Automated Fall Detection Algorithm With Global Trigger Tool, Incident Reports, Manual Chart Review, and...

Area of Science:

• Medical Informatics
• Patient Safety
• Health Informatics

Background:

• Hospital falls are frequent adverse events with significant cost implications.
• Current fall detection methods (incident reports, chart reviews) are time-consuming and error-prone.
• Electronic health record (EHR) data offers a potential for efficient and accurate fall detection.

Purpose of the Study:

• To develop and validate an EHR-based algorithm for detecting in-hospital falls.
• To compare the algorithm's accuracy against established methods like the Global Trigger Tool, incident reports, manual chart review, and patient-reported falls.

Main Methods:

• A retrospective diagnostic accuracy study was conducted in a Swiss hospital system.
• Two substudies involved algorithm development (240 patients) and validation (298 patients).
• Sensitivity, specificity, and predictive values were calculated to compare detection methods.

Main Results:

• The EHR algorithm achieved 95% sensitivity in development and 100% in validation.
• Compared to other methods, the algorithm demonstrated superior fall detection rates.
• The algorithm requires seconds per case, significantly reducing review time compared to manual methods.

Conclusions:

• The developed EHR algorithm shows high sensitivity for detecting in-hospital falls.
• This near real-time tool can enhance the development and testing of fall prevention strategies.
• The algorithm provides a cost-effective and efficient method for fall surveillance in healthcare settings.