OBJECTIVE Rotational
central composite design was employed to optimize the
overall desirability of multiple variables.METHODS
Microspheres were prepared by an o/w solvent evaporation/extraction method. Independent variables of the experimental design were polyvinylalcohol concentration in water phase(X 1),poly(d,llactide) concentration in inner phase(X 2) and theoretical drug loading(X 3). Dependent variables or responses, such as loading, loading efficiency, geometric mean diameter, span of dispersity, release percentage in day 1 and 85% release time, were transformed into desirabilities mathematically by Hassan's or Harrington's method.
Overall desirability was calculated from the geometric mean of the six desirabilities of each formulation. Data of overall desirabilities were fitted to a secondorder polynomial equation, through which three dimensional response surface graphs were produced. Optimum experimental conditions were selected from the stationary point of the response surfaces.RESULTS The regression coefficient of the nonlinear estimation was higher than 0.85. Microspheres prepared according to the optimum experimental conditions resulted in an overall desirability of 0.5004. The bias between observed and predicted values of the overall desirability of the optimum formulation was -2.09%.CONCLUSION Overall desirability can be used to represent the most desirable variables in experimental design.