Usefulness of Structural Equation Modeling in Evaluation of Prognosis of Total Hip Replacement for Coxarthrosis

Abstract

In rehabilitation medicine, it is important to understand the hierarchical structure of patients' disturbances as well as causal relations within the structure. This study developed hierarchical structural disturbance models based on patients' function, ability and health-related QOL (HR-QOL) for coxarthrosis treated by total hip replacement (THR). Whether or not the structural equation modeling (SEM) analysis could depict the disturbances structure clearly and predict the prognosis properly was investigated. We used a set of longitudinal data for 92 patients who received THR for coxarthrosis from 1995 to the end of 2000 and who had no missing data of the Japanese Orthopaedic Association's Hip Score (Hip Score) as well as Manual Muscle Testing (MMT) conducted before and 3, 6, 12 and 24 months after the surgery. Subsequently we obtained a set of cross-sectional data for 283 patients who received THR in 2001 or later, including results of the SF-36 test that rated HR-QOL, and used it to determine the relational structure among patients' function, ability and HR-QOL. SEM analysis with stepwise modification of disturbance models resulted in a model in which a patient's function explained about 70% of a patient's ability and a patient's ability explained 60% of HR-QOL. The SEM analysis used in this study made it possible to understand the disturbances structure hierarchically. An SEM analysis that allows disturbances to be understood structurally might be useful in considering patient care after THR and could be applicable to cases with other diseases. It will be necessary to consider how to represent categorical data numerically for model fit improvement. Better graphic representation should also be considered.