Is a statistical significant benefit
for a cohort always clinically
significant for a single patient ?