OBJECTIVES: The EQ-5D is often employed in clinical trials to derive quality-adjusted life years for cost-utility analyses, and in comparisons of health-related quality of life across conditions. However, there are concerns that the EQ-5D is less responsive to change particularly in oncology. Therefore the objective of this study was to determine the level of responsiveness of the EQ-5D in oncology. METHODS: A systematic review identified relevant articles reviewing responsiveness of the EQ-5D in adults (EMBASE, Medline). Effects sizes (ES) were calculated for the studies identified where not already reported. A meta-analysis was undertaken of the effect sizes: homogeneity of variance was assessed (fixed effects) and random effects models applied where there was significant heterogeneity. Responsiveness was also compared for improvement/deterioration in health status. Analyses were conducted in SPSS v18. RESULTS: Data were available from 12 studies (3 breast, 2 prostate, as well as ovarian, lung and renal cancers) each with EQ-5D data at a minimum of 2 time points leading to a total of 45 entries. The overall unweighted ES was -0.26 (95%CI: -0.31 to -0.21), however there was significant heterogeneity in terms of effect sizes (Q(44) = 427.00, p<0.001) which was accounted for using the  random effects model (Q(44) = 39.58, p>0.05). The overall weighted effect size (ES) was -0.17 (95%CI: -0.33 to -0.01). The weighted ES for improvement was 0.08 (95%CI: -0.02 to 0.18), and -0.52 (95%CI: -0.64 to -0.41) for deterioration. CONCLUSIONS: There is considerable heterogeneity in the reported effect size of the EQ-5D. Responsiveness of the EQ-5D in oncology trials as measured by effect sizes is modest at best. The instrument appears to be more sensitive to deterioration in health status than to improvements. Further work will explore the ES of the EQ-5d in comparison with responsiveness of disease-specific measures and changes in health status.