Skip to content
DraconiticResearch
Search Sources API
203,203 studies31 sourcesUpdated 4 min ago
Peer-reviewedCC BY-NC-ND 4.0NPJ Digital Medicine

Impact of an AI algorithm for multi-day prediction of incident atrial fibrillation on clinical decision-making: PROVISION-AF study

  • Yeji Kim1
  • Bogeun Kim2
  • Jaeung Yoon2
  • Jeongmin Kang1
  • Seung Woo Lee1
  • Moon-Hyun Kim1
  • Dong-Hyeok Kim1
  • Hye Ah Lee3
  • Jaemin Shim4
  • Dae-In Lee5
  • Seung-Young Roh5
  • Seung Young Shin6
10 more authors
  • Sung Hwa Choi7
  • Yae Min Park7
  • Ki Hong Lee8
  • Yoo Ri Kim8
  • Min Kim9
  • Wei-Jin Park10
  • Hyeon Yeo11
  • Jungwon Choi11
  • Tae Young Shin2
  • Junbeom Park1
11 affiliations
  1. 1Department of Cardiology, College of Medicine, Ewha Womans University, Seoul, Republic of Korea
  2. 2Department of Urology, College of Medicine, Ewha Womans University Medical Center, SYNERGY AI, Ewha Womans University Mokdong Hospital, Seoul, Republic of Korea
  3. 3Clinical Trial Center, Ewha Womans University Medical Center, Seoul, Republic of Korea
  4. 4Division of Cardiology, Department of Internal Medicine, Korea University College of Medicine and Korea University Anam Hospital, Seoul, Republic of Korea
  5. 5Division of Cardiology, Department of Internal Medicine, Korea University College of Medicine and Korea University Guro Hospital, Seoul, Republic of Korea
  6. 6Division of Cardiology, Department of Internal Medicine, Korea University College of Medicine and Korea University Ansan Hospital, Seoul, Ansan, Republic of Korea
  7. 7Division of Cardiology, Department of Internal Medicine, Gachon University Gil Medical Center, Incheon, Republic of Korea
  8. 8Division of Cardiology, Department of Internal Medicine, Chonnam National University Hospital, Gwangju, Republic of Korea
  9. 9Division of Cardiology, Department of Internal Medicine, Kangbuk Samsung Hospital, Seoul, Republic of Korea
  10. 10Acryl 11F, Seoul, Korea
  11. 11KAIST (Korea Advanced Institute of Science and Technology), Daejeon, Korea
Read full textPDFDOICite

Abstract

Atrial fibrillation (AF) is frequently asymptomatic and often remains undetected until complications arise. Although artificial intelligence (AI)-enabled electrocardiography (ECG) can predict incident AF from sinus rhythm ECGs, its influence on physician risk assessment in simulated clinical settings remains uncertain. We developed a deep learning model to predict multi-day AF risk using non-AF 12-lead ECGs. For ECG-labeled outcomes, the model achieved an AUROC of 0.79 in the internal EUMC cohort and 0.74 in the external BIDMC cohort. For Holter-labeled outcomes, AUROC values reached 0.87 in the internal EUMC subset and 0.75 in the prospective PROVISION-AF cohort. To assess decision-support utility, a multinational survey of 70 physicians evaluated how AI-derived risk estimates influenced physician risk assessment and follow-up decisions in structured simulated cases. AI assistance significantly improved physicians’ AF risk discrimination (AUROC 0.573 to 0.650) and negative predictive value (0.764 to 0.839), with significant net reclassification improvement for non-AF cases (NRI 0.127, p < 0.001). Performance gains were most notable among non-electrophysiologist cardiologists. In conclusion, AI-derived risk estimates improved physician risk discrimination in a structured simulated survey, particularly in non-specialist settings, supporting their potential role as a digital decision-support tool. Further real-world implementation studies are needed to determine whether these effects translate into improved clinical outcomes or healthcare efficiency.

Related

  1. Prediction of incident atrial fibrillation from retinal fundus images using a multimodal foundation model
    Yupeng Xu, Yige Peng, Yan Jiang +11
    NPJ Digital MedicinePeer-reviewedCC BY-NC-ND 4.0Full textMedical imaging & radiologyPMC
  2. Clinical validation of a novel catheter–tissue contact algorithm in atrial fibrillation ablation using pulsed field ablation
    Ankur Shah, Saanvi Billakanty, Andrew E. Volio +10
    Heart rhythm O2Peer-reviewedCC BY-NC-NDFull textCardiovascular medicineEurope PMC
  3. Predictive Value of P-Wave Dispersion Combined with Left Atrial Reservoir Strain for Incident Atrial Fibrillation in Patients with Hypertrophic Cardiomyopathy: A Retrospective Cohort Study.
    Li Yu, Yan Zhang, Dandan Sun +2
    International journal of general medicinePeer-reviewedCC BY-NCFull textCardiovascular medicineEurope PMCPMC
  4. Non-alcoholic Fatty Liver Disease/Metabolic Dysfunction-Associated Steatotic Liver Disease and the Risk of Incident Atrial Fibrillation: An Updated Systematic Review and Meta-Analysis of Cohort Studies
    Disha Disha, Allah Rakhio Bhurgri, Deepak Verma +5
    CureusPeer-reviewedCC BYFull textCardiovascular medicineEurope PMCEurope PMCPMC
  5. Drug Therapies and Incident Heart Failure in Atrial Fibrillation: A Systematic Review
    Amirreza Zobdeh, Gregory M. Peterson, Woldesellassie M. Bezabhe +2
    Journal of Clinical MedicinePeer-reviewedCC BY 4.0Cardiovascular medicineCrossref

Cite

BibTeXRIS
BibTeX
@article{kim2026impact,
  title = {{Impact of an AI algorithm for multi-day prediction of incident atrial fibrillation on clinical decision-making: PROVISION-AF study}},
  author = {Yeji Kim and Bogeun Kim and Jaeung Yoon and Jeongmin Kang and Seung Woo Lee and Moon-Hyun Kim and Dong-Hyeok Kim and Hye Ah Lee and Jaemin Shim and Dae-In Lee and Seung-Young Roh and Seung Young Shin and Sung Hwa Choi and Yae Min Park and Ki Hong Lee and Yoo Ri Kim and Min Kim and Wei-Jin Park and Hyeon Yeo and Jungwon Choi and Tae Young Shin and Junbeom Park},
  journal = {NPJ Digital Medicine},
  year = {2026},
  month = jul,
  doi = {10.1038/s41746-026-02950-6},
  url = {https://doi.org/10.1038/s41746-026-02950-6},
  note = {CC BY-NC-ND 4.0},
}
RIS
TY  - JOUR
TI  - Impact of an AI algorithm for multi-day prediction of incident atrial fibrillation on clinical decision-making: PROVISION-AF study
AU  - Yeji Kim
AU  - Bogeun Kim
AU  - Jaeung Yoon
AU  - Jeongmin Kang
AU  - Seung Woo Lee
AU  - Moon-Hyun Kim
AU  - Dong-Hyeok Kim
AU  - Hye Ah Lee
AU  - Jaemin Shim
AU  - Dae-In Lee
AU  - Seung-Young Roh
AU  - Seung Young Shin
AU  - Sung Hwa Choi
AU  - Yae Min Park
AU  - Ki Hong Lee
AU  - Yoo Ri Kim
AU  - Min Kim
AU  - Wei-Jin Park
AU  - Hyeon Yeo
AU  - Jungwon Choi
AU  - Tae Young Shin
AU  - Junbeom Park
PY  - 2026
DA  - 2026/07/13
JO  - NPJ Digital Medicine
DO  - 10.1038/s41746-026-02950-6
UR  - https://doi.org/10.1038/s41746-026-02950-6
AB  - Atrial fibrillation (AF) is frequently asymptomatic and often remains undetected until complications arise. Although artificial intelligence (AI)-enabled electrocardiography (ECG) can predict incident AF from sinus rhythm ECGs, its influence on physician risk assessment in simulated clinical settings remains uncertain. We developed a deep learning model to predict multi-day AF risk using non-AF 12-lead ECGs. For ECG-labeled outcomes, the model achieved an AUROC of 0.79 in the internal EUMC cohort and 0.74 in the external BIDMC cohort. For Holter-labeled outcomes, AUROC values reached 0.87 in the internal EUMC subset and 0.75 in the prospective PROVISION-AF cohort. To assess decision-support utility, a multinational survey of 70 physicians evaluated how AI-derived risk estimates influenced physician risk assessment and follow-up decisions in structured simulated cases. AI assistance significantly improved physicians’ AF risk discrimination (AUROC 0.573 to 0.650) and negative predictive value (0.764 to 0.839), with significant net reclassification improvement for non-AF cases (NRI 0.127, p < 0.001). Performance gains were most notable among non-electrophysiologist cardiologists. In conclusion, AI-derived risk estimates improved physician risk discrimination in a structured simulated survey, particularly in non-specialist settings, supporting their potential role as a digital decision-support tool. Further real-world implementation studies are needed to determine whether these effects translate into improved clinical outcomes or healthcare efficiency.
SN  - 2398-6352
ER  -