Smartphones and wearable technology: benefits and concerns in cardiology

The global expansion of wearable technology combined with smartphone access creates new questions and opportunities in the diagnosis and management of cardiac conditions

Wearable devices along with smartphone technology are becoming more common in developed nations such as Australia, with the number of connected devices expected to increase from 526 million in 2017 to over 1.1 billion worldwide in 2022.1 This new source of health information has led to novel methods of patient assessment, such as a single‐lead electrocardiograph (ECG) creating a pre‐presentation observation chart. This creates a paradigm shift where instead of symptomatic patients being referred for cardiac assessment, asymptomatic patients are now presenting with health data. In this article, we discuss the technology behind the devices, common brands and formats, potential applications, and advantages and disadvantages of wearable devices.

Devices

At the time of writing, there is a plethora of patents worldwide, with 423 devices from 132 brands in wrist‐worn trackers alone.2 However, few brands are featured in research, the most common being Fitbit (Fitbit, Inc), Garmin (Garmin Ltd), Misfit (Misfit Technologies), Apple (Apple Inc) and Polar (Polar Electro). Some of these devices are generally not classified by the United States Food and Drug Administration, such as the Garmin and Fitbit fitness trackers, as they are not marketed directly as health treatment devices. Devices can broadly be grouped into three types: chest‐strap devices, photoplethysmography (PPG), and portable single‐lead ECG (Box 1 and Box 2).

The first wearable technology devices created were chest‐strap devices which used electrodes to record heart rate.3 They differ from modern ECG based devices in that they do not record an actual rhythm, despite being able to measure chronotropic response to a correlation efficient of > 0.93 compared with ECG.4

The next set of technologies uses PPG. This technique involves optical technology to assess variations in blood volume within the microvasculature,5 effectively measuring each heart beat as a pulse, similar to pulse oximetry used in hospital settings. Accuracy is often limited by anatomical factors including location, movement, skin colour and conductivity. Fingertip PPG is highly accurate, with meta‐analysis demonstrating a mean difference of 0.32 beats per minute (99% CI, − 1.25 to 0.60) compared with ECG.6 Smartphones are the most commonly used device to feature PPG; however, wrist‐worn devices also exist, such as smart watches and heart rate tracking devices. These are reasonably accurate, with correlation coefficients > 0.93 and mean absolute percentage errors ranging from 3.3% to 6.2%.4

The latest set of devices use a built‐in electrode system to create single‐lead ECG, such as the Apple Watch Series 5, enabling more advanced analysis. These devices require the patient to hold the watch in the contralateral hand, touching the crown or band of the watch, creating a vector to generate a second lead. The other most common ECG device is a small tablet, such as the Coala (Coala Life Inc) and Zenicor‐ECG (Zenicor Medical Systems).

Applications

Wearable technology is used in heart rate tracking, with high levels of accuracy in both smartphone6 and wrist‐based formats.4 Relevant clinical assessments are similar to conventional ambulatory ECG monitors and include measuring adequate rate control in atrial fibrillation (AF), and aiding the diagnosis of bradyarrhythmia or tachyarrhythmia.

Although PPG sensors were initially only designed to track heart rate, there has been a push to use these with algorithms in the detection of arrhythmias such as AF. One analysis created an algorithm using data from 9750 participants, including 347 with AF; overall sensitivity was 98% and specificity was 90%.7 The largest such study to date has been the Apple Heart Study, a wearable PPG study using the Apple Watch, in which 419 297 people self‐enrolled.8 The aim of the study was to identify irregular pulses on PPG, and the diagnosis of AF with a confirmatory ECG patch. A total of 450 ECG patches were included based on PPG analysis by doctors, with AF documented in 34% of these patients. In patients ≥ 65 years of age, the rate of irregular heartrate notification was 3% with a positive predictive value for a final diagnosis of AF of 84. Several limitations were noted, including a self‐enrolment selection bias of possible health conscious individuals, a lower than expected target enrolment, and reporting bias with only 68% of ECG patches returned to the study organisers.

Manufacturers are now expanding beyond PPG to electrode based technology. This aims to detect AF not through beat‐to‐beat variation, but via single‐lead electrode printout. This was demonstrated in the REHEARSE‐AF Study,9 which involved 1001 at‐risk patients (CHADS‐VASc score ≥ 2), leading to the subsequent diagnosis of AF in 19 patients compared with five patients in the control arm. Orchard and colleagues found that of 1805 Australian patients screened by their local general practitioner,10 67 (3.7%) were diagnosed with possible AF by the algorithm and 19 (1.1%) were formally confirmed by 12‐lead ECG. The authors estimated that the costs would approximate $1000 in set‐up fees, as well as $10 per patient, on top of normal consultation fees.10 Although evidence for wearable technology is currently scarce, this potentially represents a cost‐effective way of implementing opportunistic age‐appropriate AF screening as per local guidelines.11

Another area in which portable technology may be useful is the correlation of symptoms with arrythmias to guide diagnosis and treatment. There is increasing evidence favouring the use of electrode‐based portable technology12 due to ease of access and availability, particularly in patients presenting with infrequent symptoms that may not be captured on traditional event recording.

One of the strongest advantages of wearable devices is their ubiquity and ease of use, which may allow for screening and further monitoring in a large population of patients. This is relevant in Australia for those located in regional and rural areas with limited access to specialist investigations. Future options include the newly approved 6‐lead ECG by AliveCor. In addition to the two‐finger electrode, this includes a third electrode on the back of the device, which extends to the patients left knee, creating Einthoven's triangle and producing limb leads (I, II, III, aVR, aVL and aVF).

Concerns and limitations

One of the limitations of wearable technology devices is the data itself, not data interpretation. PPG analysis has been shown to have significant variability in the analysis of cardiac rhythm,13 and a trial using electrode based technology resulted in 34% of the recordings being unclassifiable.

Another key limitation is that for these devices to be available in Australia, they require approval from the Therapeutic Goods Administration (TGA). Importantly, Apple has not received such approval for their ECG app on the Apple Watch, meaning it is currently unavailable in Australia. However, there are current recommendations to streamline this approval process.

Further questions arise regarding the medical implications and ethics of these newly available data, as they create previously unstudied situations. For example, does a single bout of asymptomatic AF detected on a wearable device require further investigation or warrant therapeutic anticoagulation? There are only a few studies which look at subclinical AF, all with varying results, with one subgroup analysis showing that a significant increase in stroke risk only occurs after 24 hours of sustained AF.14

If an arrhythmia is recorded by the device, does medico‐legal responsibility lie with the company, the patient or a medical professional to interpret and manage this finding? What if the technology or patient choices result in individuals being lost to follow‐up? This was seen in the Apple Heart Study, where only 68% of patients returned ECG patches.8

Ethical concerns also exist around data: are they the property of the manufacturer, and are they freely available to influence epidemiological research? From an individual point of view, will these asymptomatic findings affect future private insurance premiums? If portable technology does become universally adopted and accurate screening algorithms are identified, will governments be able to perform screening to benefit public health? The issue of health equality is also of concern, as those with higher incomes may have access to devices not attainable for others. There are also concerns regarding the limited research into such devices, and the potential risk of exaggerated marketing claims on their benefit, with multiple fraud claims documented previously.15

Finally, there are concerns around the cybersecurity and regulation of novel devices, for which there is both minimal regulation and large variance on the aforementioned factors. Security is heterogeneous between devices, with concerns regarding who physically owns the data once they are collected, and whether the data are available to the company to market, even if de‐identified.

Conclusion

Wearable devices and smartphones are a widespread, globally adapted form of technology that is only expected to increase. This presents new challenges for clinicians to assess patient data, particularly in relation to the analysis of rate and rhythm in asymptomatic patients. We suggest that clinicians be aware of the potential benefits and applications of this technology, as well as the concerns and limitations that currently exist. Further evidence and guidelines are required to clarify the role of wearable devices from a therapeutic standpoint as well as the medico‐legal and ethical implications of arrhythmia identification.

Box 1 – Selected examples of wearable technology and smartphone photoplethysmography (PPG)

---

A: Polar H10 chest strap device. B: Garmin Forerunner 735XT PPG based wristwatch. C and D: Samsung S9 smartphone and Samsung Health PPG based app. E: Coala Heart Monitor electrode based device. F: AliveKor KardiaMobile electrode based device.