Health-related quality of life (HrQOL) assessments are gaining importance as outcome measures in cancer clinical trials. A recently published clinical trial reported statistically significant (P<0.001) increases in haemoglobin (Hb) levels and significantly (P<0.01) increased HrQOL scores following the administration of recombinant human erythropoietin (r-HuEPO, epoetin alfa) versus placebo to anaemic cancer patients who received non-platinum chemotherapy. This study employed five cancer-specific HrQOL instruments. Hb and HrQOL data from this trial were analysed to estimate the minimally important difference (MID) in HrQOL measures that could be interpreted as clinically meaningful, with Hb level selected as the best external standard. Patients were assigned to two groups: improved (Hb increases of >/=1 g/dL) or stable (change in Hb of-1 g/dL to <1 g/dL). The MID was first determined as the difference between the mean changes in HrQOL in the improved group versus the stable group. By this analysis, the differences in HrQOL scores between the epoetin alfa group and the placebo group were clinically important for all Hb-sensitive, cancer-specific HrQOL evaluations. Linear regression analyses performed to provide estimates of the MID for specific values of Hb change confirmed that the differences in HrQOL scores between patient groups were clinically significant. These analyses were repeated using a data set from a separate clinical trial, which further supported the conclusion that observed HrQOL changes demonstrated in the multicentre, double-blind study were clinically important. These methods provide one means for interpreting the clinical relevance of changes in HrQOL evaluated in clinical trials.