Are Wearable ECG Devices Ready for Hospital at Home Application?

Abstract

The increasing focus on improving care for high-cost patients has highlighted the potential of Hospital at Home (HaH) and remote patient monitoring (RPM) programs to optimize patient outcomes while reducing healthcare costs. This paper examines the role of wearable devices with electrocardiogram (ECG) capabilities for continuous cardiac monitoring, a crucial aspect for the timely detection and management of various cardiac conditions. The functionality of current wearable technology is scrutinized to determine its effectiveness in meeting clinical needs, employing a proposed ABCD guide (accuracy, benefit, compatibility, and data governance) for evaluation. While smartwatches show promise in detecting arrhythmias like atrial fibrillation, their broader diagnostic capabilities, including the potential for monitoring corrected QT (QTc) intervals during pharmacological interventions and approximating multi-lead ECG information for improved myocardial infarction detection, are also explored. Recent advancements in machine learning and deep learning for cardiac health monitoring are highlighted, alongside persistent challenges, particularly concerning signal quality and the need for further validation for widespread adoption in older adults and Hospital at Home settings. Ongoing improvements are necessary to overcome current limitations and fully realize the potential of wearable ECG technology in providing optimal care for high-risk patients.

Keywords: artificial intelligence; cardiac health; patient outcomes; remote patient monitoring; wearable devices.

Figure 1

A selection of ECG wearables: (a) from left to right, FitBit © Sense, Apple © watch series 4, Samsung © Galaxy Watch 4 and Withings © ScanWatch used in [6]; (b) the Withings © body scan is a smart scale that displays ECG data acquired via a multi-electrode grip (6-lead); (c) small, portable, medical-grade personal ECG device from AliveCor used in [6]; (d) ECG patch for continuous monitoring.

Figure 2

Standard values for amplitudes and intervals, according to the Association for the Advancement of Medical Instrumentation (AAMI), for normal sinus rhythm.

Figure 3

(a) Patient-specific training. (b) Continuous ECG monitoring and heartbeat classification in real time. Adapted from Ref. [44].

Figure 4

(a) Hardware platforms. (b) Measured execution time. (c) Distribution of the execution time. Adapted from Ref. [44].

Figure 5

Flow chart detailing the development, validation, and potential future applications of the pericarditis deep learning model (DLM). Adapted from Ref. [47].

Figure 6

Automatic diagnosis of AF for the Apple Watch Series 5, Samsung Galaxy Watch Active 3, and Withings Move ECG (p = 0.02 between Withings and Apple) and (p = 0.03 between Samsung and Withings) [2].

Figure 7

Model performance as the number of leads increases. Graph was obtained by using results from the paper (12-lead sets: AUROC 0.880; 4-lead sets: AUROC 0.858, SD 0.008; 3-lead sets: AUROC 0.845, SD 0.011; 2-lead sets: AUROC 0.813, SD 0.018; single-lead sets: AUROC 0.768, SD 0.001) [35].

Figure 8

ECG Smartwatch of a patient with symptoms of chest pain and 12-lead ECG and confirmed previous STEMI. Typical use location on left wrist records a normal ECG of Einthoven lead I. Adapted from Ref. [48].

Figure 9

Placement of smartwatch for recording Einthoven and precordial leads. (A) Einthoven lead I. Recording between the left wrist and the right index finger. (B) Einthoven lead II. Recording between the left abdominal region and the right index finger. (C) Einthoven lead III. Recording between the left abdominal region and the left index finger. (D) Right Wilson lead (Wr). Recording at the right fourth parasternal intercostal space. (E) Medial Wilson lead (Wm). Recording from the midclavicular line of the fifth intercostal space. (F) Left Wilson lead (Wl). Recording from the fifth intercostal space of the left midaxillary line. Adapted from Ref. [51].

Figure 10

Electrocardiogram with ST segment elevation in Wm (V4). (Wr V1 and WL V6). Adapted from Ref. [50].

Figure 11

QTc interval assessed by AI on single-lead smartwatch ECG (AI-QTc) with QTc measured by conventional 12-lead ECG in patients with early-stage COVID-19 treated with HCQ-AZM. Adapted from Ref. [52].