AI Assistance for Critical Illness
(POC-AI-ICU Trial)
What You Need to Know Before You Apply
What is the purpose of this trial?
This trial explores how an AI tool, using a large language model (similar to a super-smart chatbot), can assist doctors in making better decisions for patients in the medical intensive care unit (MICU). The goal is to determine whether this AI tool improves the accuracy of diagnoses and overall patient outcomes compared to standard care. Critically ill patients admitted directly to the MICU from the emergency room or medical wards might be eligible to participate. As a Phase 1/Phase 2 trial, this research aims to understand how the AI tool functions in people and measure its effectiveness in an initial, smaller group, offering participants a chance to contribute to groundbreaking advancements in medical care.
Will I have to stop taking my current medications?
The trial information does not specify whether you need to stop taking your current medications. It seems that usual practice for initial evaluation and management will be followed, so you may not need to stop them.
What prior data suggests that this AI-assisted decision support is safe for ICU patients?
Research has shown that using a large language model like ChatGPT-5 for AI-assisted care in the ICU is generally safe for patients. Studies indicate that these AI tools work with data stripped of personal details to protect privacy. This AI tool provides diagnostic help and treatment suggestions without directly interacting with patients, reducing the risk of harm.
Previous findings suggest that AI systems like ChatGPT-5 are well-accepted in similar healthcare settings. The AI acts as a decision-support tool, offering doctors additional information. There are no reports of the AI itself causing harm. Instead, it aids doctors in making informed decisions, presenting a promising option for improving care without adding new risks.12345Why are researchers excited about this trial?
Researchers are excited about AI-assisted care for critical illness because it introduces a groundbreaking way to support doctors in decision-making. Unlike traditional ICU care, which relies solely on human expertise and experience, this approach uses a powerful large language model (like ChatGPT-5) to analyze patient data and provide a primary diagnosis, differential diagnoses, and treatment suggestions. This AI-driven support can help clinicians consider a wider range of possibilities and potentially improve patient outcomes by offering insights that might not be immediately apparent.
What evidence suggests that AI-assisted decision support could be effective for improving diagnostic accuracy and clinical outcomes in critical illness?
Research has shown that advanced AI tools like ChatGPT-5 can help doctors make better decisions in intensive care units. In this trial, some participants will receive standard ICU care plus point-of-care LLM-based decision support using ChatGPT-5. These tools analyze patient information to assist doctors in diagnosing and treating patients more effectively. Studies have found that AI use can lower the chances of medical mistakes and improve patient health, such as reducing death rates and shortening hospital stays. ChatGPT-5 is especially noted for its ability to quickly and accurately understand complex medical details. Although this technology is still new, early results are promising for improving care in critical situations.12456
Who Is on the Research Team?
Eric Silverman, M.D.
Principal Investigator
MetroWest Medical Center and St. Vincent Hospital
Are You a Good Fit for This Trial?
This trial is for adult ICU patients at certain hospitals. It's not clear who can't join because the exclusion criteria aren't provided, but typically those with conditions that could skew results or put them at risk are excluded.Inclusion Criteria
Exclusion Criteria
Timeline for a Trial Participant
Screening
Participants are screened for eligibility to participate in the trial
Treatment
Participants are randomized to either standard care or AI-assisted care at ICU admission
Follow-up
Participants are monitored for safety and effectiveness after treatment
Data Analysis
Masked chart review to determine primary diagnosis and extract outcomes
What Are the Treatments Tested in This Trial?
Interventions
- Point-of-care large language model decision support (ChatGPT-5)
Trial Overview
The trial tests if AI (ChatGPT-5) can help doctors by giving a diagnosis and treatment advice when patients are admitted to the ICU. Half of the patients will get this AI assistance; the other half won't.
How Is the Trial Designed?
2
Treatment groups
Experimental Treatment
Active Control
Patients receive standard ICU care plus point-of-care LLM-based decision support at admission. De-identified admission data are formatted and submitted to an LLM (ChatGPT-5). The model returns a primary diagnosis, ranked differential diagnosis list, suggested additional information, and prioritized therapeutic recommendations. This output is provided to the admitting team for consideration in ongoing management.
Patients receive usual ICU care per local practice. De-identified admission data may be processed and submitted to the LLM for research purposes, but AI output is not shared with treating clinicians and does not influence real-time management.
Find a Clinic Near You
Who Is Running the Clinical Trial?
MetroWest Artificial Intelligence Research Workgroup
Lead Sponsor
Citations
Study Details | NCT07293078 | Point-of-Care AI Assistance ...
Use of a large language model (ChatGPT-5) to analyze de-identified ICU admission data (history, physical examination, laboratory results, ...
Point-of-Care AI Assistance and Critical Care Outcomes: A ...
Use of a large language model (ChatGPT-5) to analyze de-identified ICU admission data (history, physical examination, laboratory results, ...
Recent approaches of artificial intelligence in intensive ...
Common AI model used in ICU. Real-time clinical decision-making tools have the potential to improve patient mortality and recovery prediction in ...
Impact of large language model (ChatGPT) in healthcare
This review aims to synthesise research findings to consolidate the implications of ChatGPT integration in healthcare and identify research gaps.
Large Language Models Take Center Stage: Potential for ...
Within the field of critical care, LLM-based chatbots can aid in clinical decision support, help with goals of care discussions, and summarize the large data ...
Large language models in critical care - PMC
This breakthrough holds significant promise for critical care medicine, where unstructured data and complex clinical information are abundant.
Unbiased Results
We believe in providing patients with all the options.
Your Data Stays Your Data
We only share your information with the clinical trials you're trying to access.
Verified Trials Only
All of our trials are run by licensed doctors, researchers, and healthcare companies.