Multi-parameter vital sign database to assist in alarm optimization for general care units

Area of Science:

• Biomedical Engineering
• Clinical Informatics
• Health Services Research

Background:

• Continuous vital sign monitoring on general care floors aims for early patient deterioration detection and improved rapid response team effectiveness.
• Concerns exist regarding alarm fatigue due to continuous, multi-parameter vital sign monitoring systems.

Purpose of the Study:

• To develop a methodology for optimizing alarm settings in continuous vital sign monitoring to mitigate alarm fatigue.
• To create a simulation tool using a large patient data set to predict alarm burden.

Main Methods:

• Utilized an on-body wireless monitoring system to collect heart rate, respiratory rate, SpO2, and noninvasive blood pressure data.
• Compiled a cloud-hosted database of 94,575 hours of data from 3430 patients across ten hospitals.
• Conducted simulation scenarios to assess total alarm rates based on varying alarm thresholds and annunciation delays.

Main Results:

• Predicted a total alarm rate of ten alarms/patient/day from the cloud-hosted database simulations.
• Validated the prediction by comparing it to a 10-day evaluation in an independent hospital (1550 hours, 36 patients).
• Vital sign distribution plots from the database were consistent with previously published large-scale studies.

Conclusions:

• A cloud-hosted database enables simulations to predict total alarm burden by adjusting alarm thresholds and delays.
• This methodology offers a potential future solution to reduce alarm fatigue without compromising continuous vital sign monitoring.
• Optimizing alarm parameters can enhance the effectiveness of patient monitoring systems and clinical response.