A prognostic model for temporal courses that combines temporal abstraction and case-based reasoning.

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Clinical Decision Support

Background:

• Clinical management and research are inherently time-dependent, making temporal reasoning crucial in medical informatics.
• Existing methods often rely on established standards, periodicity, or complete domain theories, which are not always available.

Purpose of the Study:

• To present a novel method for prognosis of temporal courses by integrating temporal abstractions with case-based reasoning.
• To address application domains lacking standardized protocols, known periodicity, or comprehensive domain knowledge.

Main Methods:

• Developed a hybrid approach combining temporal abstractions and case-based reasoning.
• Applied the method to two distinct prognostic applications: kidney function in intensive care and geographical medicine for infectious disease forecasting.

Main Results:

• Demonstrated utility in predicting kidney function impairments to warn against potential kidney failures.
• Showcased application in geographical medicine for early warnings of infectious diseases with irregular cycles, focusing on influenza forecasting with initial experimental results.

Conclusions:

• The proposed method offers a flexible approach to temporal prognosis in complex medical domains.
• It shows promise for early detection of critical health events and disease outbreaks, particularly in data-scarce or non-standardized environments.