Design and Optimization of Scored Tablets with Concave Surface and Application of Bayesian Estimation for Solving Scaleup Problem

From the viewpoint of self-medication, it is valuable to develop patient-friendly scored tablets that possess dividing uniformity. In this context, we attempted to optimize the preparation conditions for a tablet with a unique shape, such as a concavely curved scored tablet (CCST). Employing a design of experiment and a response surface method incorporating a thin-plate spline interpolation, and a bootstrap resampling technique, the optimal preparation conditions for CCST were successfully developed. To make it possible to scaleup the optimal solution estimated on a trial-scale, a Bayesian estimation was applied. Credible ranges of critical responses in large-scale manufacturing were estimated as a posterior probability from the trial-scale experiment as a prior probability. In terms of the large-scale manufacturing, the possibility of solving the scaleup problem was suggested using Bayesian estimation. Furthermore, a simulation study using a finite element method revealed that strong tensile stresses generated along the tip of the score line in CCST when an outer force was applied to the back surface of CCST. An advantage in dividing uniformity is indicated by the unique shape of CCST.