Experimental studies on multiple-model predictive control for automated regulation of hemodynamic variables

Area of Science:

• Critical care medicine
• Biomedical engineering
• Control systems engineering

Background:

• Hemodynamic instability is a major challenge in critical care.
• Manual regulation of mean arterial pressure (MAP) and cardiac output (CO) is complex and prone to variability.
• Inotropic and vasoactive drugs are essential for managing hemodynamic instability.

Purpose of the Study:

• To develop and evaluate an automated, model-based control methodology for regulating MAP and CO in critical care.
• To compare the performance of the automated controller against manual regulation.
• To assess the controller's ability to handle drug rate constraints and patient variability.

Main Methods:

• A multiple-model adaptive approach within a model predictive control (MPC) framework was employed.
• The control algorithm integrated inotropic and vasoactive drug administration.
• Experimental validation was conducted on canines with induced hypertension and depressed cardiac output.

Main Results:

• The model-based controller significantly improved MAP regulation (88.9% within +/- 5 mm Hg) compared to manual control (82.3%).
• Cardiac output was more precisely controlled (96.1% within +/- 1 L/min) by the automated system versus manual methods (92.2%).
• The controller effectively managed drug rate constraints and patient variability, demonstrating superior hemodynamic stability.

Conclusions:

• Model-based, adaptive MPC offers a robust solution for automated hemodynamic management in critical care.
• This automated control strategy enhances the precision and reliability of MAP and CO regulation.
• The developed methodology shows significant potential for improving patient outcomes in critical care settings.