MDSCS Radiology Companion

• 

  MDSCS Approach to Human-Centered Safe Pre-imaging Support

Inspiration

In the past three weeks, two close family members experienced sudden injuries — one slipped and fractured ribs, and another fell from stairs. I was far away and trying to help only through phone calls and unclear photos, while they did not yet have access to emergency care. Those first few minutes before reaching imaging are critical and can prevent secondary harm — yet people currently have no guided support. At the same time, my daughter, a Radiology Resident at the University of Pennsylvania, described how radiology work depends not only on reading images, but also on clinical context, reflective reasoning, and avoiding cognitive overfitting. This connected directly with my research in Multi-Dimensional Strategic Curriculum Studies (MDSCS), which focuses on how reasoning and judgment develop through guided interaction. From this intersection of real clinical need + personal experience + educational research, the idea for a human-centered AI Radiology Companion emerged.

What it does

This project provides two complementary functions:

1. Patient Pre-Imaging Guidance Helps patients capture safe, clear, standardized injury photos before professional radiology, reducing the risk of making the injury worse and improving initial communication. 2.Radiology Reasoning Companion Provides radiology trainees with a private reflective case journal that supports clinical judgment, context-awareness, and prevents cognitive overfitting. Important: This system does not diagnose. It supports safe action and reflective thinking — not medical decision replacement. ## How we built it Designed the reasoning model and user workflow based on MDSCS cognitive development principles. Defined safety + guided photography steps based on radiology resident input. Prototyped a mobile UI workflow for: (1) Take photo → (2) Safety guidance → (3) Secure private storage. Outlined a structured reflective reasoning prompt system for trainees. ## Challenges we ran into -Ensuring strict boundaries so that the AI does not generate diagnostic claims. -Designing guidance simple enough for patients in pain or stress to follow. -Balancing privacy, security, and personal case-based learning. ## Accomplishments that we're proud of -Created a unified model that connects patient experience with radiology learning. -Demonstrated how AI can support reasoning instead of replacing clinical expertise. -Built a cross-disciplinary bridge between healthcare, education, cognitive science, and human-centered design. ## What we learned -Created a unified model that connects patient experience with radiology learning. -Demonstrated how AI can support reasoning instead of replacing clinical expertise. -Built a cross-disciplinary bridge between healthcare, education, cognitive science, and human-centered design. ## What's next for MDSCS Radiology Companion -Implement the first mobile MVP with safe image capture and private journaling. -Partner with radiology residency programs for pilot testing. -Extend the reasoning model into broader health education and first-response guidance.

Built With

• coded
• figma-(prototype-only)-?-mvp-concept
• not