☰

Cardiomatics

Connecting ECG interpretation to the hospital workflow

Industry
Digital Health and Medical Device Software
Headquarters
Kraków, Poland
Public information as of
January 2026

A4BEE prepared this analysis from publicly available sources. It reflects our own reading of Cardiomatics's published strategy and is not endorsed by, or produced in cooperation with, Cardiomatics. Company website

Strategic priorities

Cardiomatics runs an AI interpretation engine for long-term ECG recordings, certified under the EU Medical Device Regulation (the European rule that governs medical devices placed on the EU market) and operating as software-as-a-medical-device from a Kraków engineering base. The company supports more than 60 ECG recording devices, serves hospitals and clinics across Europe, and is preparing for entry into the US cardiac rhythm management market, sized by FactMR at roughly $4.1 billion.

Three near-term pressures shape the platform. First, interpretation is offered on an advisory basis, so every result still needs a qualified clinician's sign-off before it reaches the patient record, which limits the practical speed gain from AI. Second, the platform runs as a standalone web application rather than embedded inside Epic or Cerner, the two dominant US hospital record systems, which slows adoption by Tier 1 (large, multi-specialty) hospital networks. Third, a competitive gap has opened against US incumbents such as AliveCor, which holds 39 FDA-cleared (US Food and Drug Administration market-authorisation) determinations, while Cardiomatics remains concentrated in European certification.

The data backbone is mature in places and thin in others. The ISO 27001-certified (international information-security management standard) cloud handles interpretation at 99.2 percent sensitivity and 92.4 percent precision for cardiac resynchronization therapy (CRT) assessment, but clinical research data from University Hospital Basel, Medical University of Warsaw and other partners sits in disparate systems, and the manual Log In → Upload → Download workflow remains the daily interface for high-volume labs. Closing those seams is the work the next phase of the platform is built around.

Funding to date is roughly $3.2 million across a 2021 seed round and Polish National Centre for Research and Development grants, with headcount reported between 23 and 31 in recent profiles. Public statements from the CEO place AF Burden quantification (measuring the percentage of time a patient spends in atrial fibrillation over a monitoring window, typically 28 days), pediatric cardiology and US market entry at the centre of the 2025-2030 roadmap.

Challenges we see

  • Operations Integration

    Moving from manual upload to automated ingestion at the lab

    Cardiomatics' documentation describes a Log In → Upload → Download workflow as the routine path for ECG recordings, requiring clinician action for every patient file. The company also publishes a 90-minute fast-track service tier for urgent cases, indicating that standard turnaround is the daily baseline against which urgent work is measured.

    Where the upload sits between the recording device and the cloud, each transfer becomes a discrete manual step, and the population that needs to be handled is every recording, not the ones that flag as urgent. Routing the device output into the platform without that intermediate hand-off moves the bottleneck upstream.

  • Digital Integration

    Embedding the interpretation result into Epic and Cerner workflows

    Cardiomatics operates as a standalone web application rather than as a SMART on FHIR application inside Epic or Cerner, the two dominant US hospital record systems. Many European hospitals still rely on HL7 v2 messaging (an older but still widely deployed hospital data-exchange standard) alongside the newer FHIR R4 (fourth release of the FHIR standard).

    Where the result of an interpretation has to be carried across systems by hand, the work that ought to be a clinical decision becomes a clerical one, and the credentialing needed to run inside a Tier 1 EHR (large hospital electronic health record) environment sits ahead of the integration itself.

  • Operations Manufacturing

    Keeping 60-plus ECG device integrations current as the fleet grows

    The platform supports more than 60 ECG recording devices from multiple manufacturers, including legacy multi-lead Holters (portable devices that record heart rhythm continuously over 24 hours or more) and OEM partnerships such as the May 2024 Biotronik agreement in Germany. Each new device family brings its own driver, data format and synchronisation behaviour.

    Where every new device is a custom integration, the engineering effort scales with the size of the fleet rather than with the value of the new clinical signal, and a standard connector layer moves that cost off the critical path.

  • Compliance Regulatory

    Bridging the gap between AI speed and clinician sign-off

    All AI interpretations are offered on an advisory basis and require confirmation by a qualified professional before clinical action, per regulatory requirements for the EU-MDR certification. Published evaluation shows sensitivity of 99.2 percent and precision of 92.4 percent on the CRT assessment task.

    Where the bottleneck sits between a high-confidence result and a confirmed one, the value of the underlying AI is gated by the capacity of the reviewing clinician, which puts the routing of borderline cases at the centre of the workflow rather than the model itself.

  • Compliance Regulatory

    Closing the regulatory gap with US-cleared competitors

    US competitors such as AliveCor hold 39 FDA-cleared determinations covering specific arrhythmia findings, while Cardiomatics remains concentrated in European certification and focuses on a narrower set of determinations, including 2nd- and 3rd-degree AV block (interruption of the electrical signal between the heart's upper and lower chambers). The US cardiac rhythm management market is sized at $4.1 billion by FactMR.

    Where clearance drives which determinations can be marketed in the US, the gap between an EU-MDR portfolio and an FDA 510(k) portfolio is the gap between the markets each one can be sold into, and the submission pipeline ahead of entry is what determines that.

Opportunities, by urgency and business impact

Each bubble is one opportunity, numbered to match the list below. Further right means it bites sooner; higher means a bigger effect on the business. A bigger bubble means a bigger implementation effort.

Source: A4BEE analysis of public sources

  1. Automated ECG data pipeline from the recording device to the cloud

    Cardiac labs upload recordings through a web portal, with several manual steps per patient, and the company runs a 90-minute fast-track service tier to compensate for that hand-off in urgent cases.

    An edge gateway in the lab that detects a connected Holter, extracts the raw signal and pushes it to the cloud over an encrypted tunnel removes the manual transfer from the routine path and makes the existing fast-track tier the default rather than the premium.

    • Cardiomatics, The 3 biggest challenges in analysing ECG signals, 2024
    • Cardiomatics, Fast-Track 90-Minute Service tier documentation

  2. SMART on FHIR integration into Epic and Cerner

    Cardiomatics runs as a standalone web application, so a clinician using Epic or Cerner has to leave the patient chart, sign in separately and copy results back across systems.

    Building the application as a SMART on FHIR component, with HL7 v2 bridging for European sites still on the older standard, puts the interpretation inside the native workflow and turns the application into something the hospital can install rather than something the cardiologist has to remember to open.

    • Itirra, Epic and Cerner EHR Integration Explained: SMART on FHIR vs Backend Systems
    • OSP Labs, How to Implement FHIR with Epic, Cerner, and other EHR/EMR Platforms

  3. Unified research data lake across Basel, Warsaw and partner sites

    Validation studies with University Hospital Basel, Medical University of Warsaw and additional French partners generate fragmented datasets across disparate systems, slowing algorithm validation and the regulatory submission cycles that depend on those studies.

    A federated research data lake with role-based access, an automated annotation pipeline for rare arrhythmia cases, and study-level audit trails lets multi-site evidence be assembled once and re-used across MDR maintenance, FDA submissions and paediatric algorithm development.

    • Cardiomatics, Paving the way for better ECG analysis in children
    • Cardiomatics, Knowledge base news

  4. Standardised device integration framework for the next 60 devices

    The platform already covers more than 60 ECG device types, and each new device family currently arrives as a custom integration of driver, format converter and synchronisation logic.

    A modular device integration framework built on OPC UA (Open Platform Communications Unified Architecture, a vendor-neutral industrial communication standard) and MTP-compliant (Module Type Package, a modular automation standard) connectors moves new devices onto a template rather than a bespoke build, so onboarding time and reliability stop moving with the size of the fleet.

    • Cardiomatics, The 3 biggest challenges in analysing ECG signals, 2024
    • Cardiomatics–Biotronik partnership announcement, May 2024

  5. AI confidence routing for advisory interpretation

    Every AI interpretation is offered on an advisory basis and confirmed by a clinician before it reaches the patient. Published evaluation shows sensitivity of 99.2 percent and precision of 92.4 percent, so most results are high-confidence and a minority are borderline.

    A confidence-scoring layer that routes high-confidence results into a streamlined review path and surfaces borderline cases with the algorithmic reasoning attached reduces the time each confirmed result takes without changing what the clinician is asked to sign off on.

    • Cardiomatics evaluation published in NIH PMC8778735, 2022

What we'd propose

  • Digital Lab

    Smart edge gateway for cardiac labs

    An on-premise device that detects a connected Holter monitor, extracts the raw ECG signal and pushes it to the Cardiomatics cloud over an encrypted tunnel, taking the manual upload steps out of the routine workflow and putting the recording directly into the interpretation queue.

    • Auto-detection of connected recorders

      Recognising what is plugged in

      Network and USB scanning that identifies a connected Holter or patch recorder by family, applies the matching ingestion profile and registers the device with the cloud without operator action, so a new recorder is usable on the same day it is connected.

    • Encrypted signal tunnel to the cloud

      Secure path from device to platform

      TLS-encrypted (Transport Layer Security, the standard encryption protocol for data in transit) upload channel from the edge gateway to the interpretation engine, with ISO 27001-aligned (international information-security management standard) key handling and audit logging so the data path satisfies the same controls as the cloud backend.

    • Real-time ingestion status dashboard

      Visibility of every recording in flight

      Operational dashboard showing recorder connectivity, ingestion queue depth, failed uploads and average end-to-end latency, so the lab team sees the same view of the pipeline that the engineering team sees.

    • The Log In → Upload → Download workflow becomes the exception, not the daily path.
    • The 90-minute fast-track tier becomes the default experience for urgent and routine work alike.
    • Lab staff see the same ingestion state the engineers do, so triage moves out of email.

    We've done this before

    See how we do Digital Lab
  • Digital Lab

    FHIR-native EHR integration platform

    A SMART on FHIR application layer that puts Cardiomatics interpretation directly into Epic and Cerner clinical workflows, with HL7 v2 bridging for European sites still on the older messaging standard, so the report reaches the chart without leaving the EHR (electronic health record) the clinician is already using.

    • FHIR R4 API and SMART on FHIR application

      Standards-based entry into the EHR

      REST APIs (web-service interfaces that exchange data over HTTP) conforming to HL7 FHIR R4, exposed through a SMART on FHIR application so the EHR launches Cardiomatics as an embedded component with the clinician's existing sign-on.

    • Patient chart integration and result push

      Reports in the chart, not next to it

      Native widgets that surface the AI interpretation, the clinician confirmation and the audit trail inside the patient's chart, eliminating the copy-back step and giving the cardiologist a single screen for decision and sign-off.

    • HL7 v2 bridge for European sites

      Older message standards, same destination

      Protocol translation layer that receives HL7 v2 messages from sites still running the legacy standard and converts them into the same FHIR-based data model, so European deployments reach the same workflow as the US build.

    • The application becomes something a hospital installs rather than something a cardiologist remembers to open.
    • Interpretation and sign-off live in the same screen, which shortens the path from result to action.
    • The same integration pattern reaches European sites on HL7 v2 and US sites on FHIR R4.

    We've done this before

    See how we do Digital Lab
  • Enterprise AI

    Unified research data lake across partner sites

    A federated data platform that brings together validation studies from University Hospital Basel, Medical University of Warsaw and additional French partners, with role-based access, automated annotation of rare arrhythmia cases and study-level audit trails suitable for MDR maintenance and FDA submissions.

    • Federated multi-site data layer

      Studies brought under one access model

      Role-based access controls over the partner datasets, with each site retaining local governance while a single query layer reaches all of them, so multi-site evidence is assembled once and re-used across submissions.

    • Automated annotation pipeline

      Labelling rare arrhythmias at scale

      Machine-learning-assisted annotation of rare arrhythmia cases, with reviewer sign-off captured as part of the audit trail, so the bottleneck in training data acquisition moves from manual labelling to reviewer throughput.

    • Regulatory documentation engine

      Evidence packages assembled from the data

      Generation of audit-ready clinical evidence packages with full traceability from a study record to the regulatory filing it supports, covering MDR maintenance, FDA 510(k) submissions and post-market surveillance updates.

    • Multi-site evidence is assembled once and re-used across MDR, FDA and post-market filings.
    • Rare-arrhythmia labelling scales with reviewer capacity rather than with dataset size.
    • Regulatory documentation arrives from the data layer rather than being compiled after the fact.

    We've done this before

    See how we do Enterprise AI
  • Digital CDMO

    Modular device integration framework

    A standardised integration architecture for ECG recorders, built on OPC UA information models and MTP-compliant connectors, so the next device family connects as a configured instance rather than as a custom integration project.

    • OPC UA device adapters

      One protocol for the whole fleet

      OPC UA adapters for ECG recorders, exposing signal data, device metadata and synchronisation events in a documented, vendor-neutral form so the interpretation engine reads every device through the same interface.

    • Data normalisation engine

      Proprietary formats into one signal model

      Automated transformation of proprietary recorder formats into a single digital signal model compatible with the interpretation engine, with format conversion verified end-to-end against representative recordings.

    • Device onboarding toolkit

      New hardware on a template, not a project

      Template-driven onboarding workflow with self-service configuration for new device families, including test corpora and acceptance criteria, so the integration time and the engineering effort stop scaling with fleet size.

    • Onboarding time for a new recorder moves from a custom build to a configured instance.
    • Reliability targets are set on the integration framework rather than on each individual driver.
    • New partnerships, including the Biotronik distribution agreement, attach to the same connector pattern.

    We've done this before

    See how we do Digital CDMO
  • Agents

    AI agents for MDR and FDA documentation work

    Narrow, reviewable agents that take the repetitive part of regulatory documentation: drafting MDR clinical-evaluation summaries from study records, pre-populating FDA 510(k) sections, mapping a standards change to every controlled document it touches, and assembling post-market surveillance updates on a schedule. A named reviewer approves every output.

    • Drafting from source records

      First drafts from the study data

      Generation of the first draft of a clinical-evaluation summary, a 510(k) section or a periodic safety update report directly from the underlying study records and submission templates, so the author edits and judges rather than assembles.

    • Template and completeness checking

      Gaps found before review

      Pre-review check of a submitted document against its MDR or FDA template and the site's own checklist, returning missing sections and inconsistent cross-references before the document enters the human review queue.

    • Change impact search across the document set

      Which documents a standards change touches

      When a harmonised standard, an EU-MDR guidance or an FDA guidance is updated, retrieval of every controlled document that references it, ranked by how directly it is affected, so the scope of an update is known on day one.

    • Regulatory documentation arrives at review complete and consistent with the template.
    • The scope of a standards change is established by search rather than by recollection.
    • Reviewer capacity is spent on judgement rather than on assembly.

    We've done this before

    See how we do Agents

Digital maturity: today and target

Scored out of 100 across six dimensions. The target is what Cardiomatics's own published ambition implies — not a perfect score.

Source: A4BEE analysis of public sources
Data integration 45 → 85
The Log In → Upload → Download workflow remains the daily path for ECG recordings and the 90-minute fast-track service exists to compensate for it, so ingestion is still a manual step rather than a continuous pipeline.
System interoperability 35 → 90
The platform runs as a standalone web application rather than as a SMART on FHIR component inside Epic or Cerner, which is the form most Tier 1 hospitals require, and the HL7 v2 / FHIR R4 split across European sites is unresolved.
Cloud infrastructure 75 → 90
The ISO 27001-certified cloud delivers interpretation at 99.2 percent sensitivity and supports 60-plus device integrations, with the remaining work concentrated in edge connectivity and multi-region deployment.
Analytics and AI 80 → 95
The interpretation engine holds the world's first EU-MDR certification for AI-based ECG interpretation, with precision reported at 92.4 percent on the CRT assessment; the next layer is confidence routing and explainability for EU AI Act transparency.
Process automation 40 → 80
Every AI interpretation is confirmed by a qualified professional before clinical action, and the bottleneck sits between a high-confidence result and a confirmed one, so the routing of cases is the central constraint rather than the underlying model.
Research platform 50 → 85
Validation studies with University Hospital Basel, Medical University of Warsaw and other partners sit in disparate systems, with no federated layer connecting them across MDR, FDA and paediatric programmes.

Check this yourself

Our Service Portal has free self-assessments and market comparisons. These are the ones that line up with what we've read above — no sales call required.

Think we've read this right?

Talk to us

Related reading

This is an independent analysis prepared by A4BEE from publicly available information as of January 2026. It reflects A4BEE's own interpretation and opinion, is not affiliated with, endorsed by, or verified with Cardiomatics, and may be incomplete or inaccurate. All company names and trademarks are the property of their respective owners. To request a correction or removal, contact [email protected].