Artificial intelligence for direct-to-physician reporting of ambulatory electrocardiography

Developments in ambulatory electrocardiogram (ECG) technology have led to vast amounts of ECG data that currently need to be interpreted by human technicians. Here we tested an artificial intelligence (AI) algorithm for direct-to-physician reporting of ambulatory ECGs. Beat-by-beat annotation of 14,606 individual ambulatory ECG recordings (mean duration = 14 ± 10 days) was performed by certified ECG technicians (n = 167) and an ensemble AI model, called DeepRhythmAI. To compare the performance of the AI model and the technicians, a random sample of 5,235 rhythm events identified by the AI model or by technicians, of which 2,236 events were identified as critical arrhythmias, was selected for annotation by one of 17 cardiologist consensus panels. The mean sensitivity of the AI model for the identification of critical arrhythmias was 98.6% (95% confidence interval (CI) = 97.7-99.4), as compared to 80.3% (95% CI = 77.3-83.3%) for the technicians. False-negative findings were observed in 3.2/1,000 patients for the AI model versus 44.3/1,000 patients for the technicians. Accordingly, the relative risk of a missed diagnosis was 14.1 (95% CI = 10.4-19.0) times higher for the technicians. However, a higher false-positive event rate was observed for the AI model (12 (interquartile range (IQR) = 6-74)/1,000 patient days) as compared to the technicians (5 (IQR = 2-153)/1,000 patient days). We conclude that the DeepRhythmAI model has excellent negative predictive value for critical arrhythmias, substantially reducing false-negative findings, but at a modest cost of increased false-positive findings. AI-only analysis to facilitate direct-to-physician reporting could potentially reduce costs and improve access to care and outcomes in patients who need ambulatory ECG monitoring.

Authors

Institutions

Publication Details

Journal
Nature Medicine
Published
2025-02-10
DOI
https://doi.org/10.1038/s41591-025-03516-x
Citations
28
Primary Topic
ECG Monitoring and Analysis
Type
article
Field-Weighted Citation Impact
38.87

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Artificial intelligence for direct-to-physician reporting of ambulatory electrocardiography

Anders Persson, Paweł T. Matusik, Martin Manninger, Alexandra Måneheim et al.
28 citations
Nature Medicine
ECG Monitoring and Analysis
38.87
article

Artificial intelligence for direct-to-physician reporting of ambulatory electrocardiography

Anders Persson, Paweł T. Matusik, Martin Manninger, Alexandra Måneheim, Philipp Krisai, Linda Johnson, Glauco Radoslovich, Sanjeev P. Bhavnani, Jeff S. Healey, Søren Zöga Diederichsen, Tigist Wodaje, Irina Mustafina, Zaza Iakobishvili, Artur Fedorowski, Gunnar Engström, Michael Rahkovich, Justinas Bacevičius, Steffen Blum, Per Hammarlund, Juan Benezet‐Mazuecos, Christian Lewinter, Ángel Miracle Blanco, Stefanos Zafeiropoulos, Christine Shen, Vytautas Juknevičius, Erik Ljungström, Jasmine Bisson, Gergana Marincheva, Raphael Spittler, Shahin Jamal, Axel Brandes, Pyotr G. Platonov, Martin Huth Ruwald, J. A. Marx, Richard G. Bennett, Alireza Oraii, Pascal Meyre, Einat Shaked, Rusne Jakaite, Ketil Haugan, A. Laish-Farkash, William F. McIntyre, Michiel Rienstra, Taya V. Glotzer, Alexander P. Benz, Dominik Linz, Stavros Stavrakis, Ada Sánchez Lozano, Jason G. Andrade, Evan S. Glazer, Emma Svennberg, Piotr Zadrozniak, Jakub Weclawski, Nick L. van Vreeswijk, Grzegorz Jasina, Agnieszka Grotek-Cuprjak, Yonatan Kogan
article en
28 citations

Abstract

Developments in ambulatory electrocardiogram (ECG) technology have led to vast amounts of ECG data that currently need to be interpreted by human technicians. Here we tested an artificial intelligence (AI) algorithm for direct-to-physician reporting of ambulatory ECGs. Beat-by-beat annotation of 14,606 individual ambulatory ECG recordings (mean duration = 14 ± 10 days) was performed by certified ECG technicians (n = 167) and an ensemble AI model, called DeepRhythmAI. To compare the performance of the AI model and the technicians, a random sample of 5,235 rhythm events identified by the AI model or by technicians, of which 2,236 events were identified as critical arrhythmias, was selected for annotation by one of 17 cardiologist consensus panels. The mean sensitivity of the AI model for the identification of critical arrhythmias was 98.6% (95% confidence interval (CI) = 97.7-99.4), as compared to 80.3% (95% CI = 77.3-83.3%) for the technicians. False-negative findings were observed in 3.2/1,000 patients for the AI model versus 44.3/1,000 patients for the technicians. Accordingly, the relative risk of a missed diagnosis was 14.1 (95% CI = 10.4-19.0) times higher for the technicians. However, a higher false-positive event rate was observed for the AI model (12 (interquartile range (IQR) = 6-74)/1,000 patient days) as compared to the technicians (5 (IQR = 2-153)/1,000 patient days). We conclude that the DeepRhythmAI model has excellent negative predictive value for critical arrhythmias, substantially reducing false-negative findings, but at a modest cost of increased false-positive findings. AI-only analysis to facilitate direct-to-physician reporting could potentially reduce costs and improve access to care and outcomes in patients who need ambulatory ECG monitoring.

Nature MedicineVol. 31(3)
University of Copenhagen (DK), Ben-Gurion University of the Negev (IL), Jagiellonian University (PL), Northwell Health (US), Hackensack University Medical Center (US), Feinstein Institute for Medical Research (US), University Medical Center Groningen (NL), Karolinska University Hospital (SE), University of British Columbia (CA), University of Groningen (NL), Vilnius University (LT), University of Southern Denmark (DK), University of Basel (CH), Lund University (SE), Johannes Gutenberg University Mainz (DE), Medical University of Graz (AT), Maastricht University Medical Centre (NL), University Hospital of Basel (CH), Copenhagen University Hospital (DK), Gentofte Hospital (DK), Rigshospitalet (DK), Vancouver General Hospital (CA), Population Health Research Institute (CA), Karolinska Institutet (SE), Zealand University Hospital (DK), Hospital La Luz (ES), University Medical Center of the Johannes Gutenberg University Mainz (DE), University Hospital of Zurich (CH), John Paul II Hospital (PL), Bashkir State Medical University (RU), Helsingborgs lasarett (SE), Hospital South West Jutland (DK), Institute of Cardiology (PL), Skåne University Hospital (SE), Centre Hospitalier de l’Université de Montréal (CA), Scripps Clinic (US), Center for Discovery (US), University of Oklahoma Health Sciences Center (US), Clalit Health Services (IL), University of Glasgow (GB), McMaster University (CA)
Lunds Universitet, Svenska Sällskapet för Medicinsk Forskning, Vetenskapsrådet
Openalex Percentile: Top 0%
ECG Monitoring and Analysis
38.87
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.