A Pipeline for the Usage of the Core Data Set of the Medical Informatics Initiative for Process Mining - A Technical...

Area of Science:

• Health Informatics
• Process Mining
• Data Science

Background:

• Process Mining (PM) offers significant potential for enhancing healthcare processes and anomaly detection.
• Widespread PM adoption in healthcare is limited by unstructured event logs and data privacy concerns.

Purpose of the Study:

• To introduce a novel pipeline for converting routine healthcare data into process mining-compatible event logs.
• To enable the application of PM in healthcare settings by addressing data standardization and privacy issues.

Main Methods:

• The pipeline leverages Core Data Sets (CDS) from Data Integration Centers (DICs).
• Routine healthcare data is converted into Fast Healthcare Interoperable Resources (FHIR) and stored locally.
• Standardized PM event logs are generated through FHIR queries applicable across different DICs.

Main Results:

• The system successfully transforms diverse healthcare data into standardized PM event logs.
• Actionable insights can be extracted from various healthcare settings without modifying existing DIC infrastructures.
• The pipeline demonstrates applicability in complex healthcare systems, overcoming data variability and constraints.

Conclusions:

• Tailored event log generation and data querying are crucial for effective PM in healthcare.
• This approach facilitates evidence-based improvements in healthcare processes.
• The pipeline offers a standardized yet flexible solution for PM applications in healthcare.