Summarizing clinical pathways from event logs.

Area of Science:

• Medical Informatics
• Health Informatics
• Clinical Data Analysis

Background:

• Clinical pathway analysis is crucial for standardized healthcare but often relies on aggregated data, limiting insights.
• Existing process mining techniques can yield overly detailed or numerous patterns, hindering comprehensive understanding.
• A need exists for methods that provide concise yet informative summaries of clinical pathways.

Purpose of the Study:

• To develop a novel approach for creating concise and comprehensive clinical pathway summaries.
• To reveal essential medical behaviors within specific time intervals of a clinical pathway.
• To improve the analysis and understanding of clinical pathway structures and practices.

Main Methods:

• Summarizing clinical pathways using event logs from hospital information systems.
• Formulating the clinical pathway summarization as an optimization problem solvable with dynamic programming.
• Segmenting the pathway's time period into continuous, overlapping intervals to discover frequent medical behaviors.

Main Results:

• The approach was evaluated using real-world data from Zhejiang Huzhou Central hospital for four diseases.
• Experimental results demonstrate the feasibility of constructing condensed clinical pathway summaries in polynomial time.
• The method exhibits linear scalability with increasing data log size and handles diverse medical behaviors.

Conclusions:

• The approach efficiently segments clinical pathways into time intervals and identifies critical medical behaviors.
• Generated summaries reveal both the global structure and interval-specific practices within clinical pathways.
• This method offers valuable insights for clinical pathway analysis and potential improvements in healthcare delivery.