Inter- and intra-observer reliability of calculating cumulative lumbar spine loads

The aim of this study was to quantify the precision of manual video digitization of three typical industrial tasks, as evaluated by the comparison of four cumulative kinetic parameters at the L4/L5 intervertebral joint: compression, joint shear, reaction shear and moment. Ten observers were recruited (five male and five female), with an undergraduate background in human anatomy. On each of three test days, each observer digitized five repeats of each of three typical industrial lifting tasks of 5 to 6 s in duration. A rigid link segment model that incorporated a single muscle equivalent model was used to calculate the cumulative loading based on the digitized coordinates. Inter-observer reliability was assessed using a mixed model ANOVA, and no significant differences were found to result from observer, gender, day or trial. Intraclass correlation coefficients (ICC) were calculated within each task to quantify intra-observer reliability. Overall, the ICCs were excellent (>0.75), with the exception of moderate values for reaction shear for Tasks 2 and 3. Compression and moment demonstrated the highest reliability of the four parameters studied, which is beneficial from an ergonomic standpoint, as compression is the most commonly used parameter for job assessments. This study demonstrated manual video digitization to be a reliable tool for the quantification of cumulative spinal loading, both within a given observer, and across days, trials and observers.