Comparison of the Hara, Harrington, and Davis hip joint center regression equations for gait analysis in children with cerebral palsy.

Abstract

Regression equations have been widely accepted in defining the location of the hip joint center in clinical gait analysis. Equations by Hara et al. require a single anthropometric measurement of leg length, allowing easy implementation in a clinical setting. This study evaluated the regression equations by Hara et al. in comparison to common equations by Harrington et al. and Davis et al. for clinical gait analysis in children with cerebral palsy. The location of the hip joint centers was defined by each of the aforementioned three models and compared for retrospective gait analysis data of 30 children with cerebral palsy (13 females, age 5-16 years). Gait kinematics at the hip and knee across those models were further compared. The location of the Hara hip joint center was defined in between the other two models: ∼1 cm anterior and medial with almost the same height as Harrington's model. Gait kinematics reflected the location of the hip joint centers, placing the curves from the Hara model in between the others. The root mean square differences between the Hara and the other models were within 2.6 degrees in all planes of motion. The regression equations by Hara et al. demonstrated similarity in the location of the hip joint center and gait kinematics to other common equations. The Hara model is simple, appears less susceptible to errors associated with soft tissue, and might be a good alternative option for clinical gait analysis in children with cerebral palsy.