Scalable Centralized Continuous Cardiac Index Analyses of Remotely Streamed Arterial Waveforms

Sunday, March 22, 2026

9:45 AM - 10:45 AM Central Time

Location: Connections Central - RST 02

Jonathan Castro, MD

Marymount Hospital

Disclosure(s): No relevant financial relationship(s) to disclose.
MO

Melinda Otero, MSN, RN, CCRN

Program Manager, eHospital
Cleveland Clinic, Ohio

Disclosure information not submitted.
DA

Dandrea Ashley, MSN, RN

Nurse Manager, eHospital
Cleveland Clinic

Disclosure information not submitted.
AB

Abigail Brown, APRN, DNP

Director, Critical Care Transport
Cleveland Clinic

Disclosure information not submitted.
SM

Steven Minear, MD

Medical Director, SICU, Cleveland Clinic, Weston, FL
Cleveland Clinic, Weston, Florida

Disclosure information not submitted.
CU

Chiedozie Udeh, MBBS, MHEcon, MBA, FCCP, FASA, FCCM

Medical Director, eHospital, Cleveland Clinic
Cleveland Clinic, Cleveland, Ohio

Disclosure information not submitted.

Introduction: Non-invasive cardiac output (NICO) monitors are increasingly used for continuous cardiac Index (CI) while limiting risks of invasive procedures. These devices use proprietary models to analyze arterial waveforms, with or without disposable accessories. However, high costs, tight bedside spaces, location tracking and storage constraints limit wider use. One such device, Retia Medical’s Argos, uses the Multi-Beat Analysis (MBA) algorithm which requires patient age, gender, height and weight, but no disposables. After 20 seconds of data flow, it updates CI every 5 seconds. This study assessed the technical feasibility of streaming arterial waveforms to a remote server for continuous centralized parallel analyses of CI for multiple patients using the MBA model.

Methods: The MBA algorithm was installed on a Virtual Machine (VM) within Cleveland Clinic’s data network. An existing server for streaming waveforms for ICU Telemedicine was reconfigured to also send arterial waveforms to the VM. For phase one, the VM was checked regularly and patient demographics manually added for newly received waveforms. CI results were sent via web-based interface to the tele-ICU and could also be seen via secure remote access. Outcome metrics were transmission time, analytical time, and uptime performance. Exceeding current bedside practice of recording CI hourly, the proportion of times the CI was updated every 15 minutes while waveforms were being streamed was also measured.

Results: Over 8 weeks, 346 patients’ waveforms were sent to the MBA server with varying concurrency based on bedside decisions. Simulated transmission time was < 1 sec. After input of demographics, median (range) dwell time of arterial catheters was 139 hours (1 minute – 701 hours). The median (range) count of opportunities for CI every 15 minutes was 556 (1- 2804). Overall, CI was calculated 53,547 times out of 53,765 opportunities as defined (99.6%). Uptime reliability was 99%.

Conclusions: Arterial waveforms were successfully streamed to a centralized remote server for parallel continuous calculation of CI using the MBA model. Transmission times were negligible. Continuous CI was updated reliably. This centralized approach could allay the constraints of bedside devices and make NICO available to more patients.