'But is the difference clinically significant?'

The statement that a difference or change found in a research study is statistically significant is frequently met with the response, 'but is it clinically significant?'. This question seems entirely reasonable and uncontentious until one asks how to determine or define clinical significance. Significance must always have an object; significant to whom? Furthermore it needs elaboration; significant in what way? Once these questions are raised, it becomes apparent that the slick, apparently sensible question is in fact extremely difficult to answer and may have multiple answers: the patient may value being free of discomfort, but the payer may only value achieving less cost in long-term care and the clinician may wish to see change in activities. The use of this question by funding organizations may disadvantage rehabilitation. This Editorial explores the substance of this question, concluding that the question can only be answered by the individual patient concerned, and that research studies should perhaps explore more fully what is actually of significance to patients.