AI-ENHANCED ICU SURVEILLANCE SYSTEM
Predictive Patient Safety for Critical Care Environments

“Predict, prevent, and protect - AI guided care for any environment.”

CATEGORY

AI-Powered Monitoring • Critical Care • Patient Safety • Software as a Medical Device (SaMD)

OPERATIONAL RELEVANCE

Unplanned falls, self-extubation, agitation, and pressure injuries remain top threats in ICUs and combat casualty care. This AI-enabled system uses multimodal video, audio, and patient-specific data to predict critical events before they occur. It supports prolonged field care, trauma stabilization, and fixed ICU settings, reducing complications, increasing staff efficiency, and enhancing survival.

TECHNOLOGY SUMMARY

This non-invasive system employs bedside video and audio monitoring with advanced AI to continuously analyze posture, motion, speech, and vitals. It predicts falls and extubation attempts minutes in advance and scores agitation and wellness trends using multiple ML models including CNNs, and RNNs.In addition, the system tracks patient lying postures, along with the duration and transition patterns of each position, to ensure timely pressure relief and prevent the development of bedsores. Custom alerts help caregivers intervene proactively, while a dashboard centralizes data across patient beds.

MILITARY-ALIGNED BENEFITS

• Predicts high-risk ICU events before they occur (fall, extubation, agitation, pressure ulcer formation)
• Supports remote triage and centralized oversight in deployed or overwhelmed care units
• Enhances prolonged field care decision-making through passive patient monitoring
• Enables safer handoff between evacuation stages
• Requires no wearables or additional patient interaction

USE SCENARIOS

• Combat support hospitals or Role 3 facilities
• Forward surgical teams with limited nurse-to-patient ratios
• Critical care aeromedical transport
• VA hospitals and military rehab centers
• Long-term care monitoring with AI-supported alerting

TECHNOLOGY READINESS & IP

• Technology: LSUHSC-S-2024-008
• Technology Readiness Level (TRL): 2 – Preliminary algorithm components developed; full integration and prototype pending
• Patent landscape assessed; multimodal AI design unique among competitors

COLLABORATION OPPORTUNITY

We are seeking partners to:

• Co-develop and validate the system in ICU and field environments
• Run clinical pilots for predictive fall and extubation event detection
• Integrate with existing patient monitoring platforms or EHR systems
• Explore joint development under SBIR, STTR, or CRADA agreements