Faculty of Medicine: 5 successful projects in 2024 Artificial Intelligence Seed Funding Program

This year’s competition, made possible by the generous contributions of donors Joseph and Amy Ip, builds on the momentum established since the program's launch in 2020. By supporting forward-thinking research and training, the AI Seed Funding Program enhances the Faculty’s capacity to integrate AI into healthcare, advancing knowledge and improving patient care outcomes. 

The program awarded a total of $50,000 to five outstanding proposals, with each receiving $10,000 to support. These projects, led by exceptional Principal Investigators from across the Faculty and affiliated hospital Research Institutes, represent diverse approaches to harnessing AI’s transformative potential in medicine: 

• Matthew Henderson (Dept. of Pathology and Laboratory Medicine/CHEO) 
  AI for Congenital Hypothyroidism Screening
  Congenital hypothyroidism (CH) is a preventable cause of intellectual disability in newborns. Early detection is crucial for normal development. The current screening method in Ontario, which measures thyroid stimulating hormone levels, often results in false positive screening determinations, leading to unnecessary diagnostic follow-up. This project aims to improve CH screening using machine learning. By leveraging extensive data from Newborn Screening Ontario, the project team developed a machine learning model that doubles the positive predictive value of initial screenings while maintaining 100% sensitivity. This project aims to further improve this advancement and offer a framework for applying machine learning to other rare disease screening programs.

• Gregory Hundemer (Dept. Medicine/OHRI) 
  Learning from an Artificial Intelligence Implementation at the Ottawa Hospital
  This project builds on a newly developed and validated state-of-the-art artificial intelligence-based prediction model that was designed to reduce the rates of unplanned dialysis among patients living with advanced chronic kidney disease. Despite the lack of established pipelines and very few precedents to guide the clinical deployment of this model, the project team is charting a path towards implementation and seek to translate new knowledge gained along the way. Through this project, the team aims to not only to improve patient outcomes, but also steer the future of ubiquitous artificial intelligence pipelines at the Ottawa Hospital.

• Kheira Jolin-Dahel (Dept. Family Medicine/ ISM) 
  Comparison of AI Generated and Human Generated Feedback Reports of Family Medicine Resident Research Project Proposals at the Question and Timeline (QT), Short Report (SR), Progress Report (PR) and Final Report (FR) Stages – An Assessment and Evaluation Project.
  The objective of this project is to assess and evaluate the use of AI in generating an initial draft feedback report for the QT, SR, PR and FR to be reviewed and edited by the Department of Family Medicine Research Team with the intention of reducing the laps time between submission and the issuance of high-quality feedback reports and test the implementation of a hybrid human-centered GenAI-assisted feedback report process.

• Ran Klein (Dept. Medicine/OHRI) 
  Precommercial Deployment of Medical Image Processing Automation Server (MIPAS)
  As AI and medical image processing algorithms continue to advance, integration with clinical systems continues to be a barrier to transitioning these applications from proof-of-concept to clinical utility. We have been developing the Medical Image Processing Automation Server (MIPAS), which can seamlessly integrate novel AI image processing applications into the clinical workflow. This has been demonstrated as a prototype at the Department of Nuclear Medicine and Molecular Imaging, integrating in-house developed and open-source AI into the clinical workflow. This project aims to enhance the implementation of MIPAS to prepare it for routine clinical deployment, wide-spread dissemination, and possibility of future commercialization.
• Jerry Maniate (Dept. Medicine/OHRI/Bruyère) 
  The CARE-AI Study: Creating Accountable and Responsive Ethics for Artificial Intelligence in Healthcare
  AI’s prevalence in healthcare offers benefits such as enhanced patient care and treatment options, but also introduces ethical and privacy challenges. Guidelines from 2011, designed for social media, do not address AI's complexities. Our research aims to update these guidelines to cover AI-specific ethical, privacy, and bias issues. Collaborating with experts, this project will use focus groups and the modified the Delphi method to refine these guidelines. The goal is to create principles that ensure responsible, ethical, and fair use of AI in healthcare, led by the Equity in Health Systems Lab at the Bruyère Research Institute.