Abstract Purpose: Atrial fibrillation (AF) is a leading cause of stroke and heart failure, yet detection remains suboptimal in rural settings due to limited resources. This study evaluates whether an enhanced screening strategy using artificial intelligence (AI)-integrated 7-day single-lead electrocardiogram (ECG) patches improves AF detection and long-term clinical outcomes compared to routine care in rural China. Methods: This cluster-randomized trial will be conducted across 128 village clinics in Quzhou, Zhejiang Province. Villages are randomized 1:1 to either enhanced or routine screening. Participants aged 60 years or older (approximately 120 per village) in both arms receive family-centered AF education and opportunistic assessments. The enhanced group undergoes screening via 7-day single-lead ECG patches, while the routine group utilizes standard 12-lead ECGs. Results: The trial features two primary endpoints. The Phase 1 endpoint is the newly diagnosed AF detection rate during a 1-year screening period. The Phase 2 endpoint is a 3-year composite outcome of all-cause mortality, stroke or systemic embolism, and hospitalization for heart failure. Conclusion: By integrating wearable AI technology into primary care, this trial seeks to overcome diagnostic barriers in resource-limited environments. The findings will determine if prolonged digital monitoring can significantly enhance AF detection and reduce major cardiovascular events in elderly rural populations.
Effectiveness of Artificial IntelliGence-Driven Single-Lead Long-Term Electrocardiograms Monitoring In Detecting New-Diagnosed Atrial FIbrillation (GEMINI) trial is a parallel, two-stage cluster randomized trial being conducted in 128 villages in Qujiang District, Quzhou City, Zhejiang Province. Village clinics serve as the primary health care units and provide essential services to rural residents, with one clinic allocated to each administrative village in China. Each village clinic is staffed by a village doctor with basic medical training (certificate-level) and operates under the supervision of physicians based at 8 township health centers. These township health centers are, in turn, overseen by a single district-level hospital, forming a three-tier hierarchical healthcare system . This decentralized structure, designed to address the shortage of specialist resources at the village level, enables consistent outreach and sustained medical engagement across rural populations. In this study, the villages (clusters) were randomly assigned to either the enhanced screening group (intervention arm) or the routine screening group (control arm). The coprimary outcomes are the 1-year detection rate of newly diagnosed AF (Phase 1) and a 3-year composite of all-cause mortality, stroke or systemic embolism, and hospitalization for heart failure (Phase 2).
Study Type
INTERVENTIONAL
Allocation
RANDOMIZED
Purpose
SCREENING
Masking
NONE
Enrollment
15,360
Participants in the long-term screening group will undergo baseline data collection and a 12-lead ECG at baseline, followed using a 7-day single-lead long-term ECG monitoring device. This device continuously collects dynamic ECG data for seven days and allows participants to view and store ECGs through the accompanying software (registration number: 20192070163).
Patients in the standard screening group will undergo baseline data collection and a standard 12-lead ECG at the start of the study, with follow-up including opportunistic ECG checks as needed. Those diagnosed with AF will receive anticoagulation and antiarrhythmic therapy according to clinical guidelines.
Beijing Anzhen Hospital
Zhejiang, Quzhou, China
RECRUITINGDetection rate of new atrial fibrillation diagnoses
Detection rate of new atrial fibrillation diagnoses in both groups. Atrial fibrillation is defined by a single-lead ECG (≥30s) or a 12-lead ECG (≥10s) showing absent P waves, replaced by irregular fibrillation waves (f-waves) of varying size, shape, and duration, and absolute irregularity of RR intervals. Two cardiologists will interpret the ECG, and if there is doubt, a third cardiologist will review the case. Opportunistic ECG is defined as any ECG performed for medical reasons during follow-up (primarily 12-lead ECG), which must be reported and diagnosed.
Time frame: 1 year
Composite endpoints
The all-cause mortality, stroke/systemic embolism, and heart failure hospitalization rate in the study population.
Time frame: 3 years
Atrial Fibrillation Burden Assessment
Assessment of atrial fibrillation burden, defined by the duration and frequency of episodes, measured through continuous monitoring and ECG evaluations.
Time frame: 1 year
All-Cause Mortality Rate
The all-cause mortality rate in the study population, measuring the number of participants who have died from any cause.
Time frame: 3 years
Cardiovascular Mortality Rate
The cardiovascular mortality rate, defined as death directly caused by cardiovascular diseases, including acute myocardial infarction, heart failure, stroke, etc.
Time frame: 3 years
Stroke/systemic embolism:
Stroke is defined as an acute episode of focal or global neurological dysfunction caused by intraparenchymal, intraventricular, or subarachnoid hemorrhage, or central nervous system infarction. Systemic embolism refers to a thrombus or other material traveling through the bloodstream and causing embolism in other vital organs (e.g., kidneys, limbs, etc.).
Time frame: 3 years
Heart Failure Hospitalization Rate
The rate of hospitalization due to heart failure in the study population.
Time frame: 3 years
Cardiovascular Hospitalization Rate
The rate of hospitalization due to cardiovascular events (excluding heart failure) in the study population.
Time frame: 3 years
Detection Rate of New Atrial Fibrillation
The detection rate of new atrial fibrillation diagnoses during the study period, defined by the criteria set for ECG findings.
Time frame: 3 years
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.