Nancy A Baker1, Nancy B Sussman, Mark S Redfern. 1. Department of Occupational Therapy, University of Pittsburgh, 5012 Forbes Tower, Pittsburgh, PA 15260, USA. nab36@pitt.edu
Abstract
INTRODUCTION: Identifying postures and behaviors during keyboard use that can discriminate between individuals with and without musculoskeletal disorders of the upper extremity (MSD-UE) is important for developing intervention strategies. This study explores the ability of models built from items of the Keyboard-Personal Computer Style instrument (K-PeCS) to discriminate between subjects who have MSD-UE and those who do not. METHODS: Forty-two subjects, 21 with diagnosed MSD-UE (cases) and 21 without MSD-UE (controls), were videotaped while using their keyboards at their onsite computer workstations. These video clips were rated using the K-PeCS. The K-PeCS items were used to generate models to discriminate between cases and controls using Classification and Regression Tree (CART) methods. RESULTS: Two CART models were generated; one that could accurately discriminate between cases and controls when the cases had any diagnosis of MSD-UE (69% accuracy) and one that could accurately discriminate between cases and controls when the cases had neck-related MSD-UE (93% accuracy). Both models had the same single item, "neck flexion angle greater than 20 degrees ". In both models, subjects who did not have a neck flexion angle of greater than 20 degrees were accurately identified as controls. CONCLUSIONS: The K-PeCS item "neck flexion greater than 20 degrees " can discriminate between subjects with and without MSD-UE. Further research with a larger sample is needed to develop models that have greater accuracy.
INTRODUCTION: Identifying postures and behaviors during keyboard use that can discriminate between individuals with and without musculoskeletal disorders of the upper extremity (MSD-UE) is important for developing intervention strategies. This study explores the ability of models built from items of the Keyboard-Personal Computer Style instrument (K-PeCS) to discriminate between subjects who have MSD-UE and those who do not. METHODS: Forty-two subjects, 21 with diagnosed MSD-UE (cases) and 21 without MSD-UE (controls), were videotaped while using their keyboards at their onsite computer workstations. These video clips were rated using the K-PeCS. The K-PeCS items were used to generate models to discriminate between cases and controls using Classification and Regression Tree (CART) methods. RESULTS: Two CART models were generated; one that could accurately discriminate between cases and controls when the cases had any diagnosis of MSD-UE (69% accuracy) and one that could accurately discriminate between cases and controls when the cases had neck-related MSD-UE (93% accuracy). Both models had the same single item, "neck flexion angle greater than 20 degrees ". In both models, subjects who did not have a neck flexion angle of greater than 20 degrees were accurately identified as controls. CONCLUSIONS: The K-PeCS item "neck flexion greater than 20 degrees " can discriminate between subjects with and without MSD-UE. Further research with a larger sample is needed to develop models that have greater accuracy.
Authors: Nancy A Baker; Rakié Cham; Erin Hale Cidboy; James Cook; Mark S Redfern Journal: Clin Biomech (Bristol, Avon) Date: 2006-10-18 Impact factor: 2.063
Authors: Fredric Gerr; Michele Marcus; Cindy Ensor; David Kleinbaum; Susan Cohen; Alicia Edwards; Eileen Gentry; Daniel J Ortiz; Carolyn Monteilh Journal: Am J Ind Med Date: 2002-04 Impact factor: 2.214