Multifactorial estimation of hip and knee joint centres for clinical application of gait analysis.

Estimation of joint centres is required in gait analysis for kinematic and kinetic elaboration. In general, the method is based on regression equations and palpation of bone prominences. A kinematic approach can also be used at the hip joint but appears to be more demanding for the patients. The method presented here combines different independent approaches and is aimed at reducing the estimation errors caused by from skin motion artefacts and statistical variance of anthropometric data. It is based on minimising an optimisation function in which data from gait kinematics, anthropometric measurements and statistically derived morphological parameters are properly weighted. The procedure is suitable for clinical applications where the range of motion might be reduced and the patients can not perform complicated exercises. Validation has been made on ten normal subjects and ten patients with spina bifida. The results show that the reproducibility of the estimation and the congruency with the external measurements is improved with respect to both the morphologically based approach and the kinematic method. Copyright 1998 Elsevier Science B.V.