This article evaluates the concept of 'minimal clinically important difference' (MCID) through the results of two studies. The first study was a randomised trial in patients with Parkinson's disease claiming a positive effect of treatment with rasagiline over placebo, based on statistical differences in ordinal outcome measures. However, the clinical relevance of the findings according to the concept of MCID was not taken into consideration. In the second study, MCIDs were defined by several methods in a large trial in patients with chronic inflammatory demyelinating polyneuropathy receiving intravenous immunoglobulin (IGIV) or placebo. In that study, the differences in outcomes between the intervention group and the placebo group were not only statistically significant, but also clinically relevant, in favour of the intervention group. This was demonstrated using various MCID cut-off values.