Statistical significance is often defined by comparing the p-value to a pre-set significance level. Assessment of clinical significance requires reliable estimates of the benefit and risk to be weighed. Clearly in order for the benefit to outweigh the risk, the benefit must be unquestionably established. We will explore through an example the relationship between the observed treatment effect size and the associated p-value, and discuss whether or not statistical significance itself may be sufficient to demonstrate efficacy even if the observed treatment effect size is small.