Related Experiment Video
Updated: Jan 13, 2026
Using Virtual Reality to Transfer Motor Skill Knowledge from One Hand to Another
Published on: September 18, 2017
Evidence of Skill Transferability From Training in Virtual Reality Robotic Surgery Simulators to Nonvirtual
Krishnamurti Ma Sarmento1, Walter B Vidal2, Hugo Na França1
1Laboratory of Experimental Surgery (CIEX), School of Health Sciences (Faculdade de Ciências da Saúde), Campus Universitário Darcy Ribeiro, Brasília, DF 70910-900, Brazil.
Objectives:
Synthesize evidence on (1) transfer of skills acquired with virtual reality robotic surgery simulators (VR-RSS) to nonvirtual settings, (2) equivalence or noninferiority versus nonvirtual training, and (3) downstream clinical outcomes.
Design:
PRISMA-conformant systematic review with a prospectively registered protocol. Risk of bias was appraised with RoB 2 (RCTs) and ROBINS-I (nonrandomized studies); certainty of evidence used GRADE with an imprecision adaptation (decision-threshold confidence intervals and scaled optimal information size).
Setting:
Searches of major bibliographic databases and grey sources through September 2025; studies conducted in academic skills labs, simulation centers, or clinical environments, with nonvirtual outcome assessments (bench/animal/cadaveric tasks, simulated or live operations, or patient outcomes).
Participants:
Medical students, surgical residents, or practicing surgeons trained on commercially available VR-RSS devices and evaluated on nonvirtual tasks or clinical endpoints.
Results:
Twenty-five studies met criteria (13 randomized, 5 nonrandomized comparatives, 7 single-group pre/post). Across designs, VR-RSS training improved proximal nonvirtual performance-faster task completion, fewer errors, and higher GEARS/OSATS/GOALS scores. Two randomized trials found no important difference versus control; these null results were internally coherent and did not contradict the overall direction of benefit. Evidence for translation to live operative performance or patient-level outcomes was limited and heterogeneous. Risk of bias varied-generally lower in recent randomized trials-while imprecision was pervasive due to small samples, absent confidence intervals, and undefined minimally important differences. Overall certainty was low-moderate for educational outcomes and low for clinical endpoints.
Conclusions:
VR-RSS improves proximal nonvirtual technical performance and merits integration within structured curricula. Claims of equivalence to nonvirtual training and of patient-level benefit remain provisional. Preregistered, adequately powered trials with blinded, standardized nonvirtual endpoints and explicit decision thresholds are needed to establish educational and clinical value.
