Turning down the noise: Smarter alarms for safer hospitals | News

Anyone who has spent more than a few minutes in a hospital has heard the cacophony of persistent alarms. Those alarms are also the cause of a persistent problem in healthcare settings: alarm fatigue. 

Hospital monitors are designed to beep when a patient’s vital signs move outside a set range, not necessarily when the patient actually needs attention. As a result, up to 99% of these alerts turn out to be false alarms. Over time, nurses become desensitized to the noise, which can lead to slower response times and an increased risk to patients when the alarm is genuine. 

New research from Dr. Hossein Piri at the University of Calgary aims to turn false alarms into smart alarms. Funded by the Natural Sciences and Engineering Research Council (NSERC), Piri is focused on improving the way bedside monitors alert nurses to changes in a patient’s condition, with the goal of making alarms smarter, more personalized and more effective.

The idea originated from direct observations when Piri spent time working with clinicians in intensive care units (ICUs). The assistant professor at UCalgary’s Haskayne School of Business observed how clinicians often face a constant stream of alerts and says, “I saw firsthand how overwhelming and distracting the current alarm systems are.” 

He also noticed how many of these alarms are triggered by temporary or harmless changes in a patient’s condition, yet they demand attention and interrupt care.

“I saw firsthand how overwhelming and distracting the current alarm systems are.”

Piri’s curiosity led him beyond the ICU, where he found that in busy pharmacies, approximately 500 prescriptions are processed daily. Each prescription can trigger up to 10 alerts warning of potential drug interactions, dosage errors, or allergy risks. 

Managing hundreds of these alerts daily, pharmacists become overwhelmed by the sheer volume and start unintentionally ignoring these alerts. Most of them, like hospital alarms, are unnecessary, but for the few that aren’t, can be a matter of life or death.

Piri shared these insights at the Haskayne Business Exchange in March, 2024. You can watch his full presentation here or catch the highlights in this aminated short. 

This newly funded project builds on those findings and focuses on developing an intelligent alarm system that can predict which patients are truly at risk and adjust alarm thresholds based on individual health data — not a pre-determined impersonal range. 

By using tools from operations research and machine learning, the system would learn to distinguish between minor fluctuations and serious warning signs.

For Piri, the goal is to reduce the number of false alarms so nurses can focus on patients who genuinely need urgent care. This could help lower stress levels among staff, improve patient outcomes and make hospital environments more efficient. 

The path forward, Piri says, is through smart, data-driven tools. 

“What excites me most is that this work sits at the intersection of operations research, machine learning, and healthcare, and has real potential to reduce nurse burnout, improve patient safety, and save lives.”

Timing of research especially relevant

Healthcare systems across Canada are under pressure, with high rates of nurse burnout and staffing shortages. Smarter alarm systems could help ease some of that burden by reducing unnecessary distractions and allowing nurses to prioritize their time more effectively.

While Piri’s work recognizes the serious risks alarm fatigue in health care workers poses to patients, it is not a reproach. It is a recognition that clinicians and nurses, like the rest of us, are human. 

The concept of alarm fatigue extends beyond the realm of healthcare. In everyday life, people are bombarded with digital alerts. 

Email notifications, app reminders and calendar alerts compete for attention. Just like nurses in a hospital, many of us have learned to ignore these signals, sometimes at the risk of missing something important.

Piri’s research highlights the need for more thoughtful and personalized alert systems, not just in healthcare but across digital environments. By designing systems that understand context and user behaviour, alerts can become more meaningful and less overwhelming.

While the project is still in development, early results suggest that data-driven alarm systems could significantly reduce false alerts. 

Piri and his team are working closely with clinicians to test and refine the model in real-world settings.

“This work sits at the intersection of operations research, machine learning, and healthcare, and has real potential to reduce nurse burnout, improve patient safety, and save lives.”

The long-term vision is to create hospital alarm systems that support, rather than hinder, patient care. 

By combining clinical insight with advanced analytics, the research offers a promising path toward safer, quieter and more responsive healthcare environments.

Ultimately, smarter alarms could help nurses do what they do best: care for patients without unnecessary distractions.