Characterization of Laboratory Flow and Performance for Process Improvements via Application of Process Mining

Area of Science:

• Clinical diagnostics
• Laboratory automation
• Data science

Background:

• Laboratory automation generates extensive event data, often underutilized for performance enhancement.
• Improving laboratory efficiency through data analysis is a key challenge.

Purpose of the Study:

• To develop a structured method for converting raw laboratory data into a process mining-suitable format.
• To apply process mining for mapping and characterizing sample flow in clinical chemistry labs to identify improvement areas.

Main Methods:

• Data extraction from instrument logs and middleware.
• Utilizing process mining for automated discovery and analysis of laboratory processes.
• Quantifying performance using key performance indicators (KPIs) like turnaround time and machine downtime.

Main Results:

• Identified opportunities to enhance laboratory efficiency by optimizing sample routes.
• Demonstrated the significant negative impact of machine downtime on laboratory performance.
• Highlighted the need to review sample pathways and maintenance policies.

Conclusions:

• This is the first application of process mining to medical diagnostic laboratory event data for automated model discovery.
• Process mining with KPIs offers valuable insights, encouraging data utilization and process intervention for performance goals.
• The vendor-independent approach is broadly applicable to medical diagnostic laboratories.