AI-Assisted Instruction Versus Peer Learning for Sustained Lumbar Puncture Competency

This study compared three teaching methods for training resident physicians in lumbar puncture, a critical medical procedure used to collect spinal fluid for diagnosing conditions like meningitis. One hundred thirty-five residents were randomly assigned to either traditional bedside teaching, peer-assisted learning (where trainees practiced in pairs and gave each other feedback), or artificial intelligence (AI)-assisted training using a smart simulator that provided real-time voice feedback on technique. All residents were tested immediately after training and again three months later to see how well they retained their skills. Right after training, all three groups performed equally well. However, after three months, both the peer-learning and AI groups scored significantly higher on practical skills than the traditional group, showing that these newer methods help slow skill decay. The AI group showed the greatest advantages in procedural standardization, pre-procedure assessment, post-procedure care, and operating speed. Meanwhile, residents in the peer-learning group reported the highest satisfaction with classroom interaction and teaching methods. In short, while all three approaches work well for short-term training, AI-assisted and peer-assisted learning are more effective than traditional teaching at helping residents maintain their lumbar puncture skills over time, with AI excelling in technical precision and efficiency, and peer learning offering a more engaging and collaborative learning experience.