This study investigated whether a machine-learning based system could monitor distal radius fracture (DRF) rehabilitation and provide real-time biofeedback to reduce secondary injury. Investigators at the University of Melbourne performed a single-center, prospective development study which recruited 20 healthy male and female volunteers, with mean age of 41 years, to execute 10-second wrist-motion sequences while hand-motion data was recorded. The investigators utilized a Leap Motion depth camera and forearm electromyography (Myo armband). A personalized musculoskeletal model and finite element method (FEM) based DRF healing model to examine how force distribution impacts healing. The model output closely matched simulated healing in a volunteer, enabling vibration alerts when predicted cartilage formation fell below threshold. This prototype integrated muscle activation, grip strength, and motion dynamics to generate individualized feedback in real time. The results also displayed that different wrist motions stimulated distinct tissue response, highlighting the need to tailor exercise type to healing stage. The training dataset was small and derived entirely from healthy adults, however, limiting generalizability and fairness assessment across ages or fracture types. The work is peer-reviewed and has not yet been tested in real patients.