Detecting False Alarms by Analyzing Alarm-Context Information: Algorithm Development and Validation

Area of Science:

• Biomedical Engineering
• Clinical Informatics

Background:

• Clinical alarm systems are critical but often generate excessive false alarms (80%-99%).
• This leads to alarm fatigue, causing healthcare providers to miss critical alerts, potentially harming patients.
• Hospitals lack standardized protocols for alarm management, exacerbating the issue.

Purpose of the Study:

• To propose a solution for mitigating alarm fatigue using an automatic reasoning mechanism.
• To develop a system that calculates the false alarm probability (FAP) for clinical alarms.
• To provide a FAP label with notifications to help healthcare teams prioritize alerts.

Main Methods:

• Developed an automatic reasoning approach to manage clinical alarm notifications.
• Implemented a reasoning algorithm to calculate FAP for sensor and multiparametric monitor alerts.
• Utilized statistical analysis of false alarm indicators (FAIs) within a simulated intensive care unit (ICU) environment.

Main Results:

• Defined a list of reusable false alarm indicators (FAIs) for researchers.
• Introduced a novel statistical method to assess FAP using multi-input alarm context.
• Created a reasoning algorithm that identifies false alarms and provides FAP labels to reduce alarm fatigue.

Conclusions:

• An intelligent notification system can effectively identify false alarms by analyzing alarm context.
• The developed reasoning system successfully attributed FAP values and provided FAP labels.
• The system notified caregivers without compromising patient safety, offering a viable solution to alarm fatigue.