The Jump ARCHES program has awarded $1.2 million in grants to 12 healthcare projects, focusing on data analytics, AI, and predictive modeling. These initiatives aim to improve medical training and healthcare delivery. Combating emergency department overcrowding, AI-assisted regulatory change management, improving medical imaging efficiency, advancing brain disease detection, optimizing medication management, and developing predictive models for diabetic patients are among the projects underway. The program also supports initiatives for health screening, nursing support, and community health equity through digital resources.
The Jump ARCHES program has granted funding to 12 projects focused on enhancing healthcare delivery through data analytics, machine learning, and predictive modeling.
The Jump Applied Research through Community Health through Engineering and Simulation (ARCHES) program has recently announced the allocation of nearly $1.2 million in grants to 12 projects. These initiatives aim to advance medical training and healthcare delivery through the utilization of cutting-edge technologies.
The Jump ARCHES program is a collaborative effort between the University of Illinois Urbana-Champaign (UIUC), the University of Illinois College of Medicine in Peoria (UICOMP), and OSF HealthCare. This program’s funds assist clinicians and engineers in the development of transformative technology and devices, overcoming healthcare barriers related to location, age, and social determinants of health.
The first project, titled ‘STREAM-ED: Simulation to Refine, Enhance, and Adapt Management of Emergency,’ focuses on developing models that integrate event simulation, machine learning (ML), and optimization techniques. The aim is to address challenges associated with overcrowding and resource utilization in emergency departments.
Another project introduces a prototype Intelligent Regulatory Change Management (IRCM) system. This system harnesses artificial intelligence (AI) and natural language processing (NLP) to monitor and evaluate public policy actions that affect OSF HealthCare. The IRCM system empowers compliance staff to determine appropriate actions and identify crucial changes to enhance quality, safety, and privacy risk management.
The project ‘Machine Learning of Standardized DICOM Metadata from Imaging Datasets’ aims to create an ML-based algorithm capable of categorizing image parameters in 2D medical images based on signal intensity variations. This innovation improves the efficiency of medical image segmentation and enables the automated characterization of previously unknown 3D imaging datasets.
In collaboration with OSF HealthCare Children’s Hospital of Illinois, another project employs imaging segmentation to generate 3D models of neuroblastic tumors for pre-surgical planning. The transition from 2D imaging to 3D modeling enhances staging analysis reproducibility and improves neuroblastic tumor models.
‘Toward Machine-Learned Aortic Arch Measured Diameters’ aims to automate the segmentation and clinical measurement of aortic arch diameters from MRI imaging. Encouraging initial results have already shown a significant correlation between standard clinical measurements and automated measurements.
The subsequent proposal seeks to advance brain disease detection and monitoring using deep learning (DL) and multimodal brain imaging data. The project aims to develop ‘brain atlases’ for AI-based brain image analysis and computational solutions for automated tumor detection. Existing research has demonstrated that DL improves the identification of brain tumors.
‘Optimizing Pharmacologic Management of Behaviors in Patients with Autism’ involves the creation of an ML model that assists clinicians in selecting appropriate medications and dosages for patients with Autism Spectrum Disorder (ASD). This model incorporates the patient’s genetic information, medical history, and clinician notes to adapt treatment protocols over time.
Another project focuses on improving medication adherence for patients with Type 2 diabetes. By utilizing predictive analytics, this initiative identifies individuals at high risk of medication non-adherence, enabling clinicians to develop strategies to address the underlying causes and enhance adherence among their patients.
‘The Knowledge Graph Construction with Large Language Models to Predict DKA Occurrence and Severity’ project aims to develop a predictive model that employs named entity recognition and language modeling-based knowledge graphs. The objective is to identify high-risk diabetic patients and prevent diabetic ketoacidosis (DKA).
Additional projects include a field experiment evaluating the effectiveness of health kiosks supported by community health workers for delivering first-line preventive health screening in underserved areas. There is also an effort to develop a personalized nursing support app to combat high nurse turnover. Lastly, a proposal is put forth to establish community health cafés that provide digital access to healthcare resources, promoting health equity.
