Many False Alarms With Smartphone Apps for AF Detection

By Reuters Staff

April 10, 2020

NEW YORK (Reuters Health) - Smartphone apps for detecting atrial fibrillation (AF) may be useful for ruling out the condition, but the risk of false alarms is considerable, according to a new systematic review and meta-analysis.

"The modeled NPV (negative predictive value) was high for all analyses, but the PPV (positive predictive value) was modest, suggesting that using these applications in an asymptomatic population may generate a higher number of false-positive than true-positive results," Dr. Jack W. O'Sullivan of Stanford University, in Palo Alto, California, and colleagues write.

Up to 700,000 people in the U.S. may have undiagnosed AF, and about one in five AF-related strokes are in people who have not been diagnosed, Dr. O'Sullivan and his colleagues write in JAMA Network Open.

"Appropriate, early diagnosis is particularly pertinent for AF given that oral anticoagulants may have a favorable risk-benefit ratio for many patients," they add.

Applications for detecting AF use the smartphone camera to track fingertip pulse, and analyze the resulting photoplethysmographic signal, Dr. O'Sullivan and his team explain. To date, studies of these apps have been small and conducted in single centers, they add.

The authors analyzed 10 primary studies including 3,852 patients, a quarter of whom had AF. Overall, combined sensitivity and specificity for the apps were 94.2% and 95.8%, respectively, and both were high for all of the applications analyzed.

However, the combined PPV of the apps was 19.3% for detecting undiagnosed AF in people 65 and older, assuming a prevalence of 1.3%, while negative predictive value (NPV) was 99.9%. PPV ranged from 13.7% to 44.3% for the individual applications.

"Regarding AF screening, if any of these applications reach a diagnosis of sinus rhythm in a healthy, asymptomatic person, it is likely this person does not have AF (ie, true-negative). Conversely, we cannot draw the same conclusion from a positive result," Dr. O'Sullivan and his team write.

"From these results, it would appear premature to use these devices among healthy individuals or to use them to screen an asymptomatic population, including those older than 65 years with hypertension," they add.

"Future research should address the accuracy of these applications when screening other high-risk population groups, their ability to help monitor chronic AF, and, ultimately, their associations with patient-important outcomes," they conclude.

The study had no commercial funding, but some authors report conflicts of interest, including ties to Apple.

Dr. O'Sullivan was not available for an interview by press time.

SOURCE: JAMA Network Open, online April 3, 2020.