key: cord-0068009-26d01v5j
authors: Kulkarni, Kanchan; Sevakula, Rahul Kumar; Kassab, Mohamad B; Nichols, John; Roberts, Jesse D.; Isselbacher, Eric M; Armoundas, Antonis A
title: Ambulatory monitoring promises equitable personalized healthcare delivery in underrepresented patients
date: 2021-06-28
journal: Eur Heart J Digit Health
DOI: 10.1093/ehjdh/ztab047
sha: 61449deca990a9d861579910e11748257e2bdb4f
doc_id: 68009
cord_uid: 26d01v5j

The pandemic has brought to everybody’s attention the apparent need of remote monitoring, highlighting hitherto unseen challenges in healthcare. Today, mobile monitoring and real-time data collection, processing and decision-making, can drastically improve the cardiorespiratory–haemodynamic health diagnosis and care, not only in the rural communities, but urban ones with limited healthcare access as well. Disparities in socioeconomic status and geographic variances resulting in regional inequity in access to healthcare delivery, and significant differences in mortality rates between rural and urban communities have been a growing concern. Evolution of wireless devices and smartphones has initiated a new era in medicine. Mobile health technologies have a promising role in equitable delivery of personalized medicine and are becoming essential components in the delivery of healthcare to patients with limited access to in-hospital services. Yet, the utility of portable health monitoring devices has been suboptimal due to the lack of user-friendly and computationally efficient physiological data collection and analysis platforms. We present a comprehensive review of the current cardiac, pulmonary, and haemodynamic telemonitoring technologies. We also propose a novel low-cost smartphone-based system capable of providing complete cardiorespiratory assessment using a single platform for arrhythmia prediction along with detection of underlying ischaemia and sleep apnoea; we believe this system holds significant potential in aiding the diagnosis and treatment of cardiorespiratory diseases, particularly in underserved populations.

Socioeconomic and geographic variances between rural and urban communities result in inequitable regional access to healthcare and disparities in mortality rates. 1 While there are multiple factors driving the community-based divide in healthcare outcomes, 2 the geographic inequality in the risk of death has shown to be higher among the elderly. 3 Hence, there is a growing need to develop and implement an accessible and affordable healthcare paradigm that provides a framework for equitable personalized medicine.

The ever-increasing availability of new technologies and an everimproving health information technology infrastructure with >90% of American adults owning a cell phone, and 81% having a smartphone (elderly 53%, rural 71%, income < $30k 71%), 4 together with the evolution of wireless devices, has initiated a new era in medicine and a transition from population-level healthcare to personalized medicine. Therefore, mobile health technologies are expected to function not only as monitoring devices, but as essential components in the delivery of healthcare in patients with chronic diseases, 5 in underserved populations. 6 Furthermore, since a majority of medical consults today are accomplished through teleconference, with no access to vital sign data, the ability of a physician to have complete (medical grade) record of the cardiac-respiratory-haemodynamic state of patients with the help of mobile monitoring devices, has taken new, unforeseen significance. 7 Figure 1 demonstrates a schematic of the telehealth paradigm offering remote medical monitoring and consultation to persons with limited access to in-hospital services.

Ambulatory monitoring devices and telemedicine offer a promising model to even disparities in healthcare and provide equitable delivery of clinical aid to patients who have limited access to in-hospital services. Telehealth aims to utilize new smartphone-based technologies to provide medical support to rural underserved populations, as well as patients across all age groups, thus paving the way towards personalized medicine.

However, despite the widespread use of smartphones across multiple age groups, utility of portable health monitoring devices has been limited due to the lack of user-friendly, non-intrusive, and computationally efficient physiological data collection and analysis platforms. Here, we present a review of the state-of-theart of current wearable smart devices that enable cardiac, pulmonary, and haemodynamic telemonitoring and discuss the shortcomings, which highlight the need for novel tools to monitor the respiratory and cardiac state of ambulatory subjects with chronic diseases, that would aim to improve patient-provider communication, and adherence to treatment and self-management, in underserved populations.

In addition, we introduce a novel smartphone-based system capable of providing complete cardiac, respiratory, and haemodynamic monitoring in real-time by assessing various physiological parameters such as heart rate (HR), blood pressure (BP), pulse oximetry,

Telehealth equity for underserved patients . . respiration rate (RR), tidal volume (TV), minute ventilation (MV), underlying ischaemia, and arrhythmia susceptibility, in a single platform. The ability of this system to detect sleep apnoea, ischaemic events, and assess arrhythmia susceptibility make it a powerful tool that holds potential to aid diagnosis and treatment of cardiorespiratory diseases, particularly in underserved populations.

The Google Scholar database was manually searched without language or time restriction for articles published on wearable physiological monitors. Our search strategy focused on '[wearable devices OR smartphones AND (cardiac monitors OR respiratory monitors OR blood pressure monitors OR pulse oximetry monitors)]' and summarizes the current in-market, commercially available wearable monitoring technologies. Furthermore, to incorporate devices not covered in the Google Scholar academic search results, a manual search was performed for all commercial wearable products offering either cardiac, respiratory, blood pressure, or pulse oximetry monitoring. We present a list of wearables, capturing the most predominant vital sign monitors, currently used for fitness and healthcare assessment, covering devices with a broad range of utilities and form factors. From smart rings and watches primarily used for fitness tracking, to dedicated life vests for physician recommended life-saving medical-grade interventions, we provide an unbiased assessment of the advantages and pitfalls of using wearable devices for healthcare monitoring.

Cardiovascular diseases are the leading cause of mortality worldwide. 8 Despite the availability of a multitude of evidence-based therapies for the treatment of heart failure (HF), the burden of this chronic disease on the US population remains unacceptably high, with an estimated 1 M admissions per year, primarily involving the elderly. HF readmission rates constitute a marker of worse prognosis and represent a significant healthcare expenditure and performance measure for payers, 9 while efforts to reduce the rehospitalization burden using conventional markers have been largely ineffective. 10 A critical step to reducing HF hospitalizations is identifying which patients will imminently decompensate by evaluating their intravascular volume status. While physicians have traditionally relied on the bedside exam of the jugular venous pressure to assess volume status, this is impractical for ambulatory patients. Implantable haemodynamic monitors have shown promise as early warning systems in HF, 11 however, they are invasive and expensive, and their cost-effectiveness remains debatable. 12 Figure 1 Schematic of a proposed telehealth monitoring system. Ambulatory monitoring of vital signals like the heart rate, blood pressure, oxygen saturation, respiratory rate, tidal volume, minute ventilation, sleep apnoea, and ischaemic or arrhythmic events is performed using a single platform telemonitoring device like the cvrPhone. Medical data are uploaded to cloud servers and transmitted to healthcare facilities as an electronic medical record. Healthcare professionals and physicians can access the data and provide clinical diagnosis along with necessary prescriptions for palliative care or recommendations for further treatments. The system provides a framework for remote monitoring and offers equitable personalized medical care for persons in rural communities and underrepresented areas. ECG, electrocardiograph; LA, left arm; LL, left leg; RA, right arm; RL, right leg.

.

especially those with milder symptoms, occur suddenly and unexpectedly. Many of these are due to abnormal heart rhythms, including ventricular arrhythmias, e.g. bradycardia and asystole. An FDA-approved electrocardiographic (ECG) patch monitor named Zio Patch, which can be worn for up to 14 days has shown improved clinical event detection compared with the conventional 24 h Holter monitor. 13 However, it offers only one lead which explains its significantly lower sensitivity compared to the Holter monitor. The NUVANT, a wireless arrhythmia event monitor 14 consists of a wearable (patch) monitoring device and a portable data transmission device. Unlike the Zio Patch, which records and stores all ECG data for retrospective arrhythmia detection, the NUVANT performs real-time analysis and transmission. 14 The Bio patch Vitalerter 15 is designed for health monitoring in care homes; it generates the complete set of vital signs, including HR, and transmits them wirelessly, to alert the staff when a resident needs assistance. Multiple other smartwatches and wearable devices are available in the market ( Table 1) , which enable 24 h vital monitoring, namely the Apple watch, Motiv Ring, Oura Ring, Moov, and Fitbit, albeit without the specificity offered by the 12-lead ECG and hence, with limited ability to detect ischaemic events or life-threatening arrhythmias. 17 A smartphone-based, wireless, single-lead real-time ECG monitoring system (AliveCor Kardia) may accurately detect abnormal atrial rhythms. 18 However, the aforementioned devices 13, 14, 18 cannot provide an assessment of the patient's respiratory state, such as in-sleep disordered breathing (SDB). The Zephyr BioHarness and the Garmin Vivoactive can monitor both, pulmonary as well as cardiac activity, but they are primarily athletic performance monitoring systems, and provide limited assessment on patient deterioration by failing to identify dizziness signals, issue alerts in situations of hypoxaemia or hyper/ hypothermia, or detect complex arrhythmias and ischaemic events. It may be noted that none of these devices (which determine HR and heart rhythm) provide medical-grade 12-lead ECGs. Since one cannot diagnose myocardial ischaemia (MI)-associated ST-elevation without 12-lead ECG, the US National Heart Attack Alert Program recommends that Emergency Medical Systems provide out-of-hospital 12-lead ECG to diagnose acute MI, and that all advanced life-saving vehicles be able to transmit the 12-lead ECG signal to the hospital. 19 It is therefore imperative that smartphone-based diagnostic devices, in the future, be capable of recording and transmitting 12lead ECG to the treating medical team.

We have recently developed a novel smartphone-based ambulatory monitoring system, cvrPhone, with the capability to perform complete cardiorespiratory assessment based on 12-lead body surface ECG. [20] [21] [22] Figure 2A demonstrates the ability of the cvrPhone to detect the onset of ventricular tachycardia (VT). 20 It has been suggested that repolarization alternans (RA), an alternation in the Twave morphology and/or duration, may be a precursor to VT and serve as a short-term 23,24 and long-term 25 predictor of susceptibility to ventricular tachyarrhythmias. Our smartphone-based system can detect and quantify the onset of RA and the corresponding K score (a statistical measure of alternans, accounting for the backgrounding noise levels), 24 as depicted in Figure 2B , where a significant rise in RA levels is observed in all leads, just prior to the onset of VT. 20 These results demonstrate the ability of cvrPhone to accurately detect the onset of RA, demonstrated by a significant rise in RA in each lead (P < 0.05) within 1 min of occlusion compared to baseline (n = 29 swine). 26 In addition, we tested the utility of the cvrPhone in detecting the onset of ischaemia using an MI swine model. Figure 3A depicts the instantaneous beat-to-beat change in ischaemic index over time, with t = 0 marking the time of MI induction by coronary artery occlusion. Summary results depicted in Figure 3B show a significant rise in ischaemic index across all leads after MI induction. 22 Similarly, the ischaemic index which significantly increases following myocardial infarction (P < 0.05) and preceding a tachyarrhythmic event, can also be successfully captured by cvrPhone. Hence, cvrPhone enables the continuous monitoring of ST-segment changes underlying acute closure of a coronary artery, which has been shown to be essential in reducing symptom-to-door time and improve outcomes. 26 Telehealth systems for respiratory monitoring Assessment of the RR and TV are integral components of patient monitoring in chronic disease management, and especially in monitoring of patients with Cheyne-Stokes respiration, 27 a form of SDB. SDB is probably the most common respiratory disorder, with recent data from the USA and Europe suggesting that between 14% and 49% of middle-aged men have clinically significant sleep apnoea. 28 Sleep apnoea has been associated with the prognosis of many diseases including HF, 27 obstructive pulmonary disease, 29 sudden infant death syndrome, 30 atrial fibrillation, 31 chronic kidney disease, 5 diabetes, 32 hypertension, 33 and obesity. 34 The need for medical-grade real-time mobile respiratory monitoring is not limited to the adult population. Premature infants routinely remain hospitalized overnight for clinical apnoea monitoring after receiving general anaesthesia for procedures that would otherwise not require hospitalization. Additionally, 2.25 million children in the USA suffer from obstructive sleep apnoea (OSA) and the gold standard remains overnight polysomnography. 35 Unfortunately, only 50% of children who snore actually suffer from OSA, and thus more effective screening tools could be invaluable. A mobile, real-time apnoea monitoring device could provide significant cost savings in the paediatric population by potentially reducing both the number of post-anaesthesia hospital admissions in infants and in-hospital sleep studies in children.

In a hospital setting, measurement of RR/TV can be accomplished either directly or indirectly, 36 using specialized hardware with features that are not often practical and convenient for ambulatory subjects, especially those with chronic conditions that live in underserved areas. Traditional TV estimation methods incorporate a spirometer or pneumotachometer that is physically attached to the mouth or nose of the patient, making it highly inconvenient for ambulatory measurement. While newer techniques for TV measurement including pitot tubes, 39 ultrasonic airflow meters, 38 respiratory inductance plethysmography, 39 electrical impedance tomography, 40 differential pressure pneumatachographs, 41 and impedance pneumography 42 have been suggested, these techniques require specialized hardware and have restricted telehealth application. Furthermore, the accuracy of pneumotachographs in TV estimation is limited 43 and most of these techniques are inadequate in detecting dynamic respiratory events such as apnoeas. 44 MV is another important respiratory parameter that plays an important role in the assessment of a patient's pulmonary activity. The most direct method of MV estimation has been from regression analysis of the HR 45 or correlations of the HR with oxygen consumption. 46 Yet these models perform poorly with patient-to-patient variability, leading to large errors, limited clinical utility, 47 and have little applicability in the ambulatory setting.

We tested the utility of our smartphone-based system, cvrPhone, in performing pulmonary assessment by estimating the RR, TV, MV, and onset of apnoeic events in real-time (n = 9 swine). 21 Figure 4 depicts a representative plot of RR, TV, and MV estimation as the true RR was varied from 6 to 0 to 6 to 0 to 6 to 0 to 6 breaths/min, TV was varied from 250 to 0 to 750 to 0 to 500 to 0 to 750 mL, and MV was varied from 1500 to 0 to 4500 to 0 to 3000 to 0 to 4500 breaths*mL/min (n = 9 swine). The smartphone-based system accurately estimated TV at different settings (0, 250, 500, and 750 mL) with statistically significant difference (P < 0.01) between any two different settings regardless of the RR (6 or 14 b.p.m.). 21 During apnoea, the estimated TV and RR values were 11.7 ± 54.9 mL and 0.0 ± 3.5 breaths/min, which were significantly different (P < 0.05) than TV and RR values during non-apnoea breathing. In addition, the time delay from the apnoea onset to the first apnoea detection was 8.6 ± 6.7 and 7.0 ± 3.2 s for TV and RR, respectively. 21 Furthermore, the cvrPhone was also accurately able to determine MV at varying TVs and RRs with median relative estimation errors of 17%, -4%, 35%, -3%, -9%, and 1%, for true MVs of 1500, 3000, 3500, 4500, 7000, and 10 500 breaths*mL/min, respectively (n = 9 swine). 48 These studies provide proof-of-concept of the utility of cvrPhone to detect multiple cardiorespiratory parameters from 12-lead surface ECG in real-time with high accuracy and relatively low error, compared to the gold standard mechanical ventilator (Ohmeda-GE, Madison, WI, USA).

Use of pulse oximetry is a clinically standardized method for non-invasively assessing the pulse oxygen saturation levels, SpO 2 . Based on changes in the optical absorption properties of haemoglobin, non-invasive SpO 2 measurements are typically spectrophotometric and performed in the visible and near-infrared spectral regions. 49 Pulse oximetry uses photoplethysmographic (PPG) signals detecting volumetric changes in blood flow that correlate with cardiac contraction and relaxation. Multiple techniques for monitoring PPG-based SpO 2 have been suggested, utilizing either finger-based 50 or forehead oximeters, 51 offering wireless 52 monitoring of SpO 2 levels. Furthermore, smartphones with optical sensors have now enabled instantaneous detection of SpO 2 in conjunction with other vital physiological parameters in mobile and remote settings. 53 The use of smartphone sensors and in particular the camera function for health monitoring has gained considerable popularity over the past few years. Multiple studies have shown the utility of acquiring a PPG like pulsatile signal from a fingertip placed in contact with a 49 (A) Representative plot of respiration rate estimates acquired using cvrPhone, as true respiration rate was varied between 0 and 6 breaths/min. (B) Representative plot of tidal volume estimates acquired using cvrPhone, as true tidal volume was varied from 250 to 0 to 750 to 0 to 500 to 0 to 750 mL. (C) Representative plot of minute ventilation estimates as true minute ventilation was varied from 1500 to 0 to 4500 to 0 to 3000 to 0 to 4500 breaths/min. Red lines denote true respiration rate, tidal volume, or minute ventilation values.

Telehealth equity for underserved patients

smartphone camera for assessing the HR, HR variability, RR, and SpO 2 levels. [54] [55] [56] [57] [58] Also, various signal processing algorithms and mobile apps are being developed to extract health parameters from smartphone camera signals and test its utility in medical diagnosis. 21, 48, 59, 60 For example, Lagido et al. 60 demonstrated that HR can be accurately estimated using the PPG signal acquired from a smartphone camera and in addition atrial fibrillation can be detected with a specificity of 97% and sensitivity of 75%, in patients with HF. Similarly, a recent prospective clinical study (DETECT AF PRO), on detection of atrial fibrillation using smartphone cameras demonstrated high specificity (99.6%) and sensitivity (91.5%). 61 However, while such modalities promised to provide basic cardiac assessment, the utility of smartphone camera-based apps for comprehensive clinical diagnosis has not yet been established. For instance, while smartphone pulse oximetry was shown to not be inferior to traditional probe-based oximetry in healthy children, 53 SpO 2 measurement has been shown to be less reliable in hypoxaemia patients. 62 Similarly, many smartphone apps have been shown to be less accurate and thus provide less reliable estimates, 63 questioning the clinical utility of these approaches. 64, 65 Commercially available wearable measurement devices like the Masimo MightySat offer user-friendly fingertip-based monitoring of SpO 2 and the pulse rate (Table 1) . However, they lack comprehensive cardiac monitoring to detect abnormal rhythms and ischaemic events. The VitalPatch, EQ02þ Lifemonitor, and VinCense WHMS devices offer continuous 24 h clinical-grade pulse oximetry monitoring, but with similar limitations of low-resolution ECG measurements, making them insufficient to assist at-risk cardiovascular patients and detect early warning signs of cardiopulmonary events. The Everion system promises a more complete physiological assessment by monitoring 22 vital signs, but is currently pending FDA approval. A convenient wristband-based SpO 2 monitoring device, Spryhealth Loop, has been designed for use by seniors, offering easy assessment of HR and respiration in addition to pulse SpO 2 levels. Finally, Current Health has designed a mobile monitoring device for continuous SpO 2 , temperature, HR, respiration, and movement measurements, for detection of warning signs of health deterioration.

Despite the surge in pulse oximetry devices for mobile monitoring, there is yet an unmet need for implementing a unified solution that provides comprehensive cardiorespiratory monitoring, in addition to pulse oximetry measurements using a single low-cost platform to aid remote patient surveillance. cvrPhone, in addition to cardiac and respiratory monitoring, offers the capability of continuous SpO 2 estimation using two-channel PPG signals. 50 We evaluated the feasibility of the system to record continuous, instantaneous changes in SpO 2 in healthy humans (n = 11). 50 Figure 5 demonstrates summary results of SpO 2 estimation using different fingers with the corresponding root mean square error values, indicating efficient detection of SpO 2 using either the index, middle, or ring fingers with no significant differences between fingers, 50 albeit the right hand ring finger exhibited the smallest error. Figure 5 Instantaneous blood oxygen saturation estimation. 51 Representative photoplethysmography and oxygen saturation signals acquired by the cvrPhone. Quantification of oxygen saturation estimation error (n = 11), acquired using different fingers (index, middle, and ring fingers), demonstrating no significant difference among fingers/hands. P-value is calculated using Wilcoxon signed-rank test. Data are presented as bar graphs denoting 10, 25, 50, 75, and 90 percentiles of oxygen saturation estimation error. PPG, photoplethysmographic; RMSE, root mean square error (%); SpO 2 , oxygen saturation.

Regular monitoring and analysis of the historical patterns of BP is essential for the detection and prediction of hypertension and subsequent cardiac arrhythmias. Portable/wearable health monitoring equipment unequivocally strikes a good balance between recording accuracy and user comfort. Notable examples for portable, continuous non-invasive BP monitoring include (i) automated oscillometric methods, 66 (ii) ultrasonic device-based methods, 67 and (iii) softwarebased BP estimation methods. 68 Oscillometric methods which use cuffs are the most commonly used devices; the difficulty with them, however, is that (i) they are uncomfortable to wear and (ii) they have an inherent error with respect to their ability in representing the arterial BP values. 69 The feasibility in estimating BP from the SpO 2 and ECG signals has been previously explored with little success 70 and ultrasonic devices are yet to become mainstream in continuous BP monitoring.

The Omron HeartGuide is a commercially available wristwatchbased BP monitor with miniaturized components for oscillometric measurement using an inflatable cuff within the watch band. It also tracks sleep and activity in real-time, offering a portable user-friendly solution for BP sensing, albeit on the costlier side. Table 1 lists some of the other currently available commercial wearable BP monitoring Figure 6 Beat-to-beat real-time blood pressure estimation. Representative, noise-free signals. A beat here corresponds to: (i) end of prior beat's T-wave to the end of current beat's T-wave, (ii) diastolic to diastolic, for blood pressure, and (iii) minimum to the next beat's minimum value for oxygen saturation. ABP, arterial blood pressure; ECG, electrocardiogram; SpO 2 , oxygen saturation. devices. A new cuffless wristband-based BP monitor developed by Biobeat is still awaiting mass production, while BPM Connect Withings offers a leadless cuff-based, smart, WiFi-enabled device but without the ability for continuous monitoring. Similarly, Care Touch offers a fully automatic wristband-based cuff monitor for intermittent tracking of BP and pulse rate, unlike the VisiMobile or Accurate 24, which provide continuous monitoring of vital signs.

We developed a BP estimation method that uses ECG and SpO 2 signals to estimate in real-time, the systolic and diastolic BP values on an almost beat-to-beat basis. Following selection of noise or artefactfree four-lead ECG, SpO 2 , and arterial BP signals from 281 intensive care unit patients, R-wave peak detection, and application of ECG waveform delineation methods, individual heartbeats and their corresponding span ( Figure 6) were obtained, resulting in 22 M beats, i.e. 4 .52 M five-beat sequences. The age and body mass index, obtained from the electronic medical records, were used as features, together with features extracted from the SpO 2 and four-lead ECG signals. These additional futures include, features characterizing the HR and its variability, the morphology of ECG signals, the duration of the QT interval, the amplitude of the T-wave, the pulse arrival time, the power spectral distribution of SpO 2 , and statistical measures over the Kaiser-Teager energy of SpO 2 signal, across windows of five consecutive beats.

With these features as the input, and the corresponding arterial waveform systolic/diastolic BP values as gold standard output, a random forest model with 400 trees was used to fit a regression model for future BP estimation. The performance of the method was assessed using five-fold cross-validation, and resulted in a mean absolute error of 4.37/2.49 mmHg, respectively, in estimating the systolic/ diastolic BP. We therefore believe that this technology can be generalized and likely enable the BP monitoring of ambulatory patients.

In a hospital setting, measurement of RR, TV, MV, SpO 2 , BP, assessment of sleep apnoea, ischaemia, and the risk of abnormal heart rhythms is accomplished either directly or indirectly, using specialized hardware and procedures with features that are often impractical and inconvenient for free-moving, ambulatory subjects, especially those with chronic conditions that live in the underserved areas. cvrPhone, encompasses critical and essential features for cardiorespiratory assessment that can drive medical decision-making, offering an alternative, efficient modality for mobile monitoring. Using the cvrPhone (Galaxy S8, 6 GB RAM, 64 GB expandable memory) we have been able to estimate the RR, TV, MV; detect ischaemia and apnoeic events; and predict cardiac arrhythmias, in real-time. The choice of the Android was made to enhance cost efficiency with the first prototype built with over-the-shelf parts at a cost of <$100. Furthermore, we believe that since an ever-increasing percentage of the population worldwide will be using smartphones or tablets in the near future, even in rural and underserved areas, cvrPhone has the potential to be an affordable, medical-grade cardiac, pulmonary, and haemodynamic monitoring solution for a wide section of the population. A further comprehensive clinical assessment of this novel system holds promise for aiding telehealth and providing personalized medicine in underrepresented populations.

The successful implementation of a telehealth paradigm involves addressing multiple shortcomings in the current telemonitoring modalities: (i) reliability-construction of efficient, accurate, real-time algorithms that are constantly refined with new data and signal processing techniques; (ii) accountability-while developers are responsible for the quality and performance of any device or algorithm, adequate training for physicians and personnel handling the data streams from the devices are warranted for accurate diagnosis; (iii) legislation and liability-involvement of all stakeholders from developers, clinicians, researchers to participant population is necessary for the successful and safe translation of a product from development, testing, FDA approval, and mass production to end consumer marketing; and (iv) data security-protection of personal patient data recorded on cloud servers and transferred via various IoT platforms needs to be ensured.

Despite evidence that remote patient monitoring and telehealth can improve patient-provider communication in remote residential populations 71 and rural healthcare settings, 72 there is a paucity of medical grade, affordable tools to monitor the respiratory and cardiac state of ambulatory patients with chronic conditions in such communities, and thus provide equitable patient-specific healthcare. Successful implementation of a telehealth paradigm requires reliable, accountable, secure, and accurate real-time remote monitoring devices. In the long term, precision medicine that employs artificial intelligence to provide clinicians with tools to enable best treatment predictions, 73 when applied to streaming vital sign-signals, is expected to impact positively all population groups, and eliminate disparities in such populations as racial/ethnic minorities, structurally disadvantaged population subgroups, sex, and sexual identity minorities, people with different levels of functional ability, rural residents, as well as other marginalized groups. 74 Funding This work was supported by a Grant-in-Aid [#15GRNT23070001] from the American Heart Association (AHA), the Institute of Precision Medicine [17UNPG33840017] from the AHA, and the RICBAC Foundation, NIH grants [1 R01 HL135335-01, 1 R21 HL137870-01, 1 R21EB026164-01, and 3R21EB026164-02S1]. This work was conducted with support from Harvard Catalyst, The Harvard Clinical and Translational Science Center (National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health Award 8UL1TR000170-05 and financial contributions from Harvard University and its affiliated academic healthcare centres). The content is solely the responsibility of the authors and does not necessarily represent the official views of Harvard Catalyst, Harvard University and its affiliated academic healthcare centres, or the National Institutes of Health.

Conflict of interest: none declared.

Data will be available to any investigator upon request.

Growth and persistence of place-based mortality in the United States: the rural mortality penalty

Closing the health equity gap: evidence-based strategies for primary health care organizations

Inequalities in life expectancy among us counties, 1980 to 2014: temporal trends and key drivers

Pew research internet project

A systematic review of central sleep apnea in adult patients with chronic kidney disease

Wearable devices for ambulatory cardiac monitoring: JACC state-of-the-art review

Secondary impact of the covid-19 pandemic on patients with heart failure

Sudden cardiac death and arrhythmias

2009 focused update incorporated into the ACC/ AHA 2005 guidelines for the diagnosis and management of heart failure in adults a report of the American College of Cardiology Foundation/American Heart Association Task Force on practice guidelines developed in collaboration with the International Society for Heart and Lung Transplantation

Ongoing right ventricular hemodynamics in heart failure: clinical value of measurements derived from an implantable monitoring system

Pulmonary artery pressure-guided management of patients with heart failure and reduced ejection fraction

Heart failure management with ambulatory pulmonary artery pressure monitoring

Comparison of 24-hour Holter monitoring with 14-day novel adhesive patch electrocardiographic monitoring

Study of arrhythmia prevalence in NUVANT mobile cardiac telemetry system patients

Bio patch vitalerter

Electrocardiographic patch devices and contemporary wireless cardiac monitoring

Wearable health devices-vital sign monitoring, systems and technologies

Wireless smartphone ECG enables large-scale screening in diverse populations

Heart Association in collaboration with International Liaison Committee on Resuscitation. Guidelines 2000 for cardiopulmonary resuscitation and emergency cardiovascular care: International Consensus on science, part 7: the era of reperfusion

Utility of a smartphone based system (cvrPhone) to predict short-term arrhythmia susceptibility

Utility of a smartphone based system (cvrPhone) to accurately determine apneic events from electrocardiographic signals

A novel point-of-care smartphone based system for monitoring the cardiac and respiratory systems

Cardiac alternans: mechanisms and clinical utility in arrhythmia prevention

Real-time closed-loop suppression of repolarization alternans reduces arrhythmia susceptibility in vivo

Prospective use of microvolt t-wave alternans testing to guide primary prevention implantable cardioverter defibrillator therapy

Initial clinical results using intracardiac electrogram monitoring to detect and alert patients during coronary plaque rupture and ischemia

Cheyne-Stokes respiration during sleep in congestive heart failure

Sleep apnoea as an independent risk factor for cardiovascular disease: current evidence, basic mechanisms and research priorities

Metabolic and respiratory variables during pressure support versus synchronized intermittent mandatory ventilation

Prenatal cigarette smoke exposure attenuates recovery from hypoxemic challenge in preterm infants

Obstructive sleep apnoea and atrial fibrillation

Diabetes mellitus prevalence and control in sleep-disordered breathing: the European Sleep Apnea Cohort (ESADA) study

Hypertension and sleep apnea

The epidemiology of adult obstructive sleep apnea

Childhood obstructive sleep apnea syndrome: a review of the 2012

Estimation of respiratory rate from photoplethysmogram data using time-frequency spectral estimation

Feasibility and validation of telespirometry in general practice: the Italian "alliance" study

Measurement of lung volume and ventilation distribution with an ultrasonic flow meter in healthy infants

Accuracy of respiratory inductive plethysmograph in measuring tidal volume during sleep

Evaluation of a novel noninvasive respiration monitor providing continuous measurement of minute ventilation in ambulatory subjects in a variety of clinical scenarios

Errors in pneumotachography with intermittent positive pressure ventilation

Ambulatory impedcardiac applications

Changes in respiratory pattern resulting from the use of a facemask to record respiration in newborn infants

Measuring tidal volume and functional residual capacity change in sleeping infants using a volume displacement plethysmograph

Dynamic ventilation and power output of urban bicyclists

Textbook of Work Physiology: Physiological Bases of Exercise

Minute ventilation of cyclists, car and bus passengers: an experimental study

Utility of a smartphonebased system (cvrPhone) in estimating minute ventilation from electrocardiographic signals

A novel time-varying spectral filtering algorithm for reconstruction of motion artifact corrupted heart rate signals during intense physical activities using a wearable photoplethysmogram sensor

Design implementation and evaluation of a mobile continuous blood oxygen saturation monitoring system

Comparison of desaturation and resaturation response times between transmission and reflectance pulse oximeters

A patch-type wireless forehead pulse oximeter for SpO 2 measurement

Accuracy of smartphone-based pulse oximetry compared with hospital-grade pulse oximetry in healthy children

Monitoring of heart and breathing rates using dual cameras on a smartphone

Physiological parameter monitoring from optical recordings with a mobile phone

Respiratory rate estimation from the built-in cameras of smartphones and tablets

Statistical analysis of heart rate and heart rate variability monitoring through the use of smart phone cameras

Annual International Conference of the IEEE Engineering in Medicine and Biology Society

Smartphone sensors for health monitoring and diagnosis

A pulse rate estimation algorithm using PPG and smartphone camera

Using the smartphone camera to monitor heart rate and rhythm in heart failure patients

Detection of atrial fibrillation with a smartphone camera: first prospective, international, two-centre, clinical validation study (DETECT AF PRO)

Reliability of smartphone pulse oximetry in subjects at risk for hypoxemia

Validation of the instant blood pressure smartphone app

Comparison of smartphone applicationbased vital sign monitors without external hardware versus those used in clinical practice: a prospective trial

Smartphone-based arrhythmia detection: should we encourage patients to use the ECG in their pocket?

Blood pressure monitoring: automated oscillometric devices

Monitoring of the central blood pressure waveform via a conformal ultrasonic device

A SVM method for continuous blood pressure estimation from a PPG signal

Accuracy of cuff-measured blood pressure: systematic reviews and meta-analyses

Wrist-worn blood pressure tracking in healthy free-living individuals using neural networks

How effective is the virtual primary healthcare centers? An experience from rural India

Evaluating the feasibility and acceptability of a telehealth program in a rural palliative care population: TapCloud for palliative care

The new era of precision population health: insights for the all of us research program and beyond

Ensuring fairness in machine learning to advance health equity