Quantifying Learning Curves in Ultrasound Training: A Real-Time Consultation Analysis Using a Novel Half-Life Metric

Background Competency-based training in clinical ultrasonography requires a clear understanding of skill acquisition trajectories. However, objective data on how residents actually learn in real-time clinical environments remain scarce. This study aims to map the learning pathway of ultrasound residents by analyzing real-time consultation patterns and introducing a novel metric—the skill acquisition half-life—to quantify the rate of skill mastery. Methods We conducted a single-center retrospective observational study in a high-volume outpatient ultrasound department. Over an eight-week period, real-time consultation requests initiated by 45 residents across three postgraduate years (PGY-1 to PGY-3) were logged. Consultations were categorized by examination type (Standard vs. Advanced assignments). Adjusted consultation rates were calculated per clinical session. A negative exponential learning curve model was fitted to derive the skill acquisition half-life (T₁/₂), defined as the time required for the consultation rate to decrease by half. Results A total of 529 consultations were recorded, with PGY-1 residents accounting for 88.5% of requests. Consultation patterns evolved from foundational examinations (e.g., pelvic and abdominal ultrasound) in PGY-1 toward complex, judgment-dependent procedures (e.g., Renal artery and Soft tissue ultrasound) in PGY-3. Learning curve analysis revealed distinct skill acquisition half-lives: pelvic ultrasound had the shortest T₁/₂ (0.33 years), followed by thyroid & cervical lymph nodes (0.51 years) and Soft tissue ultrasound (0.63 years). Renal artery ultrasound exhibited the longest half-life (2.00 years), indicating prolonged learning needs. Conclusions This study demonstrates that ultrasound skill acquisition follows a quantifiable, staged trajectory, effectively mapped through real-time consultation data. The skill acquisition half-life provides a novel, data-driven metric to benchmark procedural difficulty and inform staged curriculum design. These findings support the integration of real-time supervision systems not only as educational tools but also as sources of actionable intelligence for competency-based training in clinical ultrasound.