ObjectiveEstimates of minimal clinically important improvements (MCIIs) are larger among patients with higher values at baseline, suggesting that these patients require larger changes to appreciate improvements. We examined if baseline dependency of MCIIs was associated with specific patients across three measures, or was owing to floor and ceiling effects. Study Design and SettingWe prospectively examined 250 outpatients with active rheumatoid arthritis (RA). We used an anchor-based approach to estimate MCIIs for three measures of RA activity (patient global assessment, swollen joint count, and walking time). We examined if the same patients constituted the baseline subgroups with high MCIIs across measures. ResultsThe MCIIs were greater for those with higher baseline values of all three measures. At the ceiling, there was little opportunity to improve, and judgments were unrelated to measured changes. At midrange, improvements were balanced by worsenings, including some judged as improvements. At the floor, improvements were not similarly balanced. Patients in subgroups with high MCII for patient global assessment were not also predominantly in subgroups with high MCII for the swollen joint count or walking time, and vice versa. ConclusionVariation in MCII by baseline values is because of floor and ceiling effects rather than expectations of particular patients.