Prediction of hip joint load and translation using musculoskeletal modelling with force-dependent kinematics and experimental validation

Musculoskeletal lower limb models are widely used to predict the resultant contact force in the hip joint as a non-invasive alternative to instrumented implants. Previous musculoskeletal models based on rigid body assumptions treated the hip joint as an ideal sphere with only three rotational degrees of freedom. An musculoskeletal model that considered force-dependent kinematics with three additional translational degrees of freedom was developed and validated in this study by comparing it with a previous experimental measurement. A 32-mm femoral head against a polyethylene cup was considered in the musculoskeletal model for calculating the contact forces. The changes in the main modelling parameters were found to have little influence on the hip joint forces (relative deviation of peak value<10BW%, mean trial deviation<20BW%). The centre of the hip joint translation was more sensitive to the changes in the main modelling parameters, especially muscle recruitment type (relative deviation of peak value<20%, mean trial deviation<0.02mm). The predicted hip contact forces showed consistent profiles, compared with the experimental measurements, except in the lateral-medial direction. The ratio-average analysis, based on the Bland-Altman's plots, showed better limits of agreement in climbing stairs (mean limits of agreement: -2.0 to 6.3 in walking, mean limits of agreement: -0.5 to 3.1 in climbing stairs). Better agreement of the predicted hip contact forces was also found during the stance phase. The force-dependent kinematics approach underestimated the maximum hip contact force by a mean value of 6.68 ± 1.75% BW compared with the experimental measurements. The predicted maximum translations of the hip joint centres were 0.125 ± 0.03mm in level walking and 0.123 ± 0.005mm in climbing stairs.