Comparison of the underlying constructs of the EQ-5D and Oxford Hip Score: implications for mapping.

OBJECTIVES: Both disease-specific and generic patient-reported outcome measures provide information about the health status of patients. Generally, disease-specific measures provide more clinical information than generic measures but do not provide a utility weight. The aim of this study was to assess the comparability of the information captured by a disease-specific measure, the Oxford Hip Score (OHS), and a generic measure, the EQ-5D, and the viability of mapping between them to obtain utilities for the OHS. METHODS: Data for 439 National Health Service patients in England before and 6 months after undergoing total hip replacement were analyzed. The information provided by the OHS and EQ-5D was assessed using principal component analysis and analysis of the correlation matrix. The predictive performance of four mapping models was based on the mean absolute error. RESULTS: The results of the exploratory and confirmatory principal component analyses showed that the OHS data can be associated with three constructs relating to pain, mobility, and usual activity. Compared to the EQ-5D, the OHS items are multidimensional and the same construct is detected by different OHS items. These differences between the OHS and the EQ-5D do not impede the merits of either instrument when used for their own purposes. CONCLUSIONS: Conceptual differences between the two instruments mean that mapping is unlikely to provide an appropriate basis for estimating utilities for the OHS.