Heart Failure Monitoring Wearables: What’s Actually Being Tested in Clinical Trials Right Now

⚡ Why Heart Failure Monitoring Is a Different, Harder Problem

Heart failure isn’t a single event like a heart attack — it’s a chronic condition characterized by cycles of relative stability and periods of worsening, called exacerbations or decompensation, that often lead to hospitalization. A critical, well-established clinical fact drives most of the technology in this space: pulmonary artery pressure — a measure of how hard the heart is working to move blood through the lungs — tends to rise days to weeks before a patient notices symptoms or a clinician would catch a problem at a routine visit. That lag is exactly the window this entire category of technology is trying to catch. The challenge is that reliably detecting that early warning sign, without a lot of false alarms, has turned out to be much harder outside a hospital than measuring a heart rhythm irregularity.

🏆 The Real Established Evidence — and It’s Not a Wearable

Before looking at wearables specifically, it’s worth understanding what “good evidence” actually looks like in heart failure monitoring, because one device has already cleared that bar — just not as a wearable.

CardioMEMS, made by Abbott, is a permanently implanted wireless sensor placed in the pulmonary artery via a minimally invasive procedure, which transmits daily pressure readings to a patient’s care team. It received FDA approval in 2014, based on the randomized CHAMPION trial, which enrolled 550 patients and found a 28% reduction in heart failure hospitalizations at 6 months and a 37% reduction at 15 months in patients managed using CardioMEMS data compared to standard care — a genuinely large, randomized, controlled result. Abbott received FDA approval for an updated reader device, CardioMEMS HERO, in February 2026, continuing to build out the ecosystem around the sensor.

The reason this matters for the rest of this article: CardioMEMS works because it directly measures the pressure signal that matters, via an implanted sensor — not because it’s inferring that signal indirectly from something worn on the wrist. That’s a meaningfully easier engineering problem than what non-invasive wearables are attempting, and it explains why the evidence bar has been cleared here first.

⌚ The Strongest Wearable-Specific Evidence So Far

This is where the picture gets more genuinely promising for consumer-accessible technology. In March 2026, Nature Medicine published results from the TRUE-HF study (Ted Rogers Understanding Exacerbations of Heart Failure), led by researchers including Yuan Gao, examining whether an Apple Watch could estimate peak oxygen uptake (pVO2) — a clinically meaningful measure of cardiopulmonary fitness that normally requires an in-clinic cardiopulmonary exercise test (CPET) to obtain. The study followed an observational cohort of free-living heart failure patients over a median of about 94.5 days. A deep learning model was trained on data from 154 patients and validated on a separate held-out group of 63 patients, and the wearable-derived daily pVO2 estimates correlated strongly with actual in-clinic CPET measurements — a Pearson’s correlation of 0.85, which is a genuinely strong result for this kind of indirect, passive estimation.

🧪 Early-Stage and Conference-Presented Research

At the Technology and Heart Failure Therapeutics (THT) 2026 meeting in Boston, Afnan Tariq, an interventional cardiologist at UC Irvine, presented first-in-man results from a passive, device-agnostic AI platform designed to turn data from ordinary consumer wearables — not a specialized medical device — into clinically actionable insights for heart failure patients. The presented results reportedly showed a reduction in hospitalizations over time, attributed to earlier interventions prompted by the platform’s alerts.

This is worth taking seriously as an early signal, and also worth treating with appropriate caution: conference-presented “first-in-man” results are, by definition, an early, small-scale finding that hasn’t yet gone through full peer review and publication, the same evidentiary bar that TRUE-HF and the CHAMPION trial both cleared. It’s a promising direction, not yet a validated result.

🔬 Trials Currently Recruiting or Underway

A genuine strength of this category right now is that several concrete, registered clinical trials are actively working through exactly these open questions, rather than the field relying purely on marketing claims about future potential:

SMART-CARE, based at the University of Salerno in Italy, is studying an AI-based monitoring approach for heart failure patients that combines wearable devices with mini-invasive monitoring technology — directly investigating the middle ground between fully non-invasive wearables and fully implanted sensors like CardioMEMS.

REMOTI-HF (NCT06422832), based in Portugal, is a randomized clinical trial specifically designed to test whether remote monitoring can identify worsening heart failure earlier. As of its last recorded status update, the trial’s status was listed as not yet recruiting, with an estimated completion date in 2026 — a reminder that “a trial exists” and “a trial has produced results” are different stages worth distinguishing.

Nov-RPM-HF, a single-center prospective trial based at Washington University School of Medicine, is evaluating a clinician-co-designed remote monitoring platform built around a ballistocardiograph — a device that monitors nocturnal biometrics passively, without requiring a wearable to be worn on the body at all. The published protocol describes a 100-patient, 6-month study measuring patient adherence and clinician/patient perceptions, alongside clinical data — a useful example of monitoring technology that isn’t a wrist-worn wearable but is part of the same broader push toward non-invasive, at-home heart failure surveillance.

HERMeS (NCT03663907), completed in Spain, tested a telemedicine-based remote monitoring program involving daily self-reported symptoms, weight, blood pressure, and heart rate, plus weekly videoconference follow-up, compared with usual care — illustrating that not every “remote monitoring” study in this space is about passive wearable sensors; some rely on structured self-reporting instead, which is a genuinely different (and generally lower-tech) approach worth distinguishing from AI-driven wearable analysis.

📖 What the Broader Research Says, Honestly

Two recent, independent reviews of this field are worth summarizing directly, because they give a more complete picture than any single study.

A review published in Current Heart Failure Reports (Springer Nature, late 2025) concluded that remote monitoring — combining wearable and non-wearable devices — does reduce hospitalizations and mortality in heart failure patients, with AI increasingly integrated to enhance predictive accuracy and personalize care, and that despite high upfront costs, the approach appears cost-effective over the longer term due to reduced hospitalizations. But the same review was direct about a key limitation: whether AI-driven approaches specifically can reduce hospitalizations by enabling timely clinical interventions “remains to be established through prospective, randomized trials” — meaning the observational and pilot data so far are encouraging, but the harder, gold-standard proof is still pending.

A separate narrative review of AI-integrated wearables in cardiovascular medicine broadly reached a consistent conclusion for heart failure specifically: promising outcomes exist, particularly with FDA- or CE-approved devices, but most available studies were small, short-term, and conducted in controlled settings, which limits how confidently those findings generalize to everyday use. The review also flagged real, unresolved gaps: underrepresentation of diverse patient populations in the research, limited interoperability with electronic health records, limited explainability of the AI models involved, and incomplete cost-effectiveness evaluation.

Put together, both reviews land in a similar place: real, encouraging signal, especially for regulator-approved devices, but a field that still needs the kind of large-scale, randomized, hard-outcome trial that CardioMEMS already has behind it — for wearables specifically, that trial largely hasn’t happened yet.

⚖️ Why This Is Harder Than AFib Detection

It’s worth explicitly connecting this back to the cardiac wearable evidence we’ve covered elsewhere on this site, because the contrast is genuinely instructive. AFib detection succeeded early on consumer wearables because the underlying signal — an irregular electrical rhythm — is relatively binary and directly measurable through a single-lead ECG sensor built into a watch. A rhythm is either irregular in the specific pattern associated with AFib, or it isn’t, and a device can be validated against that comparatively clean signal in a large study, which is exactly what the Apple Heart Study did with over 400,000 participants.

Heart failure decompensation doesn’t offer that same clean signal. It’s a gradual physiological drift — fluid retention, reduced exercise capacity, subtle changes in heart rate variability and breathing patterns — that unfolds over days, driven by fundamentally different underlying mechanisms in different patients, and it doesn’t have one single, unambiguous sensor reading that confirms it the way an ECG waveform confirms AFib. That’s precisely why the most direct, established approach (CardioMEMS) works by measuring pressure directly via an implanted sensor rather than trying to infer it indirectly from wrist-worn optical or motion sensors, and it’s why wearable-based approaches have to rely on more complex, multi-signal AI models — like the deep learning approach in TRUE-HF — trained to infer a proxy measure (fitness capacity) rather than directly sensing the underlying problem. More moving parts, a noisier signal, and a harder inference problem all combine to explain why this category is several years behind AFib detection in its evidence maturity, not because the research effort has been any less serious.

🎯 What “Success” Would Actually Look Like

It’s worth being concrete about what the field still needs to demonstrate before wearable-based heart failure monitoring reaches CardioMEMS-level confidence. Based on the gaps identified in the reviews above, a genuinely validated wearable HF monitoring system would need: a large, randomized (not just observational) trial showing that acting on the wearable’s alerts actually reduces hospitalizations or improves survival, not just that the wearable’s readings correlate with a clinical measurement; validation across diverse patient populations, including different races, ages, and heart failure subtypes, rather than relatively homogeneous study cohorts; demonstrated integration with clinicians’ actual workflows and electronic health records, so alerts translate into real, timely action rather than data nobody acts on; and a clear cost-effectiveness case, given how expensive continuous AI-driven monitoring programs can be to run at scale. None of the studies covered in this article have cleared all of these bars yet — which is exactly why this remains an area to watch rather than a finished technology to rely on.

📊 Quick Comparison

📋 Questions to Ask About Any Heart Failure Monitoring Claim

• Is this an implanted sensor, a wearable, or a passive home device (like a ballistocardiograph)? Each category has a very different current evidence base, and conflating them overstates what’s actually proven for any one of them.
• Does the study show the device can accurately measure something, or that acting on the measurement improves outcomes? As TRUE-HF illustrates, these are two different, separately provable claims.
• Is this a peer-reviewed, published result, or a conference presentation of preliminary data? Both are worth knowing about, but they carry different levels of confidence.
• Was this tested specifically in heart failure patients, or in a general or healthy population? A device validated in one group doesn’t automatically transfer its accuracy to another.
• Is a specific clinical trial registration (ClinicalTrials.gov) or published study cited, or just a company description of “AI-powered monitoring”?

👤 Who This Is Actually Relevant To

If you or a family member has diagnosed heart failure, the technology with the strongest, most complete evidence behind it today remains CardioMEMS — an implanted device requiring a cardiologist’s involvement and a minimally invasive procedure, not a consumer purchase decision. That’s worth discussing directly with a heart failure specialist if hospitalization risk is a real concern.

If you’re interested in wearable-based monitoring specifically, the honest picture is that this is a genuinely active, promising area of research — not a finished, validated clinical tool yet. The TRUE-HF study is a real, meaningful step, but it measures fitness capacity, not a direct decompensation alert, and its predictive value for actually preventing a hospitalization hasn’t yet been established the way CardioMEMS’s has. If your cardiologist has recommended a specific wearable-based remote monitoring program as part of your care, that’s a different situation — clinician-guided monitoring programs, even ones using less-than-fully-proven technology, are implemented with appropriate oversight, unlike using a consumer wearable to self-monitor for heart failure decompensation without medical guidance.

❓ Frequently Asked Questions

Can a smartwatch detect heart failure getting worse?

Not reliably yet, based on current published evidence. The strongest wearable-specific study (TRUE-HF, Nature Medicine, 2026) showed an Apple Watch could accurately estimate a fitness marker (pVO2) linked to heart failure status, but this hasn’t yet been proven to predict or prevent an actual hospitalization the way the implanted CardioMEMS sensor has in randomized trials.

What is CardioMEMS, and is it a wearable?

No — CardioMEMS is a permanently implanted wireless sensor placed in the pulmonary artery through a minimally invasive procedure. It’s the most well-evidenced heart failure monitoring technology available, with FDA approval since 2014 based on a randomized trial showing reduced hospitalizations, but it is not a consumer wearable device.

Are there clinical trials currently testing wearables for heart failure?

Yes. Several trials are actively underway or recently completed, including SMART-CARE (Italy), REMOTI-HF (Portugal), and Nov-RPM-HF (a ballistocardiograph-based study in the US), each testing different approaches to non-invasive heart failure monitoring.

Should I rely on a wearable instead of my cardiologist’s monitoring plan?

No. Even the most promising wearable research published so far demonstrates measurement accuracy, not proven prevention of hospitalization or improved outcomes. If you have heart failure, follow your cardiologist’s specific monitoring recommendations rather than substituting a consumer wearable.

📚 Sources & References

• Gao, Y., et al., “Remote monitoring of heart failure exacerbations using a smartwatch,” Nature Medicine, 2026, 32:924–933
• Clinical Trials Arena, “Smartwatches could aid in monitoring heart failure exacerbations”
• Cardiovascular Business, “AI turns wearable device data into actionable insights for heart failure patients” (THT 2026 conference coverage)
• Abraham, W.T., et al., “Wireless pulmonary artery haemodynamic monitoring in chronic heart failure: A randomised controlled trial” (CHAMPION Trial), The Lancet, 2011;377(9766):658-666
• tctmd.com, “CardioMEMS Receives FDA Approval for Heart Failure Treatment”
• Abbott, “Abbott’s next-generation CardioMEMS remote heart failure monitoring reader receives FDA approval”, February 2026
• Anthem/Blue Cross Blue Shield, “MED.00115 Outpatient Cardiac Hemodynamic Monitoring Using a Wireless Sensor for Heart Failure Management” (independent clinical evidence review)
• Bhatia, A., Ewald, G., Maddox, T., “The Novel Data Collection and Analytics Tools for Remote Patient Monitoring in Heart Failure (Nov-RPM-HF) Trial,” JMIR Research Protocols, 2022
• ClinicalTrials.gov, NCT06422832 — REMOTI-HF; NCT03663907 — HERMeS
• “Non-Invasive Remote Monitoring in Heart Failure: Towards Wearable Devices and Artificial Intelligence Solutions,” Current Heart Failure Reports, Springer Nature, 2025
• “Wearable Devices and AI-Driven Remote Monitoring in Cardiovascular Medicine: A Narrative Review,” PMC12439441

Verification notes: The TRUE-HF study was independently verified via its direct Nature Medicine publication listing, including sample sizes and the reported correlation coefficient. CardioMEMS’s CHAMPION trial data (550 patients, 28%/37% hospitalization reduction) is drawn from multiple independent, corroborating sources, and the article explicitly includes the more cautious, mixed-evidence framing from an independent insurance clinical-policy review rather than presenting only the most favorable framing. The THT 2026 conference presentation is explicitly labeled as early, first-in-man, conference-presented data rather than a peer-reviewed publication. Trial statuses for REMOTI-HF and other registered trials reflect the last publicly recorded status at the time of writing and may have changed since; readers should check ClinicalTrials.gov directly for current status.

💡 The Bottom Line

Heart failure monitoring is a genuinely active, fast-moving research area with real momentum behind it — but it’s earlier in its evidence journey for wearables specifically than the AFib-detection space, and meaningfully earlier than CardioMEMS’s implanted-sensor approach, which already has randomized, hospitalization-reduction evidence behind it. The most encouraging wearable-specific result to date, TRUE-HF’s Apple Watch pVO2 study, shows real measurement accuracy, not yet proven impact on outcomes. Several concrete trials — SMART-CARE, REMOTI-HF, Nov-RPM-HF — are actively working to close that gap right now, which is exactly the kind of evidence-building process worth watching rather than a technology to rely on today.

Every diagram on this page was built specifically for this article and is not sourced from any manufacturer.