Several improvements on the target interval stochastic control (TISC) method are addressed for individualizing therapy. In particular, a global optimization control strategy is implemented to obtain the optimal dosage regimen, and weighting functions are introduced to balance the drug efficacy and the risk of toxicity. Since general guidance is often lacking in the determination of a weighting function, we introduce a systematic approach, i.e., the standard reference gamble method of medical decision theory, for the determination of the weighting function. The population model for the individualization of theophylline therapy reported by D’Argenio and Katz is applied in this research. The present method of the integration of weighting functions and global optimal strategy offer an effective and safe means to balance the drug efficacy and risk of toxicity. In addition, it also achieves better accuracy than the existing TISC method which uses a local optimal strategy.
Target interval stochastic control Weighting function Bayesian estimation Dosage regimen Global optimization