Introduction
The integration of technology into healthcare has a potential to revolutionise the management and prevention of heart diseases. Wearable devices, mobile applications, and artificial intelligence (AI) is becoming to play an important role in monitoring heart health, offering real-time data and individualised insights.
This article describes in details the advancements in heart rate monitor accuracy, the efficacy of smartwatches for electrocardiogram (ECG) monitoring, the role of AI in cardiology, and the emergence of smart technologies in cardiac rehabilitation.
It also answers two interesting questions based on recent research articles:
- Can AI help with a Safe Discharge of patients presenting with Suspected Heart Attacks?
- Does using a Mobile Phone increase Your Risk of Heart Disease?
Advancements in Heart Rate Monitor Accuracy
Accurate heart rate monitoring is essential for assessing cardiovascular health, guiding exercise routines, and detecting potential anomalies early. Benjamin Franklin in the 18th century said that `An ounce of prevention is worth a pound of cure`. And indeed that`s the goal of preventive medicine to avoid any serious health issues occurring. Wearable devices equipped with photoplethysmography (PPG) sensors have become increasingly popular for continuous heart rate monitoring. However, their accuracy has been a subject of extensive research.
A study published in the Journal of Medical Internet Research evaluated the diagnostic accuracy of smartwatches in detecting cardiac arrhythmias. The meta-analysis, which included 18 studies with a total of 424,371 subjects, reported a pooled sensitivity of 100% and specificity of 95% for detecting arrhythmias using PPG-based smartwatches. Smartwatches demonstrated high performance in detecting cardiac arrhythmias, with an overall sensitivity of 100% (95% CI: 0.99–1.00), specificity of 95% (95% CI: 0.93–0.97), and accuracy of 97% (95% CI: 0.96–0.99). The pooled positive predictive value was 85% (95% CI: 0.79–0.90), while the negative predictive value was 100% (95% CI: 1.0–1.0), indicating strong reliability in ruling out arrhythmias. (https://pubmed.ncbi.nlm.nih.gov/34448706/)
As early as 2017, data published in a Journal of Personalised Medicine showed that most wrist-worn devices accurately measure heart rate during laboratory-based activities but provide unreliable estimates of energy expenditure. This highlights the need for caution when using energy expenditure data in health improvement programs. To address this, the authors suggested reference standards for validating consumer health devices. (https://pubmed.ncbi.nlm.nih.gov/28538708/)
Another study published in 2019 assessed the accuracy of various wrist-worn heart rate monitors compared to standard ECG measurements in patients undergoing cardiac rehabilitation. In patients with cardiovascular disease, the Apple Watch demonstrated clinically acceptable accuracy in measuring heart rate (HR) during exercise. If further validated, it could be considered a safe tool for HR-guided training in cardiac rehabilitation. However, at this time, it is premature to recommend its use for this purpose. Additionally, the Apple Watch consistently overestimates energy expenditure (EE) in this patient group, warranting caution when relying on it for EE measurements. (https://pmc.ncbi.nlm.nih.gov/articles/PMC6444219/)
Smartwatches for ECG Monitoring
The evolution of smartwatches has extended beyond fitness tracking to include advanced health monitoring features, notably ECG recording capabilities. These devices enable users to perform on-the-spot ECGs, facilitating early detection of conditions like atrial tachyarrhythmias, such as atrial fibrillation (AF) and atrial flutter. Atrial fibrillation is commonly referred to as irregular heart beat.
Atrial fibrillation contributes to a risk of a stroke with (:
- In the UK, over 1 million people are affected by atrial fibrillation (AF).
- Individuals with AF face a fivefold increased risk of stroke, with the condition contributing to 1 in 5 strokes nationwide.
- The prevalence of AF rises with age, and AF-related strokes tend to be more severe, leading to higher mortality rates and greater disability.
- Despite its serious consequences, AF often goes undiagnosed.
(https://www.stroke.org.uk/professionals/atrial-fibrillation-information-and-resources)
In a study published in 2020 Seshadri et al assessed the accuracy of the Apple Watch series 4 in detecting AF in 50 post-cardiac surgery patients who were on telemetry. Exclusion criteria included the presence of a cardiac pacemaker, prior surgical use of a radial artery, chronic heart rhythm abnormalities, and wrist tattoos. The Apple Watch series 4 offers two methods for rhythm assessment: on-screen notifications and a downloadable PDF waveform. While the watch display failed to notify AF in a significant number of cases, it did not produce any false-positive AF reports. Meaning that the AF wasn`t detected when it wasn`t present. Due to the limited number of rhythm assessments, the study could not identify factors contributing to missed AF notifications. However, the saved PDF waveform allowed for accurate heart rhythm classification, distinguishing between AF and sinus rhythm (SR). Overall, the single-lead traditional electrocardiographic waveform in the PDF proved to be a more reliable tool for AF detection than the watch’s rhythm notification. (https://www.ahajournals.org/doi/10.1161/CIRCULATIONAHA.119.044126?url_ver=Z39.88-2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%20%200pubmed)
In addition to this, a very large study, with over 400 000 participants, was published in the New Journal of Medicine in 2019 by Perez et al. The study participants had to have Apple iPhone and Apple Watch and no previous history of AF. Over 50% of participants were in the age between 22 and 39 years old and 42% were female. Over a median monitoring period of 117 days, 2,161 participants (0.52%) received irregular pulse notifications. The likelihood of receiving an irregular pulse notification was low. Among those who did receive a notification, 34% were found to have atrial fibrillation on subsequent ECG patch readings, and 84% of notifications were consistent with atrial fibrillation. (https://pubmed.ncbi.nlm.nih.gov/31722151/)
Finally, another large study published in 2022 by Lubitz et al in a cardiology journal called, Circulation concluded that a novel PPG software algorithm for wearable Fitbit devices demonstrated a high positive predictive value for detecting concurrent AF and effectively identified participants likely to have AF on subsequent ECG patch monitoring. These wearable devices may aid in the early detection of undiagnosed AF. A total of 455,699 participants enrolled between May and October 2020 (median age of 47 years and 71% female). (https://pmc.ncbi.nlm.nih.gov/articles/PMC9640290/)
Artificial Intelligence in Cardiology
Artificial intelligence is emerging as a potential transformative force in cardiology with a possibility of enhancing diagnostic accuracy, predicting patient outcomes, and personalising treatment plans. AI algorithms can analyse vast datasets, identifying patterns that may be imperceptible to human clinicians. These tools may be able to help clinicians make better decisions for their patients in the future.
In the realm of ECG interpretation, AI has showed a promising potential. A study published in Lancet in 2019 entitled: `An artificial intelligence-enabled ECG algorithm for the identification of patients with atrial fibrillation during sinus rhythm: a retrospective analysis of outcome prediction`, concluded that an AI-enabled ECG recorded during normal sinus rhythm allows for the point-of-care identification of individuals with atrial fibrillation. The authors developed an AI-enabled electrocardiograph (ECG) using a convolutional neural network to detect the electrocardiographic signature of atrial fibrillation during normal sinus rhythm from standard 10-second, 12-lead ECGs. They included over 180 000 patients with 649 931 normal sinus rhythm ECGs for analysis and over 3000 patients who had verified atrial fibrillation. (https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(19)31721-0/abstract)
AI’s application extends to the detection of cardiac arrhythmias. A study published in 2019 in Nature Medicine by Hannun et al is an example. The authors hypothesised that currently available digital ECG database and the algorithmic possibility of deep machine learning is an opportunity to improve the accuracy and scalability of automated ECG interpretation. The authors developed a deep neural network to classify twelve different rhythm (heart beat) classes based on over 90 000 ECGs from over 50 000 patients who used a single-lead ECG recording devices in the ambulatory setting. Interestingly, the authors found that the sensitivity of their deep neural network exceeded the average cardiologist sensitivity for all heart beat classes. The authors concluded that an end-to-end deep learning methods can classify a range of different arrhythmias (heart beat abnormalities) from single-lead ECGs with high diagnostic performance similar to that of trained cardiologists. It also showed that, this method could reduce the rate of misinterpreted automatic ECG interpretations and help to improve the expertise of human ECG interpretation by accurately prioritising the most serious conditions. (https://www.nature.com/articles/s41591-018-0268-3)
Artificial Intelligence can help with Medical Documentation
AI can be instrumental in automating clinical documentation. The increasing demands of medical documentation, especially with the widespread adoption of electronic health records, have significantly contributed to physician burnout and reduced job satisfaction. The complexity of medical notes, coupled with the time-consuming nature of data entry and many physicians’ limited typing proficiency, has hindered their ability to focus on patient care. However, the emergence of artificial intelligence in healthcare presents a promising solution to streamline documentation and enhance clinical workflows.
Artificial intelligence holds great promise in clinical documentation, offering automation of tedious tasks, streamlined workflows, enhanced insights, and relief from physician burnout. When implemented thoughtfully and ethically, AI can serve as an intelligent assistant, freeing physicians to focus more on patient care and relationships. However, as AI reshapes healthcare, it is essential to balance its ethical and responsible use with its potential to enhance the patient experience. The future of AI in physician documentation lies in improving efficiency, accuracy, and personalized care through intelligent and automated data processing.
(https://www.mcpdigitalhealth.org/article/S2949-7612(24)00041-5/fulltext)
Smart Technology in Cardiac Rehabilitation
Smart tech for cardiac rehabilitation and cardiac rehabilitation itself are important components of recovery for patients with heart diseases, aiming to improve physical fitness and reduce the risk of future cardiac events. The integration of smart technology into rehabilitation programs has enhanced patient engagement and adherence.
Cardiac rehabilitation is a comprehensive, evidence-based program designed to improve cardiovascular health through exercise training, health education, physical activity promotion, and extensive counselling to manage risk factors. Wearable devices play an emerging role in this field by continuously monitoring physiological functions and providing biometric data, including heart rate, movement, sleep patterns, ECG analysis, blood pressure, and energy expenditure. Recent research highlights the potential of wearable technology in cardiovascular risk assessment and disease prevention. According to a review article published in 2023 home-based cardiac rehabilitation utilising wearable devices can serve as a viable alternative or complement to centre-based CR for patients with coronary heart disease and heart failure. However, further research is needed to determine its effectiveness compared to centre-based CR in patients who have undergone heart valve repair or replacement. (https://pmc.ncbi.nlm.nih.gov/articles/PMC10231816/)
Mobile applications complement wearable devices by offering platforms for education, exercise tracking, and communication with healthcare providers. These apps can deliver tailored exercise regimens, send reminders for medication and appointments, and provide educational resources about heart health. The combination of wearable technology and mobile apps fosters a comprehensive approach to cardiac rehabilitation, promoting sustained lifestyle changes. According to a literature review, published in 2023 the European Heart Journal – Digital Health, virtual cardiac rehabilitation holds the potential to enhance patient adherence and sustain long-term engagement. If these outcomes can be effectively improved, they would provide strong justification for the adoption of this technology. (https://academic.oup.com/ehjdh/article/4/2/99/7033317)
Emerging Innovations in Wearable Technology
The landscape of wearable technology for heart health is continually evolving, with innovations enhancing the accuracy and functionality of these devices. One notable advancement is the development of smart rings, which can help to track our sleep patterns. This device offers continuous monitoring of heart rate, sleep patterns, and other vital metrics, all within a discreet and comfortable form factor. In 2022 an Australian study aimed to assess the validity of 6 widely available wearable devices, i.e., Apple Watch S6, Somfit, WHOOP 3.0 Polar Vantage V, Oura Ring Generation 2, and Garmin Forerunner 245 Music, for assessing sleep. The authors concluded that all six devices are valid for assessing sleep timing and duration in real-world settings. However, when evaluating multi-state sleep categorization, all six devices require further improvement in accurately distinguishing specific sleep stages. As wearable devices with validated sleep assessment capabilities become more widespread, they open new opportunities for addressing research questions that were once impractical or impossible. (https://www.mdpi.com/1424-8220/22/16/6317)
Can AI algorithm help with a Safe Discharge of patients presenting with Suspected Heart Attacks?
ARTEMIS-POC algorithm can identify more than twice as many patients eligible for Myocardial Infarction direct rule-out as compared to ESC or ACC pathways, while maintaining patients safety, according to a study published in Lancet Digital Health journal in October 2024.
It was a retrospective study, which included patients from the US and from Australia. Over 2.5k individuals were enrolled (42% females, median age of 58 years), prevalence of MI was 6.5%, STEMIs were excluded. Safety and outcomes of heart attacks direct rule-out by this algorithm were compared to ACC and ESC 0 h pathways using Point of Care high sensitivity Troponin I levels. The safety measure was a composite of heart related death and incidence of a heart attack at 30 days follow up.
The Authors’ INTERPRETATION in full: ‘The patient-tailored, medical decision support ARTEMIS-POC algorithm applied with a single POC hs-cTnI measurement allows for very rapid, safe, and more efficient direct rule-out of myocardial infarction than guideline-recommended pathways. It has the potential to expedite the safe discharge of low-risk patients from the emergency department including early presenters with symptom onset less than 3 h at the time of admission and might open new opportunities for the triage of patients with suspected myocardial infarction even in ambulatory, preclinical, or geographically isolated care settings.’
(https://www.thelancet.com/journals/landig/article/PIIS2589-7500(24)00191-2/fulltext)
AI-Driven Predictive Analytics in Cardiology
Artificial intelligence is not only enhancing diagnostic capabilities but also enabling predictive analytics in cardiology. As early as in 2017 a study published in PLOS One journal by Weng et al showed that, machine learning significantly enhances the accuracy of cardiovascular risk prediction, enabling more precise identification of patients who may benefit from preventive treatment while reducing unnecessary interventions for others.
In addition to this, AI-based predictive models can help with a diagnostic process in coronary artery disease and other cardiovascular disease. In the UK, cardiovascular disease (CVD) is a leading health concern, accounting for 26% of all deaths and two of the top five causes of mortality. More than seven million people in the UK live with CVD, and the impact is staggering—approximately every three minutes, someone dies from a cardiovascular condition, while every five minutes, someone is admitted to a hospital due to a stroke. (https://www.bhf.org.uk/what-we-do/our-research/heart-statistics)
It has proven to be a valuable tool in guiding treatment decisions, analysing complex genetic interactions, developing clinical risk prediction models, advancing cardiac imaging, and ultimately improving patient outcomes in CAD management. (https://pmc.ncbi.nlm.nih.gov/articles/PMC8872080/#sec4-healthcare-10-00232)
Does using a Mobile Phone increase Your Risk of Heart Disease?
Making at least one call a week using a cell phone is associated with a higher risk of Heart Disease, according to a study published in the Canadian Journal of Cardiology in November 2024. In comparison to infrequent users, regular users (at least one phone call per week) had a significantly higher risk of CardioVascular Disease (primary outcome – composite of coronary artery disease, stroke, heart failure and atrial fibrillation).
The above association was especially positive in current smokers and diabetic participants.
Interestingly, the study also demonstrated that that >5%, >11% and >2% of the above association was mediated by sleep patterns, psychological distress and neuroticism, respectively. It was a huge study, which included over 400k participants with mean age 56 years, 56% females and a follow up of over 12 years.
The authors concluded: ‘Weekly mobile phone usage time was positively associated with incident CardioVascular risk, which was partly explained by poor sleep, psychologic distress, and neuroticism.’ (https://onlinecjc.ca/article/S0828-282X(24)00437-9/fulltext)
The main limitations of this study were:
– observational data, therefore no cause/effect demonstrated
– other uses of mobile phones, watching and texting not studied
Conclusion
The convergence of technology and cardiology has ushered in a new era of heart health management. Advancements in wearable devices has a potential to improve heart rate monitor accuracy and introduce accessible ECG monitoring to the public. Artificial intelligence can enhance diagnostic precision and streamlines clinical workflows, while smart technologies in cardiac rehabilitation empower patients to take an active role in their recovery. As these innovations continue to evolve, they hold the promise of improving cardiovascular outcomes and transforming the landscape of heart health care in a long term.
