The smallest change a patient would actually notice or value.
The Minimal Clinically Important Difference (MCID) is the smallest change in an outcome measure that a patient would recognise as beneficial (or harmful) and that would prompt a clinician to consider changing their management. It is a threshold for meaningfulness, not just for measurement precision, and it is specific to the outcome tool, population and condition it was derived from.
When appraising a study, the MCID lets you separate statistical significance from clinical relevance: a large trial can detect a tiny, real difference between groups that is well below the MCID and therefore irrelevant to practice. Ignoring this distinction risks adopting an intervention on the strength of a p-value alone, when the effect size is too small to matter to any individual patient.
Papers typically report the MCID (or minimal important difference, MID) for the outcome tool used, then compare the observed between-group or within-group change against it, sometimes reporting the proportion of patients who achieved the MCID rather than mean change alone. MCID values are usually derived using anchor-based methods (comparing score change to a patient-rated global improvement anchor), distribution-based methods (e.g. based on standard error of measurement or effect size), or a consensus of both, and should be reported with the population and instrument they were validated in.
MCID values vary by population, baseline severity, condition chronicity and even direction of change (improvement vs deterioration), so applying one derived in a different patient group is often invalid. Distribution-based estimates reflect statistical variability rather than patient-perceived meaning, and studies sometimes cite an MCID from an unrelated context to justify a result, or fail to report one at all, so it is worth checking that the value used matches the outcome measure and population under review.
This guide was auto-drafted and is pending editorial review.