Application of a sensorial response model to the design of an oral liquid pharmaceutical dosage form

In this paper, we discuss the application of a compartmental model to study the sensorial response, in terms of taste intensity versus time, in an oral solution for pharmaceutical use. The numerical model was developed from sensorial response curves obtained by a panel of three trained individuals. Parameter identification was carried out by means of a least-squares procedure that obtained the linear coefficients in the model by solving an exact linear least-squares problem conditional on the values of the nonlinear parameters for each iteration. Thus, nonlinear estimation was done in terms of the first-order kinetic parameters only, and ill-conditioning of the Hessian matrix present in these models was solved. Results of modeling for a set of formulations were used to determine the effects of various ingredients (sweeteners and an essence) on a baseline unflavored formulation of acetaminophen in a mixture of cosolvents. The first moment of the area under the curve of taste intensity versus time was found to be the best global indicator of taste for the purpose of product design. It was found that a mixture of sweeteners and an essence was the most efficient way of masking the bitter taste of this active ingredient.