Real-Time Intelligent Coaching Versus Human Expert Instruction in Surgical Skills Training: A Systematic Review and Meta-Analysis.

Real-time intelligent coaching systems are technology-based surgical training tools that provide automated, real-time, and objective feedback on learners' performance during simulated surgical tasks. Unlike human expert instruction, which relies on instructor-delivered guidance and feedback, these systems use automated performance monitoring to provide immediate feedback on aspects such as accuracy, motion economy, and errors. This systematic review and meta-analysis investigated the effectiveness of real-time intelligent coaching systems compared with human expert training in surgical skill instruction, with technical surgical performance as the primary outcome and cognitive load and emotional outcomes as secondary outcomes. We followed the PRISMA guidelines and conducted a comprehensive search of PubMed, the Cochrane Library, ClinicalTrials.gov, and ScienceDirect through March 2026. We considered randomized controlled trials including medical students with no prior surgical experience. We measured outcomes using the Intelligent Continuous Expertise Monitoring System (ICEMS) and the Objective Structured Assessment of Technical Skills (OSATS) scores, emotional responses, and cognitive load. We included three randomized controlled trials involving 281 participants. No statistically significant differences were observed between real-time intelligent coaching systems and human expert instruction in technical performance, as measured by ICEMS and OSATS scores; however, the ICEMS analysis showed very high heterogeneity (I² = 99%), and only two trials reported technical outcomes. Real-time intelligent coaching systems were associated with a small reduction in overall cognitive load compared with human expert instruction, with the intrinsic cognitive load domain also favoring real-time intelligent coaching systems. No significant differences were observed in positive activating or negative activating emotions, whereas positive deactivating emotions were modestly higher with real-time intelligent coaching systems. Given the small number of studies, limited sample sizes, and substantial heterogeneity, interpret the findings cautiously. Current evidence suggests that real-time intelligent coaching systems may provide technical performance similar to human expert instruction and may offer some cognitive benefits, but further well-designed, adequately powered trials are needed to establish their effectiveness and role in surgical education.