DESIGN OF DOSAGE REGIMENS - A MULTIPLE MODEL STOCHASTIC-CONTROL APPROACH

This paper presents a general stochastic control framework for determi ning drug dosage regimens where the sample times, dosing times, desire d goals, etc., occur at different times and in an asynchronous fashion . In the special case of multiple models with linear dynamics and quad ratic cost (MMLQ), it is shown that the optimal open-leap stochastic c ontrol with linear control/state constraints can be solved exactly and efficiently as a quadratic program. This provides a simple and flexib le method for computing open-loop feedback designs of drug dosage regi mens. An implementation of the MMLQ adaptive control approach is demon strated on a Lidocaine infusion process. For this example, the resulti ng MMLQ regimen is more effective than the MAP Bayesian regimen at red ucing interpatient variability and keeping patients in the therapeutic range.