what is the ai/qi program?
Our Approach
The AI/QI team drives innovation within UF's Quality and Patient Safety Initiative by creating and validating AI models that enhance patient safety and clinical outcomes. Working directly with UF Health and Shands clinicians, we rigorously test these solutions to address real healthcare challenges. Our work leverages UF's unique resources to improve access to care in Gainesville and across Florida. We combine AI expertise with clinical insights, fueled by our commitment to preventing harm and serving every patient.
featured projects
At QPSi, our projects focus on areas where we can make meaningful progress to improve patient safety and clinical quality. The featured projects below, including Milk+ and Kidney Stone Care Optimization, showcase examples of our work in action. Each project begins with questions or challenges brought to us by clinicians, guiding our efforts toward solutions that will be most useful in practice. While these are just a few highlights, they reflect areas where our expertise and tools can make a measurable difference for patients and care teams. Our approach emphasizes thoughtful, innovative solutions developed with compassion and clinical rigor.
Helping Mothers Succeed with Personalized Support
MILK+
Milk+ (Maximizing Initiatives for Lactation Knowledge) is an AI-driven clinical decision support tool designed to help providers identify new mothers at risk of lactation challenges. By analyzing factors such as medical history, delivery details, and infant characteristics, Milk+ helps care teams provide personalized, timely support tailored to each mother’s unique situation. This project highlights how thoughtful use of AI can empower families, improve early nutrition, and build confidence during a critical time.
Improving Kidney Stone Recovery with AI
Kidney Stone Care Optimization
Roughly 15 to 20% of patients nationwide return to the emergency department or require unplanned care after a kidney stone procedure. This project combines AI and process engineering to reduce that number by predicting which patients are most at risk and improving care coordination. By forecasting demand and identifying when and where care will be needed, the model helps reduce emergency room visits and avoid delays. It’s a practical example of how AI can make the system work more efficiently while still focusing on patient comfort and timely treatment.
ai/qi program iMPACT
see the impact of our project pilots in a clinical setting
Lives Touched by AI/QI Projects
Each data point represents a patient benefiting from our AI application in real clinical settings. These thousands of encounters showcase where providers accessed our models to enhance patient care decisions. Behind every number is a real person whose treatment was informed by patterns identified from anonymized past experiences.
categorized by florida health priorities
AI/QI Projects by Domain
Our projects directly address the most pressing health challenges facing Floridians. By aligning our AI/QI initiatives with identified state health priorities, we ensure our resources target the needs of those we serve. This chart shows how we distribute our efforts across critical areas such as chronic diseases, maternal health, mental well-being, and emerging infectious threats. Each slice represents the number of projects we have dedicated to that priority category, reflecting our data-driven approach to community needs. We are guided by what matters most to Florida families and driven to create AI solutions that improve care for everyone, especially our most vulnerable communities.
Scaling AI Performance with Supercomputing Power
Accelerating AI with HiPerGator
Our work leverages UF’s HiPerGator supercomputing infrastructure to process massive healthcare datasets faster and more efficiently. Using tools like NVIDIA RAPIDS, we've dramatically improved processing times, going from 12 hours for 70 features in 2023 to just 1 hour for over 25,000 features by late 2024. This graph shows the growth in standard extracted health indicators over time, reflecting both the expanding scale of our AI models and the performance gains made possible through optimized computing. With 3 to 5 model runs per week, we’re saving more than 12 hours per run as we turn complex data into actionable insights faster than ever.
Contact Us!
Want to know more about the AI/QI Program? Our team is available to answer questions or provide guidance about the AI/QI program's innovations or projects.