The study has shown the accuracy of wearable dynamic ECG recorder for AF detection from sinus rhythm in a small population trial setting. It provided a non-invasive, easy-to-use, and affordable tool to detect and discriminate AF in different positions and after exercises. The instrument used in the research is the first domestic instrument that applies an AI algorithm, and a single-lead ECG applied a wristband, which is convenient user-friendly to operate and convenient to wear. The wearable dynamic ECG recorder with an AI algorithm can accurately detect AF in different postures. As such, the specificity and positive predictive value reached 100%. When the position was changed to stand and measured after exercise, it was easier to detect the signal and make the correct diagnosis. The AI algorithm (Huami Technology) evaluated in the study has been fully trained and tested via large-scale real user data to make the algorithm reliable. AI is developed using deep convolutional neural networks. The test set's official data's sensitivity and specificity were 93.36% and 99.75%, respectively.
The single-lead ECG provided physicians with higher specificity and a clear review of the ECG records. Furthermore, the wearable dynamic ECG recorder did not affect daily activities and was waterproof, safe, battery-powered, and electrically safe. The wristband tested in the study did not require frequent communication with smartphones, thus decreasing power consumption and increasing the time of continuous data recording. It could also stand on standby for seven days on a fully charged battery.
The study provided the results of lying position, standing position and exercise to simulate the tool's test results in different states. Moreover, when the application detects AF, it promptly sends text messages to the wearer, related relatives, and a designated doctor assigned by Huami. The designated doctor would then provide patients with further diagnosis and treatment.
Early intervention and qualifiable risk factor control could reduce the incidence of AF. Hence, provide patients with or at risk an essential tool to detect AF at a particular time or over a long period, thus promoting the detection and management of silent AF early before adverse health consequences such as ischemic stroke or heart failure occur.
Nowadays, there are several studies on the detection of AF using ECG signals. It is easy to detect irregular pulse beats by palpation of the pulse. Based on simple training of 173 subjects aged ≥ 75-years, the elderly subjects were able to identify the sinus rhythm better after been educated, compared with the healthcare control group. The study showed that the proportion of slow (81.8% vs. 56.1%) and fast AF (91.9% vs.80.7%) were significantly better than the control group.  Also, older adults could adequately identify normal rhythms by self-palpation of the pulse after been educated, although for most individuals, it is more challenging to find irregular pulse without adequate training. Nonetheless, the dynamic ECG armband recorder proves a useful low-cost method for screening asymptomatic AF patients.
Presently, the 12-lead ECG remains the gold standard for the diagnosis of AF. In a study by Jonas D.E. et al., ECG screening detected more new AF cases than no screening (absolute increase from 0.6% [95% CI, 0.1%-0.9%] to 2.8% [95% CI, 0.9%-4.7%] over 12-months).  The study also showed that ECG detected no more cases than an approach using pulse palpation.  Due to the time constraint of recording, AF findings with ECGs are significantly limited. Holter makes up for the missed detection due to the time constraint and the prolonged ECG monitoring duration. In other study with 105 enrolled patients ≤ 50 years, 95 patients (90%) were cryptogenic. 
In the study, paroxysmal AF was detected in nine patients (two during 24-hour ECG Holter and seven during 3-weeks Holter monitoring). The results showed that prolonged ECG monitoring could improve the detection rate of paroxysmal AF. The implantable circulation recorder is an invasive screening method for AF, which can further improve paroxysmal AF screening with a longer monitoring time. Deshmukh et al.  conducted a study to evaluate the performance of a single-chamber implanted cardioverter-defibrillator (ICD) in detecting AF, and the research results showed that its sensitivity and specificity in the diagnosis of AF were 95.0% and 99.6%, respectively. However, its high cost, invasive, and difficulty in promotion seriously limit the possibility of its large-scale use, making it impossible to achieve universal screening in a large-scale target population.
Screening AF using traditional methods is challenging. As technology advances, the present research has confirmed that smart devices, such as mobile phones, handing devices, and wearable devices, can be used for AF detection. McManus et al.  provided an algorithm connecting the root mean square of successive RR difference (RMSSD/mean) and Shannon entropy (ShE). The pulse of seventy-six adults with persistent AF were recorded before and after cardioversion using an iPhone 4S camera. The algorithm demonstrated excellent sensitivity of 96.2%, a specificity of 97.5%, and an accuracy of 96.8% for the beat-to-beat distinction of an irregular pulse during AF from sinus rhythm.
Svennberget et al.  described a handheld ECG recorder for intermittent ECG recordings with an integrated mobile transmitter that sends 30-second ECG strip data to a database. Participants placed their thumbs on the device twice daily, and whenever they noticed palpitation. Finally, 118 cases (3.0%) were diagnosed with AF, among which only thirty-seven cases (0.5%) were detected with the first handheld ECG, and eighty-one cases (2.5%) were detected with repeated tests. The study showed that repeated routine ECG examination over a long period could improve AF detection sensitivity, which was four times higher than the number of cases diagnosed by routine ECG examination at a single time point.
In 2017, the AF-SCREEN international collaboration  confirmed that concerning the method of mass screening, handheld ECG devices have the advantage of providing a confirmable ECG trace, which is required by the diagnostic guidelines for AF, and thus it is preferred as a screening tool. Steinhubl et al.  conducted a mSToPS randomized clinical trial to confirm the impact of a self-applied wearable ECG patch in detecting AF. The results showed that the new diagnosis rate of AF in the immediate group at four months was higher than the delayed group (3.9% vs. 0.9%, the absolute difference was 3.0%, 95% CI 1.8% ~ 4.1%).
A recent technology of photo-plethysmography (PPG) application was mentioned for the detection of AF. The PPG algorithm’s sensitivity and specificity for AF detection were 97–100% and 92–94%, respectively.  The Apple-Heart Study  is a mobile application that screens participants by measuring blood flow change flow through the wristband to generate a PPG. The study showed that if periodic signals are detected with PPG technology, participants will receive notifications on the Apple Watch and Apple-Heart study application. After the participant contacts the physician in charge, the physician would decide if they should wear the ECG monitoring patch.
The study demonstrated the positive role of smart wearable devices in AF screening. HUAWEI Heart-Study-MAFA II  concluded higher sensitivity and specificity of the PPG method in AF screening and achieved a sensitivity of 100%, specificity of 99%, and a positive predictive value of 91.6%. Nevertheless, their data was collected in a supine position. Nonetheless, the supine position does not reflect the actual situation for home screening, where the movement has a more significant influence on the PPG signal. Interference caused by movement should be avoided between single-lead ECG and the PPG technology to improve accuracy. Since single-lead ECG recordings with mobile phones or wristbands for AF detection are clinically superior to PPG signals, they tended to use single-lead ECG recordings in several patients with paroxysmal AF.
With the increase of health awareness, wearable health monitoring is gaining attention, allowing patients to manage symptoms from their own homes comfortably. The wearable dynamic ECG recorder is a feasibility and accurate way for people in need to monitor and track their ECG recordings and share them with their attending physicians.
There are a few limitations to the study. Firstly, the ECG monitoring was collected in no symptoms, not in participants with symptoms. Secondly, our instruments were only used for discriminating between AF and sinus rhythm. The wearable dynamic ECG recorder cannot detect other arrhythmias, such as premature beats, atrial tachyarrhythmias, and atrial flutter. In our future study, new algorithms will be added to help identify and distinguish sinus arrhythmias and various arrhythmia forms. Thirdly, the participants were not followed up. We do not know if people with a normal ECG will develop AF later, and keep the physicians recommended AF guidance. Finally, the sample size was relatively small, and more extensive testing will be needed to be performed in the future.