Leveraging Synthetic Patient Data for Process Mining in Healthcare

Area of Science:

• Health Informatics
• Data Science
• Process Mining

Background:

• Process mining in healthcare faces challenges due to data privacy and restricted access to real-world patient data.
• Optimizing workflows and ensuring guideline compliance are key goals for healthcare process mining.

Purpose of the Study:

• To develop a methodological infrastructure for healthcare process mining using synthetic patient data.
• To evaluate the feasibility and accuracy of reconstructing healthcare processes from synthetic data.

Main Methods:

• Generated synthetic patient data using Synthea.
• Integrated HL7 FHIR AuditEvent Resources to create standardized audit logs.
• Converted audit logs to process mining formats (XES, OCEL) for analysis.

Main Results:

• Demonstrated accurate process reconstruction from synthetic healthcare data.
• Evaluated the influence of data volume on process mining accuracy.
• Identified challenges including missing events and limitations of FHIR R4 AuditEvent.

Conclusions:

• Synthetic patient data, when structured with FHIR AuditEvent Resources, can effectively support healthcare process mining.
• The proposed methodology shows promise for overcoming data privacy issues in healthcare process analysis.
• Further validation in real-world scenarios is needed to confirm practical effectiveness.