A declarative approach for interactive process discovery in the clinical domain.

Area of Science:

• Health Informatics
• Process Mining
• Clinical Informatics

Background:

• Process Mining (PM) is increasingly used in healthcare to analyze Electronic Health Record data.
• Clinical processes derived from PM often yield complex, difficult-to-interpret networks.
• Existing PM methods struggle to align with clinical realities and algorithms.

Purpose of the Study:

• To integrate specialized clinical knowledge into Process Mining discovery algorithms.
• To develop an interactive approach for clinical process discovery.
• To improve the comprehensibility and clinical relevance of discovered process models.

Main Methods:

• A declarative approach using the Declare formalism is proposed.
• Clinicians can interactively incorporate their expertise into the PM algorithm.
• The approach is implemented in the I-PALIA PM discovery algorithm.

Main Results:

• The declarative interactive approach was successfully implemented.
• A use case involving prostate cancer treatment demonstrated the approach's utility.
• The method effectively addresses common challenges in clinical PM.

Conclusions:

• A novel interactive process discovery approach for the clinical domain is presented.
• This method enables clinical experts to inject specialized knowledge into PM.
• The resulting process models are more understandable and accurately reflect treatment pathways.