The concept of a minimal clinically important difference (MCID) is well established. Here, we review the evidence base and methods used to define MCIDs as well as their strengths and limitations. Most MCIDs in chronic obstructive pulmonary disease (COPD) are empirically derived estimates applying to populations of patients. Validated MCIDs are available for many commonly used outcomes in COPD, including lung function (100 ml for trough FEV1), dyspnea (improvement of ≥ 1 unit in the Transition Dyspnea Index total score or 5 units in the University of California, San Diego Shortness of Breath Questionnaire), health status (reduction of 4 units in the St George's Respiratory Questionnaire total score), and exercise capacity (47.5 m for the incremental shuttle walking test, 45-85 s for the endurance shuttle walking test, and 46-105 s for constant-load cycling endurance tests), but there is currently no validated MCID for exacerbations. In a clinical trial setting, many factors, including study duration, withdrawal rate, baseline severity, and Hawthorne effects, can influence the measured treatment effect and determine whether it reaches the MCID. We also address recent challenges presented by clinical trials that compare active treatments and suggest that MCIDs should be used to identify the additional proportion of patients who benefit, for example, when one drug is replaced by another or when a second drug is added to a first. We propose the term "minimum worthwhile incremental advantage" to describe this parameter.