OBJECTIVES: To determine if item-level examination of 12-month changes in quality of life (QOL) ratings gives qualitatively different evidence that QOL change occurred as compared to difference in total scores. METHODS: A total of 258 AD patients in a clinical trial of nonsteroidal anti-inflammatory agents responded to 13 QOL items at two visits 12 months apart. A simple difference (later – earlier) was calculated for total scores. A qualified change algorithm was applied to the items: ratings of ‘good' and ‘fair' were treated as indistinguishable and differences per item were classified as one of four possible outcomes: improved, worsened, stayed poor, and stayed ‘OK' (fair, good, excellent). A Toulmin diagram permits evaluation of the strength of evidence supporting a claim that “QOL changed”; this method was used to compare evidence of QOL change from the total score and item levels. RESULTS: “Significant loss of QOL” was observed in the treatment (each group mean ≈ -1.0, p <0.05), but not the placebo (mean =-0.59, p >0.3), groups. Roughly 60% of all change in QOL items was worsening in each arm; 17-42% of all subjects experienced change in each item. CONCLUSIONS: Simple difference in total score over 12 months suggested little loss in QOL whereas qualifying change suggested that at least 17% of any group experienced change on every item with 60% of all change being worsening. A simple difference summarizes estimated change, obscuring evidence that change occurs. The Toulmin diagram suggests that more and stronger evidence of change was obtained in a qualified change approach, which reflects change as both improvement and worsening and contemplates alternative explanations for observed change. Qualified change provides more and stronger evidence of change in subjective measures such as QOL.