Detecting Severe Incidents from Electronic Medical Records Using Machine Learning Methods.

Area of Science:

• Health Informatics
• Patient Safety
• Clinical Data Mining

Background:

• Accurate incident reporting is crucial for patient safety.
• Many severe incidents go unreported, hindering quality improvement efforts.
• Existing methods for incident detection are often manual and time-consuming.

Purpose of the Study:

• To design and evaluate a solution for detecting non-reported incidents, particularly severe ones.
• To develop an automated method for extracting clinical notes related to severe incidents from electronic medical records.

Main Methods:

• Proposed a novel method for processing electronic medical records (EMRs).
• Developed an automated system to extract clinical notes.
• Focused on identifying descriptions of severe incidents within the notes.

Main Results:

• The implemented system successfully detected a non-reported incident.
• The incident was significant enough to be reported to the safety management department.
• The automated method proved effective in identifying critical events.

Conclusions:

• Automated processing of EMRs can effectively detect severe, non-reported patient safety incidents.
• This approach enhances patient safety by identifying events that might otherwise be missed.
• The developed system offers a valuable tool for proactive safety management in healthcare settings.