Pilot evaluation of a novel, automated ergonomics assessment tool

Abstract Gastroenterologists are prone to endoscopy-related musculoskeletal injuries (ERI). Current interventions lack real-time monitoring and feedback. ErgoGenius, a novel artificial intelligence computer-vision tool, addresses this gap by providing continuous posture assessment and feedback without wearable motion trackers. The aim of this study was to determine the feasibility of ErgoGenius, its accuracy compared with human appraisers, and its ability to detect abnormal posture. The study was conducted at two large academic centers. The Rapid Entire Body Assessment (REBA) score was used as a surrogate for ergonomic performance and risk of injury. Ten endoscopists of varying gender, height, and weight were recorded performing endoscopic tasks in optimal vs. lowered bed positions. Videos were analyzed by ErgoGenius. A paired t-test was used to compare REBA scores between bed positions. ErgoGenius was successfully deployed in a controlled endoscopy setting. ErgoGenius achieved perfect internal agreement (rho = 1) and closely correlated with human appraisers (rho = 0.987). Average REBA scores were notably higher in the lowered bed position (mean 4.64) compared with the optimal position (mean 2.55), (P= 0.006). ErgoGenius was successfully deployed to detect abnormal postures related to changes in bed position and quantify ERI risk. It performed at par with human appraisers. This tool shows promise in enhancing ergonomic practices among gastroenterologists and trainees, potentially leading to better health outcomes and reduced injury.