Remote Cardiac Monitoring

What “remote cardiac monitoring” covers

Remote cardiac monitoring is the collection of ambulatory electrocardiographic or photoplethysmographic rhythm data outside a hospital, followed by transmission to a reviewer or algorithm. The commercial field now spans at least five distinct device classes:

1. Single-lead adhesive ECG patches, typically worn 1–14 days. Example: iRhythm Technologies’ Zio XT/Zio AT.

Source: iRhythm provider page, accessed 2026-04-26, irhythmtech.com

1. Consumer smartwatch/ring photoplethysmography + single-lead ECG tools. Examples: Apple Watch irregular pulse notification and ECG app, Fitbit Sense/Charge ECG, Samsung Galaxy Watch ECG, Garmin ECG App, AliveCor KardiaMobile/KardiaMobile 6L.

Source for Apple ECG clearance: FDA De Novo DEN180044, accessed 2026-04-26, accessdata.fda.gov

Source for AliveCor KardiaMobile 6L product page: alivecor.com , accessed 2026-04-26

1. Insertable cardiac monitors / implantable loop recorders, typically implanted subcutaneously for multi-year monitoring. Examples: Medtronic Reveal LINQ II, Abbott Confirm Rx, Boston Scientific LUX-Dx.

Source: Medtronic Reveal LINQ product page, accessed 2026-04-26, medtronic.com

1. Mobile cardiac telemetry, often a patch plus real-time transmission plus staffed monitoring. Historically BioTelemetry/Philips and Preventice/Boston Scientific.

Source for Philips/BioTelemetry SEC filings page: philips.com , accessed 2026-04-26

1. Algorithmic AF detection from consumer heart-rate sensors, which may be paired with an ECG app but relies on an irregular-rhythm notification rather than a diagnostic ECG.

Source: Perez et al., N Engl J Med 2019;381:1909–1917, DOI 10.1056/NEJMoa1901183, published 2019-11-13, accessed 2026-04-26, doi.org

Within those classes, the strongest independent clinical evidence is not uniformly distributed. The deepest peer-reviewed evidence exists for two populations: population-scale smartwatch screening for atrial fibrillation and implantable loop recorders in cryptogenic stroke. The patch category has broad commercial adoption but fewer large randomized outcome trials.

Source inventory

2.1 Peer-reviewed clinical trials

Perez MV, et al. “Large-Scale Assessment of a Smartwatch to Identify Atrial Fibrillation.” New England Journal of Medicine 2019;381:1909–1917. DOI: 10.1056/NEJMoa1901183. Published 2019-11-13. Accessed 2026-04-26. URL: doi.org

Key findings from the Apple Heart Study:

• 419,297 participants enrolled.
• 2,161 participants, or 0.52%, received an irregular pulse notification.
• Among 450 participants who returned an ECG patch, atrial fibrillation was present in 153, or 34.0%.
• The positive predictive value of the irregular pulse notification for atrial fibrillation was 0.84 (95% CI, 0.76 to 0.92) among participants who were still wearing an ECG patch at the time of notification.

This is the key source for consumer-watch AF detection performance and supports the worked calculation in Section 5.

Sanna T, et al. “Cryptogenic Stroke and Underlying Atrial Fibrillation.” New England Journal of Medicine 2014;370:2478–2486. DOI: 10.1056/NEJMoa1313600. Published 2014-06-26. Accessed 2026-04-26. URL: doi.org

CRYSTAL-AF trial findings for an insertable cardiac monitor versus standard follow-up after cryptogenic stroke:

• At 6 months: atrial fibrillation detected in 8.9% of the ICM group versus 1.4% of the control group.
• At 12 months: 12.4% versus 2.0%.
• At 36 months: 30.0% versus 3.0%.

This supports the worked calculation in Section 5 and the implantable monitor comparison in Section 6.

Tison GH, et al. “Passive Detection of Atrial Fibrillation Using a Commercially Available Smartwatch.” JAMA Cardiology 2018;3(5):409–416. DOI: 10.1001/jamacardio.2018.0136. Published 2018-05-23. Accessed 2026-04-26. URL: doi.org

This earlier peer-reviewed study used a deep-learning algorithm with Apple Watch heart-rate data and reported that passive AF detection is feasible; it is weaker than the Apple Heart Study for population-scale validation but is an independent academic lineage.

2.2 Granted patent — recorded claims, not proof of performance

U.S. Patent 8,509,882 B2 — “Heart monitoring system usable with a smartphone or computer.” Assignee: AliveCor, Inc. Grant date reported as 2013-08-13. Accessed 2026-04-26. URL: patents.google.com

This granted patent is cited here as evidence of a recorded claim around a smartphone-connected heart monitor, not as proof that any AliveCor product has a specific sensitivity or specificity. Under the source-tier rules, a patent is Tier 1 for IP/claim existence and Tier 1 for “what was claimed,” but it does not establish clinical validity, reimbursement, or commercial viability.

Note: The grant date and assignee above are high-confidence but were not re-resolved live during this writing session. Treat the date as needs primary source validation until cross-checked against the USPTO Patent Center or Google Patents.

2.3 Regulatory documents

FDA De Novo DEN180044. This is the clearance event for the Apple ECG app. Source: FDA De Novo database, accessed 2026-04-26, accessdata.fda.gov

The De Novo route clears a device for specific indications; it is not proof of outcome improvement. Apple’s ECG app is cleared for use by individuals 22 years of age and older, is described as an over-the-counter single-lead ECG, and is not indicated for replacement of standard monitoring in all settings.

2.4 Standards-body references

• IEC 60601-2-47:2012, medical electrical equipment for ambulatory electrocardiographic systems. This is the principal device-safety/performance standard for ambulatory ECG recorders. The exact publication year and scope should be validated at the IEC Webstore.

URL for validation: webstore.iec.ch — needs primary source validation because the exact document number/revision was not resolved live.

This standard matters for patch and mobile cardiac telemetry hardware, but it does not by itself validate clinical detection claims.

2.5 SEC filings for adoption, reimbursement, and risk cross-check

The following SEC EDGAR query links are primary sources for company filings:

• iRhythm Technologies, 10-K filings: sec.gov
• Medtronic, 10-K filings: sec.gov
• Abbott Laboratories, 10-K filings: sec.gov
• Boston Scientific, 10-K filings: sec.gov
• Apple, 10-K filings: sec.gov
• Koninklijke Philips N.V., SEC filings page: philips.com

These are used to cross-reference company technical claims against filed risk factors. For example, iRhythm’s SEC filings disclose material reliance on reimbursement, coding, and regulatory clearance for its Zio services. Apple’s filings describe health features as subject to regulatory clearances and region-specific availability. No exact 10-K quote is reproduced here because the specific filing text was not retrieved live; the relevant risk-factor direction is marked needs primary source validation for exact wording.

How the technology works — a simplified engineering map

3.1 Adhesive patch architecture

The iRhythm Zio XT is a single-lead, water-resistant, disposable patch worn close to the sternum. The company states it can be worn for up to 14 days.

Source: iRhythm Zio XT page, accessed 2026-04-26, irhythmtech.com

The patch records a single ECG vector, stores data locally, and is mailed back or returned to a processing center. The critical technical constraint is that wear compliance, skin adhesion, and motion artifact gate data yield. Company marketing says “up to 14 days,” but the actual analyzable duration in any given patient may be shorter. SEC risk factors acknowledge that patient compliance and data quality are operational risks.

Source: iRhythm SEC EDGAR listing, accessed 2026-04-26, sec.gov

3.2 Consumer photoplethysmography plus opportunistic ECG

Apple Watch uses an optical heart-rate sensor and a proprietary irregular-rhythm algorithm. If the algorithm flags irregularity, the user can take a single-lead ECG using the watch crown. The FDA clearance is for the ECG app, not necessarily for every downstream algorithm update.

Source: FDA De Novo DEN180044, accessdata.fda.gov

The Apple Heart Study measured the performance of the irregular pulse notification pathway in a real-world enrolled population. The observed 0.52% notification rate and 34% AF yield among patch returnees are the most cited performance figures.

Source: Perez et al., DOI 10.1056/NEJMoa1901183

3.3 Insertable loop recorders

An ICM such as Medtronic Reveal LINQ II is implanted subcutaneously and continuously monitors rhythm for years. Medtronic’s product page states battery life up to 3 years for LINQ II.

Source: Medtronic Reveal LINQ page, accessed 2026-04-26, medtronic.com

The clinical advantage is extended monitoring without patient compliance. The disadvantage is an invasive insertion procedure. The CRYSTAL-AF data quantify the advantage in a cryptogenic stroke population.

Contradictions and cross-checks

Contradiction 1: “34% yield” versus “84% PPV”

The Apple Heart Study simultaneously reports that 34% of patch returnees had AF and that the positive predictive value of an irregular pulse notification for AF was 84%. These figures are not contradictory, but they are often conflated.

• The 34% is the proportion of all 450 participants who returned a patch and had AF on the patch. It includes notifications where the patch was not yet being worn at notification and may include paroxysmal AF missed by the patch if AF was not present during the patch recording.
• The 84% PPV is the fraction of irregular pulse notifications that were confirmed as AF when the participant was wearing the patch at the time of notification.

Both are source-bound to Perez et al. DOI 10.1056/NEJMoa1901183.

Contradiction 2: “Diagnostic” marketing versus informational use

Manufacturers commonly describe consumer ECG tools as capable of producing a “medical-grade” rhythm strip. But the FDA clearance language and the underlying clinical trials do not show that single-lead consumer ECGs are equivalent to 12-lead ECGs for all arrhythmia diagnoses. Apple’s ECG app is cleared for use in individuals 22 and older and does not carry an indication that it replaces standard diagnostic evaluation. Thus, the commercial claim can be broader than the approved use.

Source for clearance: FDA De Novo DEN180044

Source for clinical performance: Perez et al., DOI 10.1056/NEJMoa1901183

Contradiction 3: Continuous monitoring versus real-world use

“Continuous monitoring” is a marketing phrase. For a patch, it means the device is capable of continuous recording for the advertised wear period; it does not mean every minute yields diagnostic-quality data. For an insertable monitor, it is closer to true continuous monitoring but is limited by memory, transmission schedules, and episode-triggering logic. The company product pages support the claimed design, while the clinical data support efficacy only in specific populations.

Quantitative worked calculations

Calculation 1: Insertable cardiac monitor — number needed to test after cryptogenic stroke

Use the CRYSTAL-AF detection rates.

Formula:

\[

NNT = \frac{1}{(\text{AF detection rate}_{ICM} - \text{AF detection rate}_{control})}

\]

Inputs from Sanna et al., DOI 10.1056/NEJMoa1313600:

Calculated NNT:

• 6 months: \(1 / 0.075 = 13.3\)
• 12 months: \(1 / 0.104 = 9.6\)
• 36 months: \(1 / 0.270 = 3.7\)

Sensitivity:

• If the 12-month control detection rate were 5.0% instead of 2.0%, the difference falls to \(0.124 - 0.050 = 0.074\), and the NNT rises to \(1/0.074 = 13.5\).
• If the 12-month ICM detection rate fell to 10.0%, the difference against a 2.0% control would be \(0.080\), and the NNT would be \(1/0.080 = 12.5\).
• Therefore, the NNT is highly sensitive to the control-arm detection rate; this matters because control monitoring intensity varies across cryptogenic stroke programs.

Calculation 2: Apple Heart Study population detection yield

Formula:

\[

\text{Yield per enrolled population} = \text{irregular pulse notification rate} \times \text{AF yield among notified patients who returned a patch}

\]

Inputs from Perez et al., DOI 10.1056/NEJMoa1901183:

• Enrollment: 419,297
• Irregular pulse notification rate: 2,161 / 419,297 = 0.52% = 0.0052
• AF yield among patch returnees: 153 / 450 = 34.0% = 0.340

If every notified participant were patch-confirmed with the same 34% yield:

\[

\text{AF cases} = 2,161 \times 0.340 \approx 735

\]

\[

\text{Population yield} = 0.0052 \times 0.340 = 0.001768 = 0.1768\%

\]

\[

\text{Number of enrolled participants needed per detected AF} = \frac{1}{0.001768} \approx 566

\]

Observed confirmation in the patch subgroup:

\[

\text{Confirmed AF} = 153

\]

\[

\text{Enrolled participants per observed patch-confirmed AF} = \frac{419,297}{153} \approx 2,740

\]

Sensitivity:

• The gap between 566 and 2,740 is driven by patch-return compliance: only 450 of 2,161 notified participants returned patches, or 20.8%.
• If patch return compliance had been 60%, the observed AF count would be approximately \(2,161 \times 0.60 \times 0.34 = 441\), yielding one patch-confirmed AF per \(419,297 / 441 \approx 951\) enrolled participants.
• If the true yield among all notified participants is lower than the patch-returning subgroup because of selection bias, the population yield would be lower than 0.1768%.

This calculation is useful because it separates device performance from compliance and study follow-up.

Parameter table

Unit economics / business model

Who monetizes:

• iRhythm Technologies monetizes a service, not just a device: the Zio patch plus proprietary algorithm plus final ECG report. Revenue is recognized per Zio service delivered. The company bears the cost of the patch, logistics, algorithm development, billing, and reimbursement risk. iRhythm’s SEC filings disclose that revenue depends heavily on Medicare and commercial payer coverage for ambulatory cardiac monitoring. A concrete gross-margin percentage is not reproduced here because it was not live-retrieved; any specific margin number is needs primary source validation.

Source: iRhythm SEC EDGAR listing, sec.gov

• Medtronic, Abbott, and Boston Scientific monetize the insertable loop recorder as a device sell-through to hospitals and clinics, with the monitoring software/platform attached. Margin is nested inside the broader cardiovascular or cardiac rhythm management business, not necessarily broken out in a way that makes remote cardiac monitoring standalone margin reliable. Any specific segment margin must be labeled needs primary source validation unless sourced directly from the relevant 10-K.

• Apple, Alphabet/Fitbit, and Garmin monetize hardware, not a monitoring service. The ECG app and irregular rhythm notification are features that increase device utility and ASP, but the heart-rhythm feature is not separately a subscription. For AliveCor, the KardiaMobile business uses a hardware-plus-premium-subscription model; KardiaCare includes reports and history services, but AliveCor is private, so the unit economics are not covered by a public 10-K.

Who captures value:

The clinical value of a new diagnosis flows first to the patient and treating physician. The economic value flows mostly to the device/service vendor and, in fee-for-service medicine, to the interpreting physician and monitoring service. However, the value capture is contingent on CPT coding and payer coverage. CMS codes for remote physiologic monitoring and ambulatory ECG monitoring govern eligibility. The exact code-language and reimbursement rates were not retrieved live; any specific code number should be verified at the CMS Medicare Physician Fee Schedule search, cms.gov . The code descriptions and rates are marked needs primary source validation because they were not resolved in this session.

The practical implication is that a technically superior monitor can fail commercially if the reimbursement code does not cover the service or if the payer bundles the payment into an existing visit.

Disconfirming Evidence

The investment/scientific thesis for remote cardiac monitoring would be materially weakened by:

1. No clinical outcome benefit: If randomized controlled trials show that AF detection by patches, ICMs, or consumer wearables does not reduce stroke, hospitalization, or mortality, the health-economic case collapses. The CRYSTAL-AF and Apple Heart Study sources demonstrate detection, not outcome improvement.

1. Low real-world positive predictive value in an unselected population: If a general-population wearable screening program produces a low PPV and high downstream healthcare utilization, the cost-harm calculus could turn negative. The Apple Heart Study already shows notification prevalence of only 0.52%, meaning many people must be monitored to find one case; if confirmatory yield outside trial conditions is materially lower, the program is less compelling.

1. Reimbursement collapse: If CMS or private payers reclassify or bundle ambulatory cardiac monitoring into existing visit payments, iRhythm’s and others’ service economics would deteriorate. iRhythm’s SEC filings explicitly identify reimbursement as a risk factor. A negative National Coverage Determination or CPT change is a direct disconfirming event.

1. Algorithm drift or FDA action: If cleared consumer algorithms are shown to perform worse on updated devices or expanded demographics, or if FDA issues safety communications restricting use, the growth narrative would be disrupted.

1. Device performance and compliance degradation: For patch devices, if wear-time in real-world settings is far below the advertised 14 days and analyzable yield drops, the detection advantage over standard Holter monitoring could vanish. For ICMs, if complications from implantation or high false-positive AF burden outweigh benefits, adoption would stall.