharmaceutical formulation problem. The technique was successfully applied to a system consisting of 10 response variables (tablet properties). Analysis of the results showed that for this formulation compression pressure and lubricant level exert the greatest effect on the majority of the properties. The results obtained from this method of analysis can aid the formulator in selectively controlling the product properties of choice while leaving the others undisturbed. Selective Regression Analysis also provides a basis for understanding the underlying mechanism of the system under consideration.