An evaluation of predictive methods for estimating cumulative spinal loading

The focus of this study was to assess the amount of error present in severa l approaches that have been commonly used to estimate the cumulative spinal loading during manual materials handling tasks. Three male subjects perfor med three sagittal plane lifting tasks of varying loads and postural requir ements. Video recordings of the tasks were digitized and a biomechanical mo del was used to calculate the spinal loading (compression, joint shear, rea ction shear, and flexion/extension moment) at L4/L5 for each frame of data. The 'gold standard' for cumulative loading experienced by the subjects was obtained by integrating the resultant biomechanical model outputs for the entire lifting cycle. Five approaches that quantify cumulative spinal loadi ng, four that use discrete measures and one that reduces the number of fram es used (5 Hz), were used and compared with the gold standard. The four met hods using discrete measures to quantify the cumulative demands of a task r esulted in substantial errors (average error across task and subjects was 2 7-69%). Reducing the number of frames of data processed to 5 frames/s prese rved the time varying information and was the only approach examined that d id not induce significant error into the cumulative loading estimates. This study indicates that errors in cumulative spinal loading estimates can be large depending upon the approach used, which will hinder any progress in d eveloping a dose-response link between cumulative exposure and an increased risk of low-back pain or injury.