Design of dosage regimens: a multiple model stochastic control approach

Area of Science:

• Pharmacokinetics and Pharmacodynamics
• Stochastic Control Theory
• Computational Medicine

Background:

• Traditional drug dosage regimens often assume synchronous sampling and dosing.
• Asynchronous events in drug administration can lead to suboptimal therapeutic outcomes and increased variability.
• Existing control methods may not adequately address the complexities of real-time, asynchronous drug delivery.

Purpose of the Study:

• To develop a general stochastic control framework for asynchronous drug dosage regimens.
• To present an efficient computational method for optimal open-loop stochastic control under linear constraints.
• To demonstrate the efficacy of the proposed method in a clinical setting.

Main Methods:

• A general stochastic control framework was formulated for asynchronous sampling and dosing.
• The Multiple Models with Linear dynamics and Quadratic cost (MMLQ) approach was derived for optimal open-loop control.
• The MMLQ method was solved efficiently using quadratic programming.
• An adaptive MMLQ control strategy was implemented for Lidocaine infusion.

Main Results:

• The MMLQ approach provides an exact and efficient solution for optimal open-loop stochastic control with linear constraints.
• Implementation on a Lidocaine infusion process showed superior performance compared to a Maximum A Posteriori (MAP) Bayesian regimen.
• The MMLQ regimen significantly reduced interpatient variability.
• Patients managed with the MMLQ regimen were more consistently maintained within the therapeutic range.

Conclusions:

• The proposed stochastic control framework offers a flexible and effective method for designing drug dosage regimens.
• The MMLQ adaptive control approach is a powerful tool for optimizing drug delivery, particularly in asynchronous scenarios.
• This method demonstrates potential for improving patient outcomes by minimizing interpatient variability and enhancing therapeutic efficacy.