MitPlan: A planning approach to mitigating concurrently applied clinical practice guidelines.

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Computational Medicine

Background:

• Aging populations lead to increased patient complexity and multi-morbidity.
• Current clinical practice guidelines (CPGs) address single diseases, creating challenges for multi-morbid patients.
• Existing frameworks for mitigating adverse interactions in multi-morbid patients rely on complex logic and theorem proving.

Purpose of the Study:

• To simplify the process of identifying and addressing adverse interactions in multi-morbid patients.
• To develop a novel framework (MitPlan) for managing complex patient care using planning techniques.
• To optimize patient management plans based on physician-defined horizons, patient preferences, and costs.

Main Methods:

• Representing clinical practice guidelines and medical knowledge as a planning problem using the Planning Domain Definition Language (PDDL).
• Utilizing durative planning actions to model clinical actions like medication administration and patient testing.
• Implementing optimization criteria including total treatment cost and revision costs.

Main Results:

• The MitPlan framework successfully identifies and addresses adverse interactions in multi-morbid patients.
• MitPlan generates interpretable sequences of timed actions for patient management plans.
• Demonstrated capabilities through illustrative and clinical case studies.

Conclusions:

• MitPlan offers a simplified and effective approach to managing complex, multi-morbid patients.
• The framework facilitates the creation of optimized and interpretable patient management plans.
• This approach has the potential to improve care for aging populations with multiple chronic conditions.