WEAR 2

Objectif(s) de la recherche et intérêt pour la santé publique

Finalité de l'étude

Objectifs poursuivis

Domaines médicaux investigués

Bénéfices attendus

Public health significance
The retrospective study addresses an important public health need: improving the timeliness and efficiency of ambulatory ECG interpretation in the context of increasing ECG data volumes and limited specialist cardiology resources. It evaluates whether an AI-enabled Fast Track workflow can shorten Holter analysis and report turnaround times while preserving the validity of arrhythmia assessment. By supporting earlier identification of atrial fibrillation and other clinically relevant rhythm disorders, the project may contribute to faster diagnosis, more efficient use of healthcare resources, and improved cardiovascular care pathways, including for patients assessed after cryptogenic stroke.

Study objectives
The primary objective is to evaluate the validity and usefulness of an AI-enabled Fast Track validation workflow in routine clinical practice, with particular focus on identifying Holter recordings without major abnormalities.
- Assess the validity and usefulness of a dedicated algorithm and workflow for atrial fibrillation identification, especially in post-stroke patients.
- Measure the reduction in Holter analysis time and report turnaround time compared with the baseline workflow.
- Evaluate diagnostic performance using sensitivity and negative predictive value for recordings without major abnormalities, and sensitivity, specificity, and predictive values for atrial fibrillation assessment.
- Compare Fast Track results with those obtained using the standard workflow and commercially available AI algorithm.

Methodological approach
This is a retrospective, observational, non-interventional, comparative study for research purposes. Existing ECG data collected during the prospective WEAR study will be reprocessed using an investigational version of the Cardiologs Holter Platform that includes Fast Track labelling, an accelerated validation workflow, and algorithm improvements. Results will be compared with the baseline analyses performed using the standard workflow and the commercially available algorithm.

Fast Track eligibility will be assessed using predefined strict and lenient criteria for two principal recording categories: recordings without major abnormalities and recordings obtained for atrial fibrillation detection. The analyses will examine workflow efficiency and diagnostic performance. Descriptive statistical methods will follow the approach specified for the prospective study.

Study population
The retrospective population will consist of the same adult patients enrolled in the prospective WEAR study, which was approved by an ethic comittee, using the same inclusion and exclusion criteria. The target is 1,500 patients from the Cardiology and Neurovascular Departments of Institut Cardiovasculaire Paris-Sud (ICPS), Massy, and Assistance Publique - Hôpitaux de Paris (AP-HP), Hôpital Pitié-Salpêtrière, Paris. The population includes patients prescribed ambulatory Holter monitoring for any indication, including patients monitored for atrial fibrillation after cryptogenic stroke. Patients with implanted pacemakers or implantable cardioverter-defibrillators and patients who are pregnant are excluded.