Multi-level process mining methodology for exploring disease-specific care processes

Area of Science:

• Health Informatics
• Process Mining
• Healthcare Systems Analysis

Background:

• Traditional process mining tools struggle with the complexity and variability of public healthcare protocols.
• Specialized methods are needed to effectively analyze intricate healthcare processes and patient journeys.

Purpose of the Study:

• To develop a universally applicable process mining methodology for exploring disease-specific care pathways.
• The goal is to enhance the analysis of healthcare processes for better insights and system improvements.

Main Methods:

• The proposed Process Mining Methodology for Exploring Disease-specific Care Processes (MEDCP) uses a multi-level expert taxonomy for systematic event log refinement.
• Domain-specific rules define treatment processes, identifying key start and end events.
• Taxonomy concepts support process definition for meaningful expert conclusions.

Main Results:

• MEDCP was validated in cardiology and oncology care studies in Hungary over ten years.
• The methodology identified high-level event patterns and system anomalies, like provider volume differences in oncology.
• Cardiology study revealed geographically clustered providers with similar care spectra.

Conclusions:

• MEDCP enhances standard process mining, providing interpretable, high-level conclusions for domain experts.
• The methodology offers system-level insights crucial for optimizing and planning healthcare delivery.
• This approach facilitates evidence-based decision-making for healthcare system efficiency.