Postural Sway Parameters are Useful to Discriminate Biomechanical Risk Classes Associated with Load Lifting

Work-related musculoskeletal disorders are a significant health issue in the workplace, contributing to absenteeism, disability, and decreased productivity. These disorders result from biomechanical overload due to repetitive movements and extended actions. Several quantitative and semi-quantitative methods have been developed in the occupational ergonomic field to assess biomechanical risk. Still, they tend to be complex and heavily reliant on the operator's expertise. Recently, wearable sensors coupled with artificial intelligence have proven to be effective for biomechanical risk assessment, offering practical benefits for enhancing worker health and safety. Based on these considerations, this study aimed to explore the feasibility of a logistic regression model to classify two biomechanical risk classes defined by the Revised NIOSH Lifting Equation. The model was built using postural sway parameters extracted from linear acceleration data acquired by a single inertial measurement unit placed on the lumbar region. Eight participants performed two sessions consisting of 20 load lifting tasks, respectively. The results, although preliminary, were promising, indeed the logistic regression model, properly trained with postural sway parameters, achieved an accuracy and an area under the receiver operating characteristic curve equal to 79% and 80%, respectively. Study limitations due to the small sample size and limited age range make this study preliminary. For this reason, future research involving a larger and more varied study population and different work scenarios is needed to validate the effectiveness of the proposed methodology.