À compter du 1er décembre 2023, l’accès à POEMs et à Essential Evidence Plus ne fera plus partie des avantages offerts aux membres de l’AMC.
Question clinique
Can a smartwatch correctly diagnose atrial fibrillation?
L’Essentiel
In this large pragmatic diagnostic trial, the positive predictive value for a notification of possible atrial fibrillation (AF) by a smartphone app was 34%. However, the use of the app led to lots of false alarms. An end-of-the-study survey indicated that the vast majority of new diagnoses of AF were given independent of the app (n = 3070) versus detected by the app (n = 153). Scaling this up to a population of 40 million adults would result in 200,000 notifications that would require follow-up. The question of whether the app results in improved patient-oriented outcomes due to earlier diagnosis remains unanswered. Gotta go, my watch just told me to stand up! 4
Référence
Plan de l'etude: Diagnostic test evaluation
Financement: Industry
Cadre: Population-based
Sommaire
As we all know, devices like the Fitbit and Apple Watch can monitor our heart rate. In this study, Apple developed an app that could monitor a wearer's heartbeat and identify the irregular pattern of AF. They invited any US resident with an Apple Watch who was at least 22 years of age to download the app. Over 8 months, they recruited more than 419,000 participants who were monitored for a median of 117 days. The mean age of participants was 41 years, with only 16% older than 54 years. If the app detected what it thought was AF (n = 2161), the user was asked to schedule a telemedicine visit. If they had chest pain, dyspnea, or other "urgent symptoms" they were advised to go to the emergency department. Otherwise, they were mailed an electrocardiography (ECG) patch that they were asked to wear for one week. Of the 2161 watch wearers with an irregular pulse notification, only 945 made a telemedicine visit, of whom 20 had urgent symptoms, 174 already had a diagnosis of atrial fibrillation, and 101 were excluded because they were taking an anticoagulant (or for other reasons). That left 658 of whom only 450 successfully returned the ECG patch for analysis. The ECG patch confirmed a diagnosis of AF in 153 (34%). The authors speculate that this positive predictive value is low because the ECG patch was applied on average approximately 2 weeks after the initial notification, and many episodes of lone or sporadic AF may have passed by that time. They also compared app notifications that occurred while the patch was being worn with the ECG patch reading to calculate a positive predictive value of 71% overall and 60% in adults 65 years and older. At the end of the study, 70% of the 291,995 participants who received no app notification completed a survey and 3070 reported being given a new diagnosis of AF independent of the smartwatch monitoring.
Reviewer
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA