The Pulse
GE HealthCare Launches AI System to Forecast Hospital Bottlenecks
GE HealthCare has announced CareIntellect for Operations, an AI-enabled software service that forecasts hospital capacity constraints up to 72 hours ahead. The Queen’s Health Systems and Duke Health will be the first clinical evaluation sit
AI.info Team ·
GE HealthCare says hospitals should stop waiting for capacity problems to appear before responding to them. On September 15, the company announced CareIntellect for Operations, an AI-enabled software service designed to forecast operational constraints up to 72 hours in advance and recommend actions tied to beds, staffing, patient movement and clinical services.
The announcement places GE HealthCare’s product against a problem hospitals still often handle through fragmented data gathering and manual escalation. The company says operational leaders can spend several hours each day pulling information from multiple systems, reconciling it and trying to determine which pressure point needs attention first. CareIntellect for Operations is intended to produce a single view of emerging strain across units, departments and an entire health system.
The product is not yet being presented as a widely deployed system with independently demonstrated results. The Queen’s Health Systems and Duke Health will be its first clinical evaluation sites, where frontline teams will provide feedback that GE HealthCare says will inform later enhancements.
Two models target pressure before it spreads
CareIntellect for Operations analyzes hundreds of patient-level and operational data points, including bed availability, patient delays, staffing, wait times and ancillary services. It draws on electronic medical records and other resource-management systems to identify potential constraints across hospitals, departments and units.
Its Pressure Forecast model focuses on system demand and resource strain. GE HealthCare says it can project census, staffing pressure, emergency-department boarding, therapy demand, procedural-recovery capacity and inbound-transfer activity up to 72 hours ahead.
A second model, called Estimated Day of Discharge, examines longitudinal patient data to predict when individual patients are likely to leave the hospital. The estimate updates hourly as new information becomes available. The application then connects those forecasts with recommended actions and prioritized workflows intended to help teams address capacity pressure before it affects more parts of the hospital.
Queen’s and Duke will test the operational case
The first deployments will take place at The Queen’s Health Systems and Duke Health. Alex Wroe, chief operating officer at The Queen’s Medical Center, said the health system wants to use the software to improve access and make better use of its beds, teams and other resources.
“We’ve seen firsthand how this type of advanced operational software can help us free up more capacity and better use our resources. We look forward to taking this work to the next level, helping us move beyond reacting to capacity challenges and toward proactively ensuring that our beds, teams, and resources are available for the patients who need them most.”
Alex Wroe, Chief Operating Officer, The Queen’s Medical Center
Duke Health’s participation builds on a three-year partnership with GE HealthCare, according to Katie Flanagan, the health system’s associate vice president for patient flow and care coordination. She said the three-day view could help teams move more quickly from analysis to action.
“Managing demand and capacity across our system in a coordinated, predictable way is essential to delivering exceptional care. CareIntellect for Operations builds on three years of partnership with GE HealthCare, giving our teams a three-day view of what’s coming so they can move faster from analysis to action, for the benefit of our patients.”
Katie Flanagan, Associate Vice President, Patient Flow and Care Coordination, Duke Health
GE HealthCare builds on Command Center software
The new application draws on more than two decades of GE HealthCare experience in hospital operations. The company says its Command Center software is used by nearly 500 hospitals and medical facilities globally.
GE HealthCare cites customer-reported results from Command Center deployments, including up to $20 million in first-year cost savings, a reduction of more than one day in length of stay and capacity to serve 19,000 additional patients annually. Those figures come from separate case studies and are not presented as results from CareIntellect for Operations. The company also cautions that outcomes can vary by user, electronic medical-record system and level of adoption.
That distinction matters because the new product’s first phase is an evaluation, not a completed outcomes study. Hospitals will need to determine whether the forecasts are accurate in their own settings and whether recommended actions fit local staffing, clinical and governance processes.
AWS availability puts the product in procurement channels
CareIntellect for Operations is part of GE HealthCare’s broader CareIntellect family of clinical and operational applications. The company says the applications share a cloud-first infrastructure intended to let customers add future products without repeating a separate product-by-product integration process.
The operations application runs on Amazon Web Services and is available for purchase directly from GE HealthCare and through the AWS Marketplace. Taha Kass-Hout, GE HealthCare’s global chief science and technology officer, said the system is designed to make its forecasts understandable and connect them to decisions hospital staff already make.
“Health systems need to know where pressure is building and what to do next. CareIntellect for Operations brings together patient and operational signals to give teams a 72-hour view of emerging constraints and help them act earlier.”
Dr. Taha Kass-Hout, Global Chief Science and Technology Officer, GE HealthCare
The immediate test will come at Queen’s and Duke: whether a 72-hour forecast gives hospital teams enough time to adjust discharges, imaging queues, transfers, staffing and other services before a local delay becomes a broader access problem.