The self-aware diabetic patient software agent model

Area of Science:

• Biomedical Engineering
• Computational Biology
• Artificial Intelligence in Healthcare

Background:

• Diabetes management requires accurate modeling of blood glucose dynamics.
• Existing models often lack consideration for patient behavior and daily life variability.
• The seminal Ackerman model provides a foundation for blood glucose level simulation.

Purpose of the Study:

• To develop a 24-hour, stochastic, and self-aware diabetic patient software agent.
• To extend the Ackerman model to include night-time periods and patient self-awareness.
• To evaluate the impact of lifestyle management and self-monitoring on blood glucose control.

Main Methods:

• Extension of the Ackerman mathematical model for blood glucose levels.
• Incorporation of stochastic daily living effects and night-time dynamics.
• Development of a self-aware patient agent simulating patient behavior and condition awareness.

Main Results:

• Lifestyle management (diet, exercise, medication) reduced average blood glucose by up to 51%.
• Self-monitoring of blood glucose further decreased average blood glucose by 25%.
• Combined lifestyle management and self-monitoring achieved perfect blood glucose control within the target range.

Conclusions:

• The self-aware patient agent effectively represents diabetic patients and evaluates dynamic behaviors.
• The model demonstrates the significant impact of patient lifestyle and self-monitoring on glucose control.
• The agent can be integrated into multi-agent systems for broader healthcare insights (quality, cost, performance).