Using pattern recognition to detect differences in movement strategy between high and low relative biomechanical exposure lifts and lifters: Application to backboard lifting

Abstract

Introduction: Backboard lifting is a demanding aspect of paramedic work that results in higher low back moments and sagittal trunk angles compared to other paramedic tasks. Movement strategy in a backboard lift affects resultant biomechanical exposure at the low back so there is a need to identify differences in movement strategies that yield lower relative biomechanical exposure. Pattern recognition methods can be used to objectively identify features of movement related to biomechanical exposure. In particular, principal component analysis (PCA) is a pattern recognition technique that can identify whole body features of movement that explain variance in a data set. This approach is conceptually compatible with Optimal Feedback Control theory (OFC), which provides a theoretical motor control framework in which to contextualize the pattern recognition analysis. Research Questions: 1) How do features of movement differ between high and low relative biomechanical exposure lifts, and as a function of relative demand in backboard lifting? 2) How do features of movement differ between high and low relative biomechanical exposure lifters, and across relative demand in backboard lifting? Methods: Twenty-eight participants performed 10 backboard lifting trials within each of a light, medium and heavy relative demand condition. Relative demands were scaled to participants’ one-repetition max backboard lift. Full body kinematics and ground reaction forces were collected for backboard lifting trials. A whole-body kinematic model was created in Visual3D to calculate low back moments and sagittal angles for dichotomizing lifts and lifters as low vs. high relative biomechanical exposure, and to provide positional data. PCA was applied on positional data as a pattern recognition technique. For retained principal movements (PM), PM scores were calculated as dependent variables. Six PMs were retained for analysis. A two-way ANOVA with independent factors of relative biomechanical exposure and relative demand was used to test for differences in PM scores for retained PMs across all lifts for research question 1. A two-way mixed ANOVA with a between factor of relative biomechanical exposure and a within factor of relative demand was used to test for differences in mean of PM scores in lifters to answer research question 2. Results: Movement strategies associated with high and low relative biomechanical exposure lifts: Significant main effects of relative biomechanical exposure were detected in 5 of the 6 PMs. PMs were interpreted to deduce that low exposure lifts positioned the body closer to the load, used a distal to proximal strategy and maintained an upright trunk. Significant main effects of relative demand were seen in 4 of the 6 PMs. Heavy relative demand lifts were interpreted to have the body further from the load, use a distal to proximal strategy, use a more stoop-like strategy and had differences in timing of the lift. Movement strategies associated with high and low relative biomechanical exposure lifters: High exposure lifters positioned the body further from the load than low exposure lifters. Significant main effects of relative demand were seen in PMs 2 and 5 within lifters, which are interpreted to have a distal to proximal, and more stoop-like strategy in the heavy relative demand lifts. Discussion: The application of a pattern recognition technique identified differences in movement strategies between those who experienced relatively less and greater biomechanical exposure. Pattern recognition also revealed how relative demand influenced movement strategies during backboard lifting. Based on effect sizes, the horizontal distance of the body to the load was the most important determinant of relative low back exposure. The influence of relative demand revealed that a distal-to-proximal strategy was more likely when lifting a heavier relative demand, a finding that is consistent with past literature. The strong relationship of horizontal distance to the load as identified via the pattern recognition approach suggests that some lifters consider biomechanical exposure in their OFC control law by positioning themselves closer to the load. With no significant interaction effects, assessment of backboard lifting can be conducted by evaluating a lifters proximity to the backboard prior to lifting without considering relative demand.