Incidence and delays of a new clinical atrial fibrillation diagnosis following an atrial fibrillation detection signal on a personal electrocardiogram device available to consumers

Abstract

Abstract Background Patient and consumer-led detection of atrial fibrillation (AF) using medical devices available to consumers such as personal electrocardiogram (ECG) devices is becoming increasingly common. However, the AF patient journey after an AF signal observed on the personal ECG device is variable and poorly defined. Purpose To report the incidence and temporality of a new clinical AF diagnosis among individuals with a device-detected AF signal. Methods This retrospective cohort study was conducted using the OMNY Health open-source claims database linked to data recorded from a medical grade, United States Food and Drug Administration (FDA)-cleared, personal ECG device available to consumers. Adults (≥18 years) with ≥1 device-detected AF signal were selected between July 1, 2017, and December 31, 2020. The date of the first device-detected AF signal was defined as the "index date." Individuals were required to have continuous healthcare activity for 12 months before (baseline period) and 12 months after (follow-up period) the index date, and ≥2 device recordings (of AF or not) during a 12-month follow-up period. Individuals with a history of AF diagnosis in the baseline period were excluded. Outcomes evaluated during follow-up included: 1) New clinical AF diagnosis (defined as ≥1 non-diagnostic medical claim with an AF diagnosis [ICD-10 I48.0-I48.2, I48.91] in any position); 2) Time from index date to new clinical AF diagnosis; and 3) Number of device-detected AF signals. Results The study included 32,484 individuals with ≥1 device-detected AF signal (mean 11.8, SD: 23.7 signals). Mean age was 68.7 (SD: 11.9) years; 71.4% were 65 years or older; and 43.1% were female. The mean baseline CHA2DS2-VASc risk score was 2.0 (SD: 1.3). Overall, about a quarter (24.1%; n=7,822) of individuals had a new clinical AF diagnosis within 1 year of index date (Figure 1). The likelihood of a new clinical AF diagnosis was lower for younger age groups (age 18-44 = 8.0%; age 45-54 = 16.6%; age 55-64 = 23.0%; age 65-74 = 26.1%; age 75+: 26.2%) (Figure 2), and individuals with certain CHA2DS2-VASc risk factors (congestive heart failure: 26.4%; hypertension: 26.3%) had a slightly higher likelihood of a new clinical AF diagnosis than the overall population. In those with a new AF diagnosis, this was delayed until after 90 days from the index date in 46.7% of the patients (Figure 1). Conclusions This observational study offers unique real-world insights into the use of personal ECG devices and their potential ability to lead to clinician-diagnosed AF. Only a quarter of patients had a new clinical AF diagnosis within a year of their first device-detected AF signal, and diagnosis was delayed by >90 days in almost half of cases. This may relate to false positives of the device AF diagnosis, and/or poorly defined AF pathway for patient- or consumer-led AF detection.