Clinical Evaluation and Assessment of AI-enabled Transfusion Decision Support Systems for Personalized Trauma Care: Phase II Study

Anticipated timeline and budget

Background

Hemorrhage with trauma‑induced coagulopathy (TIC) is a major cause of preventable death in trauma. Early blood product use is key, but identifying patients needing massive transfusions, or at risk for TIC is challenging, especially in austere settings. AI and ML models show promise in predicting transfusion needs from trauma data but are mostly retrospective, binary, and triage focused. Many of these models are considered pre prospective data validation prototypes, defined as “Phase I, while few progress to Phase II, which involves validation and implementation.

This funding opportunity seeks “Phase II” studies to prospectively validate, which would include Phase I prototype model validation, refinement with prospective data, as well as integrating real-time data streams with Point-of-Care (POC) systems, such as ROTEM (Rotational Thromboelastometry), FloPatch, POC INR (International Normalized Ratio). Proposals should focus on exploring and assessing the limitations of these models via validation, as well as identifying methods to improve, refine, and generalize them, thereby contributing to broader efforts in medical translational research.

Research objectives

• Validate and refine existing pre- and in-hospital transfusion triage models using prospective data.
• Validate and refine transfusion decision-making models focusing on: Trauma-induced coagulopathy (TIC), Massive haemorrhage protocol (MHP) activation, massive transfusion requirements, and Dynamic blood product needs over time
• Develop real-time models using hemodynamic and coagulation data from POC systems to guide blood product selection and timing throughout trauma resuscitation.

Desired Outputs and deliverables

Phase II studies should be built on the previously completed Phase I study where prototype AI/ML models were developed to support personalized transfusion in trauma care. The work will focus on developing a deployment framework and further validating and refining the models using both retrospective and prospective data, evaluating their ability to predict MHP activation, transfusion requirements, and patient outcomes within the first 24 hours post-injury, thereby supporting future clinical implementation and follow-on studies. Applications should present a clearly defined and achievable research plan for each year. Ultimately, the proposal will reach the stated research objectives.

During the funding period, the research team must:

• Submit regular progress reports in a timely manner as required.
• Model validation and assessment to reduce blood product wastage, optimize resource utilization, and support early (e.g. pre-hospital, emergency room) clinical decision-making.
• Model validation and assessment to support effective in-hospital transfusion decision-making and hemorrhage control.
• Collection of real-time POC datasets (ROTEM, FloPatch, and/or POC INR) for model development, with de‑identified data shared with DRDC TRC.
• Development of a prototype real-time transfusion decision-making model.

Applicant Qualifications and Requirements

• Possess prototype transfusion decision-making models for validation and refinement 
• Be affiliated with at least one Level 1 trauma centre.
• Include a multidisciplinary team with expertise in trauma resuscitation, AI/ML, and clinical trial design, led by a senior MD with trauma care expertise.
• Have access to high-performance computing resources and large military or civilian trauma datasets. 
• Demonstrate a clear and feasible plan for model validation, refinement, and clinical deployment.
• Maintain an established and functional data-sharing mechanism with DRDC TRC.
• Collaborate with the DRDC TRC (Toronto Research Centre) research team.
• Attend project meetings to provide updates and share preliminary results.
• Submit regular progress reports in a timely manner as required.
• Research ethics board approval from an accredited Canadian institution will be required prior to the release of funding.

Application deadline

Please download and submit the Research Funding Application form

Applicants are advised to include eight weeks of administrative lead-time in their timelines. This will allow for ethics board submission and approval (if required), and the drafting and signing of the funding agreement.

Enquiries

Questions about this funding opportunity can be sent to the VAC Research office at research-recherche@veterans.gc.ca