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, ldquo;neck flexion angle greater than 20°rdquo;. In both models, subjects who did not have a neck flexion angle of greater than 20° were accurately identified as controls.Conclusions
The K-PeCS item ldquo;neck flexion greater than 20°rdquo; can discriminate between subjects with and without MSD-UE. Further research with a larger sample is needed to develop models that have greater accuracy.