BACKGROUND AND OBJECTIVES: There are several preference-based measures (PBMs) used to generate utilities for cost-utility analysis. Psychometric evidence on their performance in different populations exists but a single overview of this evidence is missing. This study aims to provide a comprehensive summary of the overall validity and responsiveness of the EQ-5D, SF-6D, HUI 3, 15D and AQoL, and a critical analysis of the methods employed for assessing PBMs psychometric characteristics. METHODS: A systematic search was undertaken in 4 databases as well as grey literature. Inclusion/exclusion criteria were applied. Data were extracted using a customized template. Quality assessment used a modified version of the AMSTAR checklist. A narrative summary of results is provided based on a three-step approach that generated a strong and weak body of evidence. RESULTS: Thirty reviews were included, 56.7% of which were of good or excellent quality. Most reviews focused on the EQ-5D (n=29), SF-6D (n=12) and HUI 3 (n=8), while only 2 on the 15D and 3 on the AQoL. There was evidence to support EQ-5D validity and responsiveness in respiratory, endocrine and skin diseases, but not in mental health and in diseases of the nervous system. SF-6D was valid in visual disorders and hearing impairment but showed problems in cardiovascular diseases, while HUI 3 was valid in diseases of the nervous system but not in some neoplasms. However, the evidence across populations was limited. Few studies had complete psychometric results and/or comparative evidence between PBMs. Most of the known groups and responsiveness tests were considered weak (e.g. patients vs. general population) and there was poor reporting of psychometric criteria (e.g. validity cut-offs) and results. CONCLUSIONS: Although evidence generally supports PBMs performance, this is often from weak tests and it is rarely comparative, making it difficult to determine with certainty which measure performs better in different disease areas.