Sustainability of bundled approaches to alarm fatigue: Lessons from a long-term quality improvement study.

Abstract

Physiological monitor alarms are critical for detecting abnormalities but often cause alarm fatigue due to excessive, nonactionable events. This issue is pronounced in the pediatric setting, where movements and crying increase the alarm frequency. The long-term feasibility of countermeasures to reduce alarm fatigue remains underexplored. To evaluate the long-term outcomes of alarm countermeasures implemented through a quality improvement initiative. From January 2016 to February 2024, two intervention phases were implemented in nine wards of a tertiary pediatric center, with each phase being followed by a nonintervention period. The first intervention involved monthly data assessment, daily electrode replacement, and discussions about monitoring discontinuation criteria. The second intervention introduced patient-specific alarm thresholds and reminders to reconsider monitoring settings. Alarms were categorized as Advisory, Warning or Crisis. The primary outcome was daily alarm frequency per patient (ppd). A negative binomial regression was used for interrupted time series analysis. Overall, 46,302,740 alarms in 65,720 patients were analyzed. The alarm frequency declined from approximately 143 ppd to 99 ppd during the first intervention phase but rose to 124 ppd in the subsequent nonintervention phase. In the second intervention phase, the frequency declined to 86 ppd, the lowest value recorded, but rebounded to 165 ppd, the highest level, during the nonintervention phase. Advisory and Warning alarms showed a similar trend while Crisis alarms decreased consistently in the last phase. The interventions were associated with reduction in the alarm frequency but were unsustainable. Long-term success requires integrating countermeasures into a durable framework.