Improving Automatic Smartwatch Electrocardiogram Diagnosis of Atrial Fibrillation by Identifying Regularity within Irregularity

Scientific News

Velraeds, A.; Strik, M.; van der Zande, J.; Fontagne, L.; Haissaguerre, M.; Ploux, S.; Wang, Y.; Bordachar, P. Improving Automatic Smartwatch Electrocardiogram Diagnosis of Atrial Fibrillation by Identifying Regularity within Irregularity. Sensors 202323, 9283.

One of the most common cardiac arrhythmias is atrial fibrillation (AF), resulting in an entirely irregular ventricular rhythm on the electrocardiogram (ECG) due to a complete loss of organized atrial contractility with signals redirected intermittently to the ventricles. In their 2023 article, Anouk Velraeds and colleagues present a novel algorithm that automatically detects AF in a large population. They underscore the crucial role of a second step in the analysis for improved AF detection. Their study is a call to action, highlighting the urgent need to develop an improved automatic AF detection algorithm for patients with or without coexisting ECG abnormalities. The authors propose a promising two-step approach: first, identify any irregular rhythm and then exclude ECGs that show patterns of regularity. This approach holds the potential to be more inclusive and significantly reduce the number of missed AF cases of false positives, thereby improving AF detection.


Atrial fibrillation (AF) is one of the most common cardiac arrhythmias, characterized by an entirely irregular ventricular rhythm due to a complete loss of organized atrial contractility. This condition leads to significant health risks, including stroke and heart failure, if not detected and managed promptly. In 2023, Anouk Velraeds and colleagues presented a novel algorithm designed to improve the automatic detection of AF using smartwatch electrocardiograms (ECG). This article delves into the innovative approach of identifying regularity within irregular rhythms to enhance diagnostic accuracy.

The Need for Improved AF Detection

Early detection of atrial fibrillation is crucial for preventing severe complications. Wearable technology, particularly smartwatches with ECG capabilities, has emerged as a valuable tool for continuous health monitoring. However, existing algorithms often struggle with accuracy, leading to false positives and missed diagnoses. Velraeds and colleagues highlight the urgent need for a more reliable automatic AF detection algorithm that can be applied to a diverse patient population. In an article published in the review Sensors Anouk Velraeds and colleagues (2023) reported on improving Automatic Smartwatch Electrocardiogram Diagnosis of Atrial Fibrillation by Identifying Regularity within Irregularity. Atrial fibrillation (AF) is a common cardiac arrhythmia that leads to irregular ventricular rhythms on an electrocardiogram.

The Two-Step Algorithm: Identifying Regularity within Irregularity

The proposed algorithm employs a two-step approach:

  1. Initial Detection of Irregular Rhythms: The first step involves identifying any irregular rhythms in the ECG data.
  2. Exclusion of Regular Patterns: The second step focuses on excluding ECGs that exhibit patterns of regularity within the identified irregular rhythms. This method reduces false positives and improves the accuracy of AF detection.

Methodology and Patient Population

The study involved a diverse group of 723 individuals, encompassing various ages and health conditions. Patients were selected from different healthcare settings to ensure a representative sample. The algorithm was tested on this dataset, demonstrating superior performance compared to existing algorithms.

Key Findings and Comparative Analysis

The algorithm’s effectiveness lies in its ability to cluster ECG data points and identify regular patterns within these clusters. This process significantly enhances diagnostic accuracy, outperforming traditional methods. Compared to previous studies, including those involving popular smartwatches like the Apple Watch, the novel algorithm shows promise in reducing missed AF cases and false positives.

Challenges and Limitations

Despite its potential, the study acknowledges several limitations:

  • Single ECG Per Patient: Recording only one ECG per patient can lead to multiple diagnoses with repeated tests, affecting user confidence.
  • Dataset Representativeness: The dataset was not entirely representative of real-world conditions, resulting in a higher prevalence of ECG anomalies.
  • Specific Patient Groups: The algorithm faced difficulties with RR wave detection in patients with pacemakers and cardiac resynchronization therapy (CRT), where continuous monitoring is already in place.

Future Research Directions

The authors express optimism about the future of AF detection using this innovative approach. Further research will focus on refining the algorithm, particularly for patients with pacemakers and CRT. Additionally, integrating data clustering techniques and identifying regular patterns within irregular rhythms hold promise for broader applications in cardiac health monitoring.


The development of this novel algorithm represents a significant advancement in the field of cardiac health. By improving the automatic detection of atrial fibrillation through the identification of regularity within irregular rhythms, this approach offers a more accurate and reliable diagnostic tool. As wearable technology continues to evolve, the potential for early detection and intervention in cardiac arrhythmias will only grow, ultimately improving patient outcomes and quality of life.

Was this article helpful?