Abstract 190: A Systematic Alarm Management Pilot to Reduce Alarm Fatigue at a Tertiary Care Center

Abstract

Background: While alarms on hospital telemetry units are vital to patient safety, alarm fatigue is a well-established and potentially dangerous consequence. Our goal was to implement an alarm management pilot to systematically determine how to safely reduce alarm burden. Methods: The Brigham and Women’s Hospital Monitoring Steering Committee (BWH MSC) studied the effect of alarm fatigue from March to November 2014 in a multiweek, phased pilot (Figure 1). The intervention groups were two medical/surgical stepdown units (group 1) and three cardiac stepdown units (group 2). The control group was a combination of two medical/surgical units and one cardiac stepdown unit. Various interventions were performed in a stepwise and cumulative manner. Electrodes were switched on a daily basis. The high heart rate (HR) alarm threshold was increased from 120 to 130 beats per minute (bpm) and the low HR threshold was decreased from 50 to 45 bpm. Duplicate alarms (e.g. “tachy” alarm same as high HR parameter alarm) and non-actionable alarms (e.g. “couplets”) were switched to messages (non-audible). Alarm data was pulled and analyzed on a weekly basis. Results: At baseline, there was wide fluctuation in alarms monitored/week across all three groups; however, the peak alarm unit averaged roughly 400 audible, unique alarms/bed/day. Heart rate alarms decreased by 54% after phase 2 in group 1 and 48% after phase 2 in group 2 (Figure 1). Audible alarms were reduced in phase 3 and 4, most notably by making non-actionable alarms inaudible (34% decrease in phase 3 in group 1, 64% decrease in phase 4 in group 2). Electrode alarms appeared to be mildly reduced in phase 1 and 2 (4-24%), however, these findings were not consistent in phase 3 and 4. Conclusions: Alarm type and burden varied widely across time and hospital units, making it difficult to generalize interventions. However, based on the pilot results, the BWH MSC implemented the following changes across all telemetry units: 1) high HR limit and low HR limits were made more permissive as above, and 2) duplicate and non-actionable alarms were made non-audible. The latter intervention was thought to provide the greatest impact based on our experience. While this pilot represented a systematic approach to reduce alarm burden, ongoing efforts to safely improve alarm burden are warranted.