OBJECTIVES: Psychometric analyses to derive a health state classification system in multiple myeloma using the European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire (EORTC QLQ-30) were used to estimate a preference-based measure for health state utility derivation (Rowen et al. 2011). This study sought to apply a modified Rowen framework to identify EORTC QLQ-30 and the breast module (QLQ-BR23) items amenable to valuation to estimate a preference-based measure in metastatic breast cancer (MBC). METHODS: Exploratory factor analysis, Rasch analysis and other psychometric analyses were undertaken on the 53 item EORTC tools using baseline Phase 3 clinical trial data on 1063 patients with MBC to identify items amenable to valuation. Criteria for item identification were: factor loading > 0.4; < 50% of respondents at ceiling/floor; mean square fit statistics between 0.5 and 1.5; and ordered category response. RESULTS: Nine dimensions were identified on the EORTC dataset via exploratory factor analysis. Potential items, stratified by dimension, deemed useful for valuation include: physical (QLQ-30 items 1, 2, 6, 7, 27 and QLQ-BR23 item 6), cognitive/emotional (QLQ-30 items 21, 22, 23 and 24), self-image (QLQ-BR23 items 9, 10, 11 and 12), breast symptoms (QLQ-BR23 item 20), fatigue/sleep disturbance (QLQ-30 items 10, 12 and 18), arm pain (QLQ-BR23 item 17), and general pain (QLQ-30 item 9).  Items related to gastrointestinal and sexual symptoms were not considered due to the measurement criteria. CONCLUSIONS: Twenty of 53 items suitable for valuation were identified over seven dimensions of the EORTC QLQ-30 and QLQ-BR 23. Items that did not meet the specified criteria for inclusion may be utilized for health state derivation if they represent clinically significant components of MBC. Results of the created health state classification system for valuation and derivation of a preference-based measure for metastatic breast cancer will be presented.