Clinical trial health outcome data intended for the purposes of economic evaluation must be capable of being represented by a single, summary index. When benefits are to be estimated in terms of quality-adjusted life years that index needs to represent social preference weights/utilities. However clinical trials often fail to include utility-weighted measures in their protocols. OBJECTIVE: To describe the strategic options available to the researchers' facing the task of deriving a utility weighted index from a condition specific measure, using the FACT-L as a case study. METHODS: The FACT-L is a widely used instrument that defines health-related quality of life in terms of four dimensions: physical, social, emotional and functional well-being. As far as is known, utility weights for the items/levels within the FACT-L have not so far been published. The complexity of the FACT-L makes it impossible to estimate utility weights using traditional MAUT methods. Hence the simplification of the descriptive system is an essential prerequisite. Two separate strategies do this were adopted, based firstly on the analysis of pre-existing responses on the FACT-L, obtained in a clinical study of 431 patients with non-small cell lung cancer and secondly on the results of five qualitative interviews with oncologists experienced in the use of FACT-L were performed. RESULTS: The factor analytic approach yield limited scope for item reduction although a more efficient clustering of items within dimensions was achieved. The qualitative review produced greater scope for simplification of the descriptive system. CONCLUSIONS: An efficient subset of FACT-L items that have clinical descriptive salience can be achieved as a preliminary to the estimation of item/level utility weights.