Translate this page into:
Multi-sensor wearables re-shaping care of chronic heart-failure: A narrative review
Present address: †Lee Business School, University of Nevada, Las Vegas, United States
$Alphacrucis University College, Sydney, Australia
For correspondence: Dr Enibokun Theresa Orobator, College of Medicine and Veterinary Medicine, University of Edinburgh, Edinburgh EH16 4SB United Kingdom e-mail: theresaorobator@gmail.com
-
Received: ,
Accepted: ,
Abstract
Heart-failure decompensation often evolves between visits, driving recurrent admissions and cost. Multi-sensor wearables, adhesive patches, smart watches, and garments capture electrocardiography, thoracic impedance, photoplethysmography, respiration, activity, and speech for near-real-time review. Evidence synthesis indicates two signals. First, integrated telemonitoring, combined with structured clinical intervention during the post-discharge vulnerable phase, reduces cardiovascular events and heart failure hospitalisations in randomised programmes. Second, device-level performance metrics (lead time, alert burden, detection accuracy) demonstrate early warning capability but do not alone establish outcome benefit. Implantable multi-sensor algorithms offer a median lead time of approximately one month, accompanied by manageable alert rates. External systems that estimate lung fluid or fuse wearable signals demonstrate promising feasibility, although large pragmatic trials remain limited. Consumer smartwatches achieve high accuracy for detection of atrial fibrillation in general population; however, evidence is lacking for detection of heart failure decompensation, or for improving the outcome. Key implementation issues include AI-validation, workflow-linked triage, cost-effectiveness, cybersecurity, and equitable access for low- and middle-income settings. Multi-sensor monitoring warrants targeted deployment within intervention pathways and rigorous evaluation focused on patient-important outcomes.
Keywords
AI-driven HF readmission prevention
digital health in cardiology
heart failure remote monitoring
multi-sensor wearable devices
thoracic impedance congestion detection
Heart failure is a global epidemic with increasing prevalence and healthcare burden. Simple telemonitoring based on telephone transmission of weights and symptoms did not reduce death or rehospitalisation in large trials1. In contrast, programmes that pair remote physiologic data with protocolised tele-intervention during the vulnerable post-discharge phase have reduced the composite of cardiovascular death or worsening heart failure at six months [n=506; HR (hazard risk) 0.35, 95% CI 0.24–0.50]2. Multi-sensor wearables: patches, watches, and smart textiles capture electrocardiography, photoplethysmography, thoracic impedance, respiration, and activity, transmitting data to clinician dashboards in near real time. Smartwatch atrial-fibrillation detection is accurate in general population, but evidence does not support heart failure decompensation detection or improved outcomes of heart failure3. This review synthesizes physiologic targets, the sensor/analytics landscape, and clinical effects of multi-sensor wearables in chronic heart failure, distinguishes device-performance metrics from patient-important outcomes, and identifies implementation constraints.
Materials & Methods
Design and sources
A comprehensive narrative review was conducted to synthesise evidence on multi-sensor wearable technologies for the care of chronic heart failure. Searches were executed on April 30, 2025, in PubMed/MEDLINE, Embase.com, CINAHL Complete, and IEEE Xplore. The representative PubMed search strategy included the following terms: (“heart failure”[MeSH] OR “cardiac failure” OR CHF) AND (wearable* OR telemonitor* OR “remote monitoring” OR multisensor) AND (ECG OR impedance OR photoplethysmography OR accelerometer). Equivalent controlled vocabulary and syntax were adapted per database. Only studies in the English language were included. Records between January 1, 2019 to April 30, 2025, were retained. Reference-list snowballing was used to identify additional studies.
Eligibility
Included studies enrolled adults (≥18 yr) with established chronic heart failure, used an external wearable capturing ≥2 physiologic signals, and transmitted data to clinicians. Eligible designs were randomised trials, prospective/retrospective cohorts, and published systematic reviews.
Exclusions
Purely technical validations without clinical endpoints, implantable-sensor studies, single-signal monitoring, conference abstracts, and duplicate analyses.
Screening and data extraction
Two independent reviewers screened titles/abstracts, with full-text assessment in duplicate; disagreements were resolved by consensus. Extracted items included study design, sample size, sensor suite, follow-up, primary endpoints, usability measures, and economic data.
Synthesis
Findings were organised narratively under four themes: physiologic targets, technology landscape, clinical outcomes, and emerging signals. No de novo meta-analysis was undertaken; effect estimates and precision were reported as in the source studies. A prespecified distinction was maintained between device performance metrics (e.g., lead time, alert burden) and patient-important outcomes.
Results
Pathophysiological targets monitored by wearables
Congestion and volume status
Thoracic bioimpedance, seismocardiography, and radio-frequency dielectric sensing (ReDS) enable non-invasive estimation of pulmonary fluid and intrathoracic congestion. In a pre-post cohort of patients recently hospitalised for heart failure, ReDS-guided therapy reduced three-month readmissions versus usual care (observational design)4. Multi-sensor algorithms in implantable cardiac devices extend this concept: in the Multisense validation, the HeartLogic™ index (heart sounds, thoracic impedance, respiration, night HR, and activity) detected ⁓70 per cent of worsening events with a median 34-day advance notice and an unexplained alert rate of ⁓1.47 alerts per patient-year (threshold 16)5. Real-world registries report similar alert burdens (⁓0.4–1.0 alerts/patient-year) and workflow actions in ⁓40 per cent of alerts, while randomised outcome trials remain ongoing (pre-post post-reductions in heart failure hospitalisations have been described; causality is unproven)5. Daily weight remains commonplace but shows limited clinical impact when used alone in telemonitoring programmes1.
Arrhythmia surveillance
Continuous single-lead ECG patches and smartwatch ECGs uncover silent atrial fibrillation and ectopy. In cryptogenic-stroke populations, implantable loop recorders detected atrial fibrillation in 12.4 per cent at 12 months versus 2 per cent with conventional follow-up, illustrating the advantage of continuous sensing (the population differs from heart failure, but the mechanism generalises)6. High premature atrial contraction burden (>500/24 h) signals increased AF risk, supporting proactive rhythm surveillance in high-burden phenotypes7. Rhythm data complements congestion metrics to inform anticoagulation or rate-control adjustments.
Cardiorespiratory fitness
Accelerometers, gyroscopes, and photoplethysmography quantify activity and chronotropic response. Lower free-living step counts are associated with worse functional status in heart failure and can degrade ahead of clinical decompensation in observational cohorts8. In a prospective heart failure trial (LacS-001), a sweat-lactate patch identified a lactate threshold that correlated with a ventilatory threshold (r = 0.651), with a mean bias of –4.9±15.0 W, supporting feasibility for exercise prescription9.
Composite physiologic panels
Contemporary programmes aggregate heart rate variability, cuffless blood pressure surrogates, respiratory rate, and symptom inputs into multimodal dashboards, enabling near-continuous ‘physiologic storyboards’ that support protocolised, early interventions (see Table for device diversity and study designs).
| Device/platform | Signals (modalities) | Key evidence (design; n; headline metrics) | Regulatory status | Notes/limitations |
|---|---|---|---|---|
| HeartLogic™ (Boston Scientific; ICD/CRT-D algorithm) | Heart sounds, thoracic impedance, respiration, night HR, activity | Validation (MultiSENSE): sens ⁓70%; median lead-time ⁓34 days 14; unexplained alerts ⁓1.46/pt-yr2; real-world: ⁓0.4–1.0 alerts/pt-yr2 9 | Within FDA-approved ICD/CRT-D systems | Performance ≠ outcomes; outcomes trial (MANAGE-HF) ongoing 15 |
| ReDS™ dielectric vest (Sensible medical) | RF dielectric sensing of lung fluid | Lower odds of HF readmission ≤3 months vs. no-ReDS (meta-analysis) 16 | FDA 510(k) for non-invasive assessment of lung-fluid content (e.g., K150095) 17 | Observational designs; no large RCT showing outcomes; body habitus/posture can affect readings |
| μCor™ HFMS (ZOLL) | RF thoracic fluid (patch/belt) | Concurrent-control programme (BMAD-HF): 90-day HF readmission ↓ 38% (HR 0.62) 18 | US-marketed Class II; verify K-number in final refs | Vendor-supported workflow; non-randomised; causal inference limited |
| VitalPatch® + LINK-HF analytics | Single-lead ECG, impedance/respiration, accelerometer, temperature | Prospective cohort: predict HF hospitalisation with lead time of ∼6.5–8.5 days; sens 76–88%, spec ∼85% 19 | Investigational for HF prediction | No RCT outcomes |
| CordioHearO (voice analytics app) | Smartphone speech features (congestion proxy) | Hospitalised HF cohorts: voice-derived index tracks wet→dry state; studies suggest pre-event warning window (wk) 20 | Investigational | Language/accent & ambient noise sensitivity |
| Acorai Heart Monitor (handheld multisensor) | Seismo-/phono-cardio + tonometry fusion | Feasibility: r ≈ 0.75 vs. invasive pressures in multi-centre cohorts 21 | Investigational | Prospective validation/outcomes pending |
| Cardiosense CardioTag (chest wearable) | Seismocardiography + accelerometry | Early feasibility: ML estimates PCWP; pilot/observational data 22 | Investigational | Not yet peer-reviewed outcomes |
| Sweat-lactate patch (LacS-001) | Sweat lactate (exercise threshold) | Prospective HF trial: sLT correlates with ventilatory threshold (r ≈ 0.65); bias −4.9±15.0 W13 | Investigational | Exercise/rehab adjunct; not diagnostic |
| Consumer smart-watches (e.g., Apple Watch) | PPG + single-lead ECG | AF detection accuracy: pooled sens 94.8%, spec 95% in general adults 7 | FDA-cleared for AF detection; not for HF decompensation | No evidence for HF decompensation detection or HF outcomes |
AF, atrial fibrillation; HR, heart rate; HF, heart failure; PCWP, pulmonary capillary wedge pressure; pt-yr, patient-year; RF, radiofrequency; RCT, randomised controlled trial; sLT, sweat-lactate threshold; ECG, electrocardiography; FDA, Food and Drug Administration
Technological landscape
Devices now entering routine practice are catagorised in fall into three broad families with a common goal: capture multiple heart failure-relevant signals and deliver them to clinicians in near real time.
Chest-worn adhesive patches
Patches such as VitalPatch sit over the sternum and record single-lead ECG, respiratory surrogates/impedance, triaxial accelerometry, and skin temperature for extended wear. In the multicentre LINK-HF observational study, a cloud-based machine-learning backend using multisensor streams predicted imminent heart failure hospitalisation with sensitivity 76–88 per cent, specificity ⁓85 per cent, and lead time ⁓6.5–8.5 days10.
Smart textiles
Conductive yarns and stretch sensors woven into vests or T-shirts can provide multi-lead ECG and high-fidelity respiratory signals, sometimes paired with dielectric measurements of lung water. For dielectric vests specifically, evidence for reduced early readmissions comes from observational and pre–post cohorts and a meta-analysis showing lower odds of 30–90-day heart failure readmission (OR ⁓0.40); a large positive RCT has not been established11. Consumer smartwatches achieve high detection accuracy of atrial fibrillation in general adult population; no evidence demonstrates heart failure decompensation detection or improvement in heart failure outcomes. In heart failure, detection of atrial fibrillation may support anticoagulation or rate-control adjustment.
Consumer wearables
Wrist-worn devices use photoplethysmography to log heart-rate trends, rhythm strips via single-lead ECG, oxygen saturation, and activity. These tools show high accuracy for detection of atrial fibrillation in general adult population (pooled sensitivity ⁓94.8%, specificity ⁓95%), but heart failure decompensation detection and heart failure outcome benefit are not established3. Hybrid programmes often combine a clinical-grade chest patch for physiology with a smartwatch for prompts and engagement.
Across form factors, five sensor modalities dominate: continuous ECG (rhythm), photoplethysmography (rate/SpO₂), thoracic bioimpedance (fluid), accelerometry (activity/posture), and dielectric sensing (absolute or relative lung-fluid estimation). Most platforms now fuse signals at the edge or in the cloud to mitigate artifacts (e.g., photoplethysmography motion, impedance drift); LINK-HF blended 26 variables to raise early warnings roughly a week before admission with ⁓85 per cent specificity10.
Interoperability and data pathways are crucial. Raw data typically travels via bluetooth low energy to a phone or hub, then through Transport Layer Security-protected Application Programming Interfaces to cloud dashboards for clinical review. Integration with the EHR (Electronic Health Record) increasingly relies on HL7 FHIR (Fast Healthcare Interoperability Resources) with SMART-on-Fast Healthcare Interoperability Resources OAuth2 authorisation for secure app access, while cross-vendor harmonisation of telemonitoring data remains inconsistent, according to a recent scoping review12,13. The complete remote-care loop, integrating multi-sensor data capture, cloud-based analytics, clinical triage, and protocolised intervention, is illustrated in the figure.
- Remote-care loop for multi-sensor heart-failure wearables. HF, heart failure; PPG, photoplethysmography; GDMT, guideline-directed medical therapy; EHR, electronic health record; FHIR, fast healthcare interoperability resources; TLS, transport layer security.
Composite indices and alerting
To reduce alert fatigue, vendors combine metrics into composite scores. The HeartLogic™ implantable algorithm (heart sounds, thoracic impedance, respiration, night heart rate, activity) provides a median ⁓34-day advance notice of worsening heart failure5, and in prospective practice evaluations, non-clinically-meaningful alerts occur @ ⁓0.37 per patient-year14. Clinical teams typically embed thresholds and call-to-action workflows directly in dashboards to streamline tele-visits.
Clinical outcomes—What the trials and registries show
Early proof-of-concept is based on LINK-HF study, where a disposable multisensor chest patch plus machine learning correctly anticipated many rehospitalizations ⁓6.5–8.5 days before presentation with specificity ⁓85 per cent (multicenter observational study)10. These signals demonstrated feasibility and tolerability for continuous, multi-stream physiology outside the hospital10.
Moving from prediction to outcomes, programmes that use multisensor data to guide therapy have reported harder endpoints. In the multicenter BMAD-HF/Tx pathway, ambulatory thoracic impedance, ECG, and activity-guided post-discharge care were associated with a 38 per cent relative reduction in 90-day heart failure readmissions (HR 0.62) versus usual care, alongside more timely diuretic adjustments (concurrent-control, non-randomised)15. Because treatment allocation was not randomised, the observed benefit should be interpreted as a programme-level association rather than device causality15.
A stronger causal signal comes from a vulnerable-phase RCT. In HERMeS-HF study (n=506), adding a bundled mHealth package, remote telemonitoring plus structured tele-intervention, to guideline-directed care reduced the composite of cardiovascular death or worsening heart failure at 24 wk (17% vs. 41%; HR 0.35, 95% CI 0.24–0.50)2. This effect size contrasts with earlier weight/symptom-only telemonitoring, aligning with the hypothesis that streamed objective physiology and responsive workflows are the key active components (Yun et al2, 2025).
Real-world programmes echo these patterns yet remain susceptible to selection and coding bias. A U.S. payer-sponsored initiative reported ⁓52 per cent lower monthly total cost of care and fewer heart failure admissions over six months with comprehensive remote patient monitoring and centralised titration (observational analysis)16,17. Similarly, a Finnish pre-post evaluation found that the hospitalisation-related costs decreased by ⁓49 per cent and the proportion experiencing ≥1 heart failure HF admission decreased by ⁓70 per cent (P=0.002) after app-plus-device remote patient monitoring was added to usual care18. These implementation signals are encouraging but non-causal.
Patient-centred outcomes trend positive when objective physiology drives the care pathways. Several modern telemonitoring trials reported clinically meaningful Kansas City Cardiomyopathy Questionnaire (KCCQ) gains (⁓5–8 points), whereas controls were neutral or smaller; device-related adverse events were uncommon, with mild adhesive dermatitis the most frequent complaint19. Wear-time feasibility in LINK-HF trial (median ⁓23 h/day) supports usability for continuous monitoring10.
A 2025 comprehensive meta-analysis of non-invasive remote patient monitoring concluded that programmes lower first heart failure hospitalisation and all-cause mortality overall, with larger effects when data streams include objective physiology and protocolised responses (heterogeneity modest)20. A complementary 2024 analysis estimated heart failure-hospitalization RR ⁓0.78 and all-cause mortality RR ∼0.84, again favouring physiologic remote patient monitoring over weight/symptom-only designs21. While small-study asymmetry appears in some strata, pooled hospitalisation benefits persist after adjustment, suggesting publication bias does not fully explain the effect20.
When multi-sensor wearables are embedded in responsive clinical workflows, evidence has progressed from prediction to reductions in heart failure events, with supportive (if imperfect) real-world cost data. The most robust causal signal to date comes from post-discharge RCTs coupling remote data with structured tele-intervention2,20.
Emerging sensors and analytical horizons
Chemical sensing is starting to join physiologic monitoring at the wrist and chest. Flexible microfluidic electrolyte patches have matured to the point that benchtop validation now shows excellent agreement with laboratory instruments (intraclass correlation 0.998) and low signal variability (coefficient of variation <4%), while maintaining continuous streaming during exercise22. Although these devices were not built specifically for heart failure, sodium trends can complement congestion and diuretic titration signals, especially when interpreted alongside weight, impedance, and symptoms.
From ions to metabolites, sweat-lactate sensing has crossed into heart failure clinics. In the prospective LacS-001 trial of ambulatory heart failure patients, a wearable lactate patch identified the anaerobic threshold within a mean –4.9 ± 15.0 W of gas-exchange testing and reported no device-related adverse events, supporting its use to individualize exercise prescriptions in cardiac rehabilitation9. This kind of metabolic window sits naturally beside activity and heart-rate recovery, giving clinicians a richer picture than steps alone.
A step further is continuous molecular monitoring via microneedle electrochemical aptamer sensors that sample interstitial fluid. Human-feasibility work has already demonstrated real-time, on-body analyte tracking with microneedle-enabled aptamer sensors23, and materials advances have shown week-long operational stability of aptamer interfaces under repeated scanning, an essential prerequisite for chronic disease use24. While NT-proBNP-targeting aptamer platforms have achieved femtomolar-level detection in whole blood on benchtop chips, translation to an on-body, continuous microneedle format in heart-failure populations has not yet been reported; this field remains a near-term developmental frontier rather than a clinical reality25.
Analytically, the field is also moving beyond single thresholds toward mechanistic ‘digital twins’. In a 343-patient heart failure study, adding 29 digital-twin–derived physiologic features to a random-survival-forest model improved internal prognostic performance (out-of-bag C-index 0.724 vs. 0.707 using clinical variables alone) and showed transferable discrimination in an external cohort (C-index 0.671 for the combined-feature model), while preserving interpretability of dominant hemodynamic drivers for each patient26. For remote-monitoring programmes, the study suggests a path from more data to explainable, patient-specific risk signals that clinicians can act on.
Finally, innovation is widening access, not just capability. USSD-based self-care platforms adapted for basic mobile phones (e.g., Medly Uganda) have shown acceptability and feasibility among patients and clinicians, with a pilot clinical trial completed and community adaptations documented; these models illustrate how remote monitoring can be context-fit even where smartphones and broadband are scarce27. As these programmes layer simple symptom triage with physiologic signals from low-cost peripherals, they offer a realistic route to equitable deployment.
Together, these strands point to a near future in which multi-modal wearables pair physiology (rhythm, activity, and impedance) with chemistry (electrolytes, metabolites, and eventually proteins), and analytics elevate the stream into individualised, interpretable guidance, shrinking the gap between early deterioration and timely action.
Discussion
Remote multi-sensor wearables are shifting heart failure care from episodic checks to continuous surveillance that can surface hemodynamic drift days before admission. Early prediction with multisensor patches have shown feasible lead times and specificity for impending decompensation in, indicating signal-level readiness for triage rather than outcome claims on its own10. By contrast, simple telephone-based weight/symptom telemonitoring failed to improve death or rehospitalisation in a large randomised trial, underscoring that data without structured response does not change outcomes1. Causal benefit emerges when streamed physiology is paired with protocolised tele-intervention in the vulnerable post-discharge window: the HERMeS randomised trial reported a marked reduction in cardiovascular death or worsening heart failure over 24 wk (17% vs. 41%; HR 0.35, 95% CI 0.24–0.50)2. Contemporary meta-analyses of non-invasive remote patient monitoring corroborate lower heart failure hospitalisations and all-cause mortality, with larger effects in programmes that combine objective physiology with predefined actions20.
Translating signals into care depends on service design. Daily centralised review by nurses, explicit thresholds, and pre-authorised medication adjustments keep alert-to-action latency short while containing workload. Multiparameter indices can help limit non-actionable alerts to a rate compatible with routine dashboards; real-world HeartLogic™ reports illustrate how combining streams can preserve a useful early-warning window with a manageable alert burden, though calibration and clinical oversight remain mandatory5. Embedding remote data within the electronic health record via HL7 FHIR/SMART-on-FHIR reduces login friction and improves order adherence by keeping triage and order sets in the same context28. For safety and trust, programmes should document transport security, app authorisation, audit trails, and rapid patch cycles rather than rely on generic assurances.
Equity and measurement bias require direct mitigation. Pulse-oximetry error varies by skin pigmentation and can misclassify hypoxemia in darker-skinned patients; single-stream decisions tied to SpO₂ should be avoided when signals conflict29. Optical heart-rate sensing via photoplethysmography also shows device- and context-dependent accuracy with skin-tone considerations, reinforcing the value of multi-wavelength hardware and transparent validation by Fitzpatrick type30. Access barriers can be lowered with device-agnostic kits (e.g., cellular hubs) and low-bandwidth pathways; the Medly-Uganda program demonstrates that USSD-based self-care plus clinician triage is acceptable and feasible where smartphones and broadband are limited27.
A credible path forward focuses on high-risk periods and measurable actions. The largest absolute yield concentrates in the first 30–90 days after discharge, where risk and actionability peak; stable, low-risk outpatients likely realize smaller gains. As a construct for causality, invasive hemodynamic monitoring shows that admissions can be reduced by titrating therapy to pulmonary-artery pressures; external wearables should emulate the discipline, algorithmic suggestions, explicit clinician confirmation, and auditable changes while evidence matures31,32. Priorities include pragmatic trials and service evaluations that report patient-important outcomes and costs, longitudinal audits of algorithm performance across populations, and interoperable data flows. Biochemical sensors and advanced analytics are promising but remain investigational for decompensation management. Selective deployment during the post-discharge period, paired with protocolised response, Electronic Health Records (EHR) integration, and equity-focused design, offers the greatest clinical return today. Further gains depend on pragmatic trials with patient-important outcomes, external algorithm audits, and interoperable, secure data flows.
Financial support & sponsorship
None.
Conflicts of Interest
None.
Use of Artificial Intelligence (AI)-Assisted Technology for manuscript preparation
The authors confirm that there was no use of AI-assisted technology for assisting in the writing of the manuscript and no images were manipulated using AI.
References
- Telemonitoring in patients with heart failure. N Engl J Med. 2010;363:2301-9.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Evaluation of mobile health technology combining telemonitoring and teleintervention versus usual care in vulnerable-phase heart failure management (HERMeS): a multicentre, randomised controlled trial. Lancet Digit Health. 2025;7:100866.
- [CrossRef] [PubMed] [Google Scholar]
- Diagnostic accuracy of apple watch electrocardiogram for atrial fibrillation: a systematic review and meta-analysis. JACC Adv. 2025;4:101538.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Evaluation of remote dielectric sensing (ReDS) technology-guided therapy for decreasing heart failure re-hospitalizations. Int J Cardiol. 2017;240:279-84.
- [CrossRef] [PubMed] [Google Scholar]
- HeartLogicTM: Real-world data—efficiency, resource consumption, and workflow optimization. Eur Heart J Suppl. 2023;25:C331-6.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Cryptogenic stroke and underlying atrial fibrillation. N Engl J Med. 2014;370:2478-86.
- [CrossRef] [PubMed] [Google Scholar]
- Frequent premature atrial contractions as a signalling marker of atrial cardiomyopathy, incident atrial fibrillation, and stroke. Cardiovasc Res. 2023;119:429-3.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Use of free-living step count monitoring for heart failure functional classification: Validation study. JMIR Cardio. 2019;3:e12122.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Sweat lactate sensor for detecting anaerobic threshold in heart failure: a prospective clinical trial (LacS-001) Sci Rep. 2024;14:18985.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Continuous wearable monitoring analytics predict heart failure hospitalization. Circ Heart Failure. 2020;13:1323310.
- [CrossRef] [Google Scholar]
- Efficacy of remote dielectric sensing (ReDS) in the prevention of heart failure rehospitalizations: a meta-analysis. J Community Hosp Intern Med Perspect. 2021;11:646-52.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- HL7 International. SMART App Launch: Launch and Authorization - SMART App Launch v2.2.0. Available from: https://build.fhir.org/ig/HL7/smart-app-launch/app-launch.html, accessed August 17, 2025.
- Interoperability of telemonitoring data in digital health solutions: a scoping review. Front Digit Health. 2025;7:1502260.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Prospective evaluation of the multisensory HeartLogic algorithm for heart failure monitoring. Clin Cardiol. 2020;43:691-7.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Impact of a novel wearable sensor on heart failure rehospitalization. JACC Heart Fail. 2024;12:2011-22.
- [CrossRef] [PubMed] [Google Scholar]
- Fierce Healthcare. Cadence finds reduced cost and better outcomes for heart failure patients through remote monitoring program. Available from: https://www.fiercehealthcare.com/digital-health/exclusive-cadence-finds-reduced-cost-and-better-outcomes-heart-failure-patients, accessed August 17, 2025.
- Leveraging remote patient monitoring to effectively put the heart failure guidelines to practice. J Card Fail. 2024;30:1166-9.
- [CrossRef] [PubMed] [Google Scholar]
- Effectiveness of telemonitoring in reducing hospitalization and associated costs for patients with heart failure in Finland: Nonrandomized pre-post telemonitoring study. JMIR Mhealth Uhealth. 2024;12:e51841.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Wearable devices in cardiovascular medicine. Circ Res. 2023;132:652-70.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Remote patient monitoring in heart failure: a comprehensive meta-analysis of effective programme components for hospitalization and mortality reduction. Eur J Heart Fail. 2025;27:1670-85.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Comparative effectiveness of complex telemedicine support in prevention of hospitalizations and mortality in patients with heart failure: a systematic review and meta-analysis. Life (Basel). 2024;14:507.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Wearable device for continuous and real-time monitoring of human sweat sodium. Sensors. 2025;25:3467.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Continuous molecular monitoring of human dermal interstitial fluid with microneedle-enabled electrochemical aptamer sensors. Lab Chip. 2023;23:3289-9.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Week-long operation of electrochemical aptamer sensors: New insights into self-assembled monolayer degradation mechanisms and solutions for stability in serum at body temperature. ACS Sens. 2023;8:1119-31.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Monitoringcardiac biomarkers with aptamer-based molecular pendulum sensors. Angew ChemInt Ed Engl. 2023;62:e202213567.
- [Google Scholar]
- Identification of digital twins to guide interpretable AI for diagnosis and prognosis in heart failure. NPJ Digit Med. 2025;8:110.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- A digital self-care intervention for Ugandan patients with heart failure and their clinicians: User-centred design and usability study. Digit Health. 2022;8:20552076221129064.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- SMART on FHIR: a standards-based, interoperable apps platform for electronic health records. J Am Med Inform Assoc. 2016;23:899-908.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Racial bias in pulse oximetry measurement. N Engl J Med. 2020;383:2477-8.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Investigating sources of inaccuracy in wearable optical heart rate sensors. NPJ Digit Med. 2020;3:18.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Wireless pulmonary artery haemodynamic monitoring in chronic heart failure: a randomised controlled trial. Lancet. 2011;377:658-66.
- [CrossRef] [PubMed] [Google Scholar]
- Sustained efficacy of pulmonary artery pressure to guide adjustment of chronic heart failure therapy: Complete follow-up results from the CHAMPION randomised trial. Lancet. 2016;387:453-61.
- [CrossRef] [PubMed] [Google Scholar]