A Comprehensive Review of Next-generation Electrocardiogram Technologies, Trends, and Outlook

INTRODUCTION

Statistics as per the World Health Organization (WHO) reveal that cardiovascular disease (CVD) has been the major cause of deaths globally, reaching 17.9 million deaths each year.^[1] The recent COVID-19 pandemic has also posed a new challenge for health care providers whose cardiac patients are at risk due to unmonitored, untreated cardiac condition which may lead to more number of deaths compared to people affected by COVID-19. It is therefore imperative that healthcare providers should think ahead and protect hearts to reduce the mortality rate.^[1-3] Electrocardiogram (ECG) has always been the first choice to monitor and diagnose CVDs. Driven by the urge to continuously monitor and to keep a check on one’s heart health status, in this digital era, smart ECG devices are continuously evolving. It is therefore required to understand, compare, and evaluate the performance of these health monitoring devices with respect to their design and capability to enable end-users to make a proper choice. A comprehensive and systematic literature review is presented in this paper based on the latest trends, technologies, challenges, and outlook in this domain from various perspectives which can set the base to analytically understand the ECG monitoring and analysis units. Figure 1 depicts the complete ECG acquisition, processing, analysis, and monitoring trends and technologies from every viewpoint.

Close

Figure 1: Depiction of electrocardiogram (ECG) acquisition, processing, analysis, and monitoring trends and technologies. (AI: Artificial intelligence, PDAs: Personal digital assistant)

Export to PPT

Continuous monitoring of the heart is a major concern in modern-day healthcare scenario. It has been proven with facts that CVDs could be diagnosed, measured, analyzed, and prevented through continuous monitoring of ECG signals and presents a holistic paradigm for the assessment of CVDs. With the advent of smart and enabling technologies, the ECG monitoring and analysis systems are available for monitoring at remote sites,^[4-16] tele-monitoring, and ambulatory situations,^[17-31] at home environment^[32-39] and in hospitals or intensive care units (ICUs),^[40-44] and Holter monitoring.^[45-49] Advancements in sensor technology have led to the development of Internet of Things (IOT)-based sensors,^[11-16,31] Wireless body area networks,^[50-56] and smart clothing and wearables.^[57-75] Wireless and Information and Communication Technology (ICT) facilities,^[76-84] data analytics, simulations and modeling, machine learning for data mining, signal processing algorithms, feature extraction, classification, neural networks (NN), artificial intelligence (AI), deep learning, etc.^[85-108] have introduced an era of innovative system on chip, robot-assisted devices, implants, and smart devices that can ensure health care better than ever before.^[109-125]

Looking at the expanse in the domain of cardiac monitoring and analysis systems available and the ever-growing need to make it even better by each passing day, an attempt has been made to exhaustively review literature and present the recent trends, technologies, and modern outlook with respect to ECG systems and its analysis approach. Based on the analysis, the following research questions (RQ) are being framed:

• RQ1: What are the similarities and differences in the diagnostic accuracy, use, and patient compliance between next-generation wearable ECG devices and conventional Holter monitors?

• RQ2: How might AI-based algorithms improve the analysis of ECG signal?

• RQ3: What needs to be done to optimize IoMT structures to enable real-time ECG monitoring?

• RQ4: How can interoperability and standardization of ECG data formats across different devices and healthcare information systems be achieved?

• RQ5: What has the COVID-19 pandemic done to the uptake of wearable and remote ECG monitoring technologies, and what does this mean to the future of cardiac care?

INNOVATIVE AND SMART WEARABLES AND DEVICES

Holter technology for ECG monitoring has existed since the 1940s, having its own advantages and limitations.^[45-49] Numerous smart and less obtrusive wearables offering qualitative information comparable to Holter readings are getting clearance from the U.S. Food and Drug Administration (FDA) for long-term monitoring of subjects to gather ECG data. Few such wearables approved are the Cardiac Insight’s Cardea SOLO, ScottCare novi+ Patch Holter, Biotricity Bioflux, iRhythm Zio patch, etc. The reason for adapting to these compact and long-term wearables is the fact that they can be worn for weeks and so have the ability to notice subtle variations in the ECG signal and can assist in timely diagnostics and therapy.^[126]

To quote the Cardiac Insight CEO Brad Harlow, “the spread of COVID-19 pandemic and the need to maintain social distancing have led to the development of next-gen, user-friendly, wearable ECG monitoring devices like Cardea SOLO ECG System which has the capability to reduce the time required to diagnose and treat cardiac patients.”^[127] The device comprises compact, light-weight, water-resistant sensors that permit the user freedom to carry and move with the device, and the recordings captured are comparable to the Holter monitor’s results. Figure 2 shows the comprehensive review of ECG Technologies.

Close

Figure 2: The comprehensive review of electrocardiogram systems (ECG) technologies.

Export to PPT

In this digital era, monitoring of heart rate is also done using fitness trackers, smart watches, etc., which can ably detect tachycardia and atrial fibrillation in healthy individuals and can send an alarm to the physicians, but do not have the FDA approval; hence, doctors have to depend on standard techniques and methods for continuous monitoring and diagnostics. Flexible electronics can help in reducing this challenge with the deployment of electronic skin patches and textiles, as reported by IDTechEx, and have made a prediction that this market may go up to $2 billion by 2030.^[128] These skin patches are in contact with the skin, offering greater mobility, and can be a compromise between the precise, safe, non-ambulatory 12-lead ECG recordings and the recordings collated from consumer electronic health devices which are ambulatory, safe, less expensive but not so accurate. E-textiles combine the functionality of electronics with the esthetics and comfort of textiles to produce smart wear that can continuously monitor the heart health status, respiratory rate, etc., ensuring quality of care, economic, and enhanced patient satisfaction. With the spread of scare due to COVID-19, remote monitoring of patients has become the new norm and the future of e-textiles and patches for ECG monitoring is promising.^[128]

Atrial Fibrillation (AFib) being the major cause of cardiac arrests, a study was performed on more than 60,000 subjects in Belgium, to screen for Atrial Fibrillation using a smartphone app. The results were reported at Heart Rhythm Society’s 40^th annual Scientific Sessions in 2019, and the team demonstrated that data collection was flexible, scalable, and cost-effective. Around 791 new cases of Atrial Fibrillation could be reported. Figure 3 shows the graphical representation of the above analysis. This digital technology enabled health care providers to reach a larger population faster.^[129,130] The technology giant “Apple” has developed a watch which is in sync with an app which has the ability to monitor ECG and notice subtle variations to draw diagnostic inferences. As reported, a person in Germany and in Seattle could be saved using this device from a serious heart condition of atrial fibrillation, which went unnoticed by the physician.^[131] Researchers have come up with innovative ideas to help drivers, military forces, athletes, etc., to track their heart conditions using signals captured from the ears, which as claimed during the annual congress of the European Heart Rhythm Association 2019 in Lisbon can prevent heart strokes and reduce hospitalization. Smart clothing involving only three ECG textile electrodes knitted into the fabric has also been demonstrated recently to capture the ECG signals.^[132] Figure 4 presents the depiction of the medical wearable sensors, smart devices, implants, etc.

Close

Figure 3: The population study as per the Belgium study.

Export to PPT

Close

Figure 4: Depiction of medical wearable sensors, smart devices, and implants. (EEG: Electroencephalography, ECG: Electrocardiography, EOG: Electrooculography, EMG: Electromyography)

Export to PPT

AI-BASED ALGORITHMIC APPROACH

Many facets of the healthcare domain are being explored nowadays, including AI involving machine learning, deep learning, etc., to interpret ECG. ^[133-140] Health care providers can dive deeper into understanding not-so-apparent variations in the ECG waveforms using AI and can save lives. Dave Fornell, the editor of “Feature-ECG” has predicted that AI-inspired algorithms are the future of next-gen ECG devices based on the research presented at the American Heart Association (AHA) meeting over the past 2 years. FDA, in November 2020, cleared the first-ever informative AI-based algorithm AliveCor’s Kardia AI V2 for personal monitoring of ECG as shown in Table 1.^[141] To quote the AliveCor CEO Priya Abani, this is a major breakthrough toward developing a sophisticated AI-based personal ECG monitoring device that offers professional services and covers a wide range of cardiac interpretations, beyond AFib. The system can detect Premature Ventricular Contractions, Premature Atrial Contractions, and wide Q wave: Initial negative deflection, R wave: First positive deflection, S wave: Negative deflection following the R wave intervals reduce the number of false positives and negatives, reduce unclassified readings, and offer better sensitivity and specificity. Table 2 shows the comparative analysis of different technologies with respect to their strength and limitations.

Clinic found that applying AI to a simple and cost-effective ECG monitoring device enables early and accurate detection of left ventricular dysfunction, which is a precursor to a serious heart condition. Left ventricular dysfunction could earlier be diagnosed only through an expensive, invasive process like computed tomography or magnetic resonance imaging scans.^[142,143] AI involving convolutional NN can be trained to assess the “physiologic age” of an individual, which can give information about the overall status of cardiac health status. This research in the future may foster an innovative area of science to comprehend human biological conditions in a better way.^[144,145] A recent study presented at the AHA’s Scientific Sessions 2019 in Philadelphia claims that AI has the potential to assess ECG to identify subjects who may be at a higher risk of developing arrhythmia even before it develops and can predict death in the near future. To quote Brandon Fornwalt, M.D., Ph.D., chair of the Department of Imaging Science and Innovation at Geisinger in Danville, Penn, “This is exhilarating and offers confirmation that we are on the threshold of a revolution in the field of medicine where computers will be employed alongside physicians to progress patient care.”^[146]

AI can be quoted to be “The secret weapon for Healthcare,” as its impact is going to be life-changing. Application of AI in healthcare has varied applications ranging from drug development, robotic surgery, to clinical research as depicted in Figure 5.^[147,148] Figure 6 shows the AI versus human diagnostics accuracy analysis.

Close

Figure 5: Artificial intelligence | Application areas in healthcare.

Export to PPT

Close

Figure 6: Artificial intelligence versus human diagnostic accuracy . (AI: Artificial Intelligence ECG: Electrocardiography)

Export to PPT

Reports from “Rock Health” claim that AI and machine learning have had a revenue generation of $2.7 billion for 121 organizations.

MEDICAL IOT MODEL

Just like modern technologies, AI and ML, the IoT is also finding a place in healthcare. IoT in Healthcare is a bridge between devices, sensors, humans, and systems and develops a connected digital ecosystem.^[11-16,31] IoMT has improved health outcomes and patient satisfaction in various ways such as creating a unified platform for various stakeholders, interoperability with electronic medical records and health information system (HIS), patient data security, drug administration, remote monitoring, preventive care, and reduced downtime. IOT has also enabled innovation in the field of medical care by building a data repository and offering a real-time setting for analyzing medical information. Figure 7 presents the technological features of IoMT which comprises the sensor and device layer, the network layer, the cloud, and the application layer.^[149]

Close

Figure 7: The Internet of Medical Things setup.

Export to PPT

The first layer comprises smart devices, implants, wearable sensors, smart wear, etc., that transmit raw data through the sensor nodes to the computing setup through the network layer using smartphones and developmental boards involving Arduino, Raspberry Pi, etc. The most widely used microcontrollers in IoMT are the ATmega328p microchip and STM32F107 from megaAVR and Cortex-M3 families, respectively, due to their performance rating and low power consumption. The software platforms used for IoMT applications are MATLAB, Python, Java, and C. The subsequent network layer includes various short and long-range communication protocols such as Bluetooth, WiFi, and cellular networks. Data are also transferred to the cloud through a Gateway. The most recent finding is the 4G network owing to its speed and bandwidth availability. The cloud layer comes in between the network layer and the end user due to the storage and computing facility that it offers. As enormous data is being collated from the IoT sensors, it is required these days to use fog/edge intermediate computing to expand the cloud. The final application layer is of the end user, the physician, or the patient. Figure 8 represents the different layers. Analysis of different layers as shown in Table 3.

Close

Figure 8: Representation of different layers. (PPG: Photoplethysmography, AI: Artificial Intelligence)

Export to PPT

INTEROPERABILITY

A major reform in the health care sector has been superior interoperability among the various IT systems so as to ensure improved workflow, easy transfer of patient e-information, and compatible data format outputs for physiological parameters such as ECG and electroencephalogram. With the recent technological advancements and trend toward user-friendly, compact, and mobile monitoring units, ease of connectivity through universal serial bus has become a prime requirement while selecting an ECG system. Many medical image data have standard formats like digital imaging and communications in medicine (DICOM) and Health Level Seven medical device communication standard formats. The ECG data format however still requires standardization which results in an array of formats leading to interoperability issue.^[150] However, vendors of ECG market are now making available systems that can address this issue.

Modern-day hospitals are extremely digitized these days with efficient IT infrastructure. In today’s time, physicians require accurate and quick access to patient information, which is transmitted using the hospital infrastructure including hospital information system (HIS), Laboratory Information System (LIS), Radiology Information System (RIS), etc. In the past, ECG data were stored using binary encoding. XML files are being explored to replace the binary format for encoding the ECG.^[88,151]

CONCLUSION

The spread of COVID-19 pandemic has impacted the number of CVD patients and so it is of utmost importance that we think beyond and guard our hearts. Recent post by “IDTechEx Research” on CVD talks about the trends and technologies in ECG and predicts that the CVD technology and devices market shall surpass $40 Billion by 2030. Newer technologies are on a constant increase to tackle the disease at every stage. Modern developments include cardiac rhythm management, bio-printing, and cardiovascular tissue generation. Scientific research has always linked weight with the risk of AFib, so keeping a check on our weight is advisable. It is also better to have a good night sleep to avoid Afib. Studies have revealed that sleep apnea has up to 400% of increased risk of cardiac arrest. Lack of exercise, improper diet, stress, anger, and depression also can impact the risk up to 500%. Recent study has revealed that the amount of Vitamin D is also associated with AFib. These are steps that one can take to enhance the quality of life and can also save lives.

This literature highlights the changing nature of the wearable ECG device as a complement and in certain situations, as a superiority of the more traditional Holter monitoring, but disparities in diagnostic validity, usability, and compliance persist (RQ1). AI-based algorithms show the ability to enhance ECG diagnosis and lessen diagnostic uncertainty (RQ2) but require strong IoMT infrastructures that guarantee real-time transmission of signals in a safe way (RQ3). There are still no standard practices and interoperability among the devices and health information systems which hinders clinical integration (RQ4). Remote monitoring has become increasingly popular due to the COVID-19 pandemic, which makes positioning wearables part of the cardiac care model of the future (RQ5). In general, the integration of wearable devices, AI, and interoperable IoMT systems will be an important factor in the development of scalable and patient-centered cardiac care.