No new hardware. No retrofitting. Just an intelligent layer that unlocks proactive, precision care from systems hospitals already own.

Predict risk. Prevent loss.

• eCART is a research-based, AI-driven software as a medical device (SaMD) that utilizes a machine learning algorithm to continuously assess hospitalized patients’ risk of impending death or intensive care unit (ICU) transfer, assisting medical staff in swift and accurate recognition of patients requiring increased medical attention. eCART draws upon readily available patient data from the EHR, rapidly quantifies disease severity, and predicts likelihood of critical illness onset, hours before it happens.
• eCART received FDA 510(k) clearance in 2024, with clinical performance data of nearly two million hospitalizations from 21 hospitals, including an unprecedented level of real-world prospective data, to ensure consistent accuracy across geography, age, race, and top medical conditions (e.g. sepsis, COPD, heart failure). 
• Built upon more than a decade of ongoing scientific research, eCART was developed at UChicago Medicine by Dr. Dana Edelson, MD, MS and Dr. Matthew Churpek, MD, PhD and is well chronicled in published literature. The analytic and workflow have been in use in clinical practice since 2016.

At AgileMD, we are building the most advanced real-time predictive analytics and clinical pathways platform for hospitals.

To date, AgileMD tools have been used on more than 4 million patient encounters by more than 135,000 clinical users in over 250 hospitals, including some of the nation's leading academic medical centers. AgileMD's eCART clinical deterioration prediction software is FDA cleared and the first AI-powered early warning device for ward patients to go through the FDA process with real-world prospective data. 

AgileMD is backed by premier start-up partners MATTER, YCombinator, and Rock Health, and our cutting-edge analytics are built on more than decade of ongoing research at the University of Chicago. In 2021, AgileMD garnered nearly $3 million in combined federal research and development funding from the U.S. Department of Health and Human Services (HHS), including the National Institutes of Health (NIH) and the Biomedical Advanced Research and Development Authority (BARDA).

Unlike traditional EHR systems that primarily document patient information for retrospective review, Wavital delivers real-time, proactive insights. EHRs typically rely on manual charting, delayed vital updates, and rule-based alerts; Wavital continuously streams high-fidelity vitals directly from bedside monitors, uses predictive AI instead of static thresholds, and generates actionable alerts within seconds — not hours.

Wavital is patient-level, device-agnostic, and cloud-native, making it faster to deploy, easier to scale, and far more effective at detecting early signs of clinical deterioration — without needing massive customization or "bolt-on" EHR modules.

Wavital uniquely combines real-time, high-fidelity streaming from any monitoring device (vendor-agnostic) with predictive AI that delivers patient-specific alerts in seconds — not retrospective dashboards or delayed reviews. Unlike competitors that often require complex IT setups, work only with curated device kits, or are locked into proprietary ecosystems, Wavital offers plug-and-play compatibility, cloud-native deployment, and continuous local model refinement using each hospital’s own patient data. It’s the only platform designed to turn any bed into an ICU-grade monitoring station — without ripping and replacing existing systems.

• Continuously monitors EHR data in real time
• Stratifies hospitalized patients by risk of impending all-cause clinical deterioration and sepsis
• Notifies inpatient providers when patient may need additional assessment and management
• Recommends clinical pathways with integrated order entry

• Authorized: FDA 510(k) cleared
• Tested: Derived and validated in nearly 3 million patient encounters across 28 hospitals
• Experienced: 8+ years in clinical use; 10+ years of published academic literature on value over other early warning scores
• Accurate: High sensitiviy and specificity, meaning fewer false alarms and more efficient focus on the highest risk patients
• Integrated: Embedded cross-team workflow ensures that risk scores are not ignored and action is taken
• Benchmarked: Robust reporting, benchmarking, and implementation support drive quality improvement