Smartwatches can now detect cardiac arrest

The world of wearable technology continues to evolve, bringing new possibilities for health monitoring. What started with step counting and heart rate tracking has now expanded to life-saving features. One of the most significant advancements is the ability of smartwatches to detect cardiac arrest.

A recent study led by Google Research has demonstrated that a machine learning algorithm can identify sudden pulse loss with remarkable accuracy.

The system, designed to recognize cardiac arrest events, can automatically call emergency services when it detects pulselessness. This feature could be life-changing, especially for individuals who experience cardiac arrest without anyone around to help.

Early detection and quick medical intervention are key to survival, making this innovation a potential game-changer in emergency response.

Challenge of unwitnessed cardiac arrest

“Wearable technology has the potential to revolutionize emergency response,” noted Google Research Team.

Out-of-hospital cardiac arrest (OHCA) is a major cause of sudden death worldwide. In many cases, survival depends on immediate recognition and rapid medical response.

Unfortunately, a significant number of these incidents occur without witnesses, making timely intervention nearly impossible. Studies show that 50–75% of OHCA cases happen with no one around to call for help.

This delay in emergency response greatly reduces the chances of survival. Without CPR or defibrillation within minutes, brain damage can occur, and the likelihood of recovery diminishes rapidly.

Researchers aimed to solve this problem by creating a system that allows smartwatches to detect pulselessness automatically. By doing so, they hope to improve survival rates for those who suffer cardiac arrest when no one is there to assist.

Smartwatches detect cardiac arrest

To develop a reliable system, researchers trained an algorithm using photoplethysmography (PPG) and motion data. The study involved extensive testing across six distinct groups, ensuring the algorithm worked in both controlled and real-world environments.

The first phase of testing took place in a clinical setting. In an electrophysiology lab, 100 patients undergoing defibrillator testing experienced induced ventricular fibrillation.

This allowed researchers to gather data on pulseless states. Another 99 participants experienced pulselessness through a tourniquet-induced arterial occlusion model, further refining the algorithm’s ability to detect a loss of pulse.

In addition to clinical trials, the researchers needed to evaluate how the system would perform in everyday life.

To accomplish this, they collected data from 948 users who wore the smartwatch under normal conditions without experiencing cardiac arrest. This helped the team measure false alarms and refine the algorithm’s accuracy.

Testing in real-world conditions

Once the initial data was collected, further testing was needed to assess the system’s reliability in different scenarios. To determine how often false positives occurred, 220 participants wore the smartwatch in their daily routines.

This group helped establish how frequently the algorithm might mistakenly detect a loss of pulse when no real emergency was happening.

To further validate the system, 135 participants underwent testing in both free-living conditions and controlled settings. These individuals experienced real-world activity while also being exposed to controlled pulseless events through the tourniquet-induced arterial occlusion method. This allowed researchers to compare results across different environments.

A unique part of the study involved professional stunt performers. Twenty-one trained stunt persons simulated out-of-hospital cardiac arrest collapses, helping researchers understand how well the algorithm could detect pulse loss during sudden, high-motion events.

Since real-world collapses involve unpredictable movements, this test was essential to ensuring the smartwatch could still detect cardiac arrest in dynamic situations.

Detecting cardiac arrest with 99.99% accuracy

The study produced promising results, showing that the smartwatch algorithm could accurately detect pulselessness while keeping false alarms to a minimum.

The researchers found no significant differences between PPG signals from ventricular fibrillation and those from arterial occlusion-induced pulselessness.

The system achieved a 72% sensitivity rate for motionless pulseless events, meaning it correctly identified cardiac arrest in nearly three out of four cases. Sensitivity for simulated collapses was slightly lower at 53%, reflecting the challenge of detecting pulselessness during high-motion scenarios.

However, the algorithm’s specificity reached an impressive 99.99%, meaning it generated very few false alarms. The study estimated one false emergency call per 21.67 user-years, a promising figure for real-world application.

When detecting a loss of pulse, the system identified pulselessness within 57 seconds. It then waited an additional 20 seconds to check for user response before initiating an emergency call.

This short response time could make a critical difference in survival rates, ensuring help arrives as quickly as possible.

Future of wearable emergency detection

Smartwatches equipped with cardiac arrest detection have the potential to save countless lives.

For individuals who suffer cardiac arrest alone, this technology could be the difference between life and death. By ensuring immediate medical response, smartwatches may soon become essential tools in emergency healthcare.

Despite its high specificity, reducing false positive rates remains a crucial next step. While the algorithm performed well in controlled settings, real-world scenarios can be unpredictable. The researchers acknowledge that more refinements are necessary to enhance accuracy in everyday conditions.

One challenge lies in distinguishing between true pulseless events and temporary pulse irregularities caused by movement, stress, or sensor issues. Continuous data collection from smartwatch users may help improve the algorithm, allowing it to adapt to different conditions more effectively.

Smartwatch-based health monitoring

As wearable technology advances, the line between fitness tracking and life-saving medical devices continues to blur. Smartwatches are no longer just tools for step counting and heart rate monitoring – they are evolving into critical health companions.

The ability to detect cardiac arrest autonomously represents a major leap forward in healthcare innovation.

By combining machine learning with real-time health data, researchers are pushing the boundaries of what wearable devices can achieve. The goal is not just to monitor health but to actively respond to life-threatening emergencies.

With further refinements and broader adoption, smartwatch-based emergency detection could become a standard feature, offering peace of mind to millions.

The study is published in the journal Nature.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–