Dynamically generated models for medical decision support systems

Area of Science:

• * Computational modeling
• * Biomedical engineering
• * Respiratory physiology

Background:

• * Mechanical ventilation requires balancing patient benefit and risk.
• * Mathematical models can simulate patient responses to ventilation changes.
• * Existing models may lack the complexity to capture intricate physiological interactions.

Purpose of the Study:

• * To introduce a framework for dynamically combining diverse mathematical models.
• * To create a complex, interacting model system for ventilation simulation.
• * To investigate the impact of model complexity on simulation outcomes.

Main Methods:

• * Developed a framework to integrate submodels from different families.
• * Submodels vary in dynamic formulation complexity and anatomical resolution.
• * Simulated patient responses using the combined, interacting model system.

Main Results:

• * Model system interactions yield qualitatively different results based on complexity.
• * Detected realistic overlaying of respiratory and cardiovascular rhythms.
• * Observed these interactions in simulated blood gas concentrations.

Conclusions:

• * The proposed framework enables sophisticated simulation of mechanical ventilation.
• * Model complexity is a critical factor influencing simulation accuracy and insights.
• * This approach can reveal complex physiological dynamics, aiding clinical decision-making.