The use of Rasch analysis as a tool in the construction of a preference based measure: the case of AQLQ

The majority of quality of life instruments are not preference-based measures and so cannot be used within cost utility analysis. The Asthma Quality of Life Questionnaire (AQLQ) is one such instrument. The aim of this study was to develop a health state classification that is amenable to valuation from the AQLQ. Rasch models were applied to samples of responders to the AQLQ with the aim of i) selecting a number of items for a preference based utility measure (AQL-5D), ii) reducing the number of levels for each item to a more manageable number of levels for establishing AQL-5D. Selection of items for the evaluation survey was supported with conventional psychometric criteria for item selection (feasibility, internal consistency, floor and ceiling effects, responsiveness and regression against overall health). The role of Rasch analysis in reducing the number of item levels to a preconceived target number of levels proved unsuccessful. However, Rasch analysis proved to be a useful tool in assisting in the initial process of selecting items from an existing HRQL instrument in the construction of AQL-5D. The method is recommended for use alongside conventional psychometric testing to aid in the development of preference-based measures.