There is always exciting research being conducted at IC3 labs. Find out more about our exciting research or read one of our many published articles below.


CHoRUS (bridge2ai)

A Patient-Focused Collaborative Hospital Repository Using Standards (CHoRUS) for Equitable AI

While AI in critical care (AICC) has helped improve patient care, AICC research lacks a large generalized dataset, standards that support the exchange and use of health information (interoperability) between sites and researchers, and support from a workforce trained in AI and general public familiar with medical AI. CHoRUS, a multimillion dollar collaboration of 18 top medical research institutions, aims to solve these problems by creating a 100,000 patient dataset from a sample population with varied backgrounds, ethical standards that enable interoperability, and AI education for AICC researchers and the general public. The expanded AICC ecosystem and education efforts will help improve critical care for all patients.


Intelligent Intensive Care Unit (I2CU)

Pervasive Sensing and Artificial Intelligence for Augmented Clinical Decision-Making

Although close monitoring and dynamic assessment of patient acuity are key aspects of ICU care, both are limited by time constraints imposed on healthcare providers. Dynamic and precise assessment of patient acuity relies almost exclusively on physicians’ clinical judgement and vigilance, and details indicating patient condition may be partially captured or completely missed by overburdened nurses. We aim to develop deep learning models for sensing, quantifying, and communicating any patient’s condition in an autonomous, precise, and interpretable manner. This tool for autonomous and precise patient quantification may enhance clinical workflow and early intervention.



Autonomous Delirium Monitoring and Adaptive Prevention

Delirium prevention in the ICU is often achieved with non-pharmacological approaches, but some methods are difficult to implement due to their dependence on sporadic human observations. ADAPT, the Autonomous Delirium Monitoring and Adaptive Prevention system, will autonomously quantify two indicators of delirium (patient mobility and circadian desynchrony) and predict delirium trajectories, as well as prompt actions that decrease delirium risk factors (e.g., ambient light, nightly disruptions). Successful application of ADAPT would augment clinical decision-making and promote more targeted interventions.



Digital Rehabilitation Environment Augmenting Medical System

The purpose of Digital Rehabilitation Environment Augmenting Medical System (DREAMS) is to research the feasibility and clinical potential of an immersive digital reality-augmenting system in reducing the occurrence of cognitive, behavioral, and emotional consequences of critical illness (e.g., pain, anxiety and insomnia) and environmental exposures (noise & light) that are risk factors for the development of delirium, a common and devastating complication in the intensive care unit (ICU).

DREAM project, AR with medical research

Pain recognition

Autonomous Pain Recognition in Non-Verbal and Critically Ill Patients

The under-assessment of pain response is one of the primary barriers to adequate treatment of pain in critically ill patients and is associated with many negative outcomes. However, pain intensity ratings are limited by ICU patients’ ability to self-report and manually-administered, potentially variable behavioral pain scales. Our objective is to develop and validate pain recognition algorithms that autonomously assess facial pain expressions and patient activity and quantify patient pain. We rationalize that this autonomous pain assessment system will reduce nurse workload and enable real-time pain monitoring.



Automated Algorithm Identifies and Communicates Risk of Acute Kidney Injury among Health Care Providers and Patients

Acute kidney injury (AKI) is a sudden reduction in kidney function that may go unrecognized and is associated with increased adverse events. We will develop and implement an automated algorithm that feeds findings to the EHR in real-time, allowing AKI diagnosis and severity to be communicated to healthcare providers using interface on EPIC platforms for early management for AKI. We will assess its impact on knowledge, awareness and communication about AKI among healthcare providers.