Developing a similarity searching module for patient safety event reporting system using semantic similarity measures

Area of Science:

• Healthcare
• Patient Safety
• Data Analysis

Background:

• Patient safety events (PSE) require effective reporting and analysis to prevent recurrence.
• Identifying patterns in PSEs is crucial for learning and improving care.
• Existing reporting systems lack dynamic similarity measurement capabilities.

Purpose of the Study:

• To develop a patient safety event (PSE) reporting system with dynamic similarity measurement.
• To improve the analysis and learning effect from PSEs.
• To integrate semantic similarity algorithms for enhanced PSE reporting.

Main Methods:

• Implemented three semantic similarity algorithms to analyze 366 PSEs.
• Evaluated algorithm performance using expert domain review on a Likert scale.
• Integrated algorithms into a self-learning and self-updating reporting system.

Main Results:

• Similarity scores demonstrated high consistency with expert evaluations.
• The reporting system incorporated a mechanism for learning and updating based on PSE similarity.
• Semantic similarity algorithms improved the identification of patterns among PSEs.

Conclusions:

• Integrating semantic similarity algorithms enhances PSE reporting systems.
• The system facilitates learning from past events and provides knowledge support.
• This approach promises to educate providers and prevent PSE consequences.