HRQoL measures are becoming important for evaluating the effects of arthritis treatments. Interpreting changes in HRQoL scores, however, has not been fully evaluated. OBJECTIVE: To determine the clincally-meaningful improvement of HRQOL as measured by the SF-36 among OA patients. METHODS: Data were obtained from three 12-week randomized clinical trials among 3,369 OA flare patients. The SF-36 and Patient Global Assessment (PtGA) were administered to patients at baseline and week-12. PtGA was measured on the following 1 to 5-point scale: very good, good, fair, poor, or very poor. Change scores were computed by subtracting patient’s baseline from week-12 follow-up scores. Clinically-meaningful improvement was estimated as the mean change score corresponding to a one-level improvement in PtGA. A similar interpretation of clinically-meaningful changes was also performed using physician-global assessment, pain, and functional-status. RESULTS: At week-12, the following patients experienced improvements in their PtGA rating: one-level (1,158), two-levels (745), three or more levels (311), which corresponded to standardized physical and mental component scores (PCS & MCS) improvements of: one-level (3.8 & 1.4), two-levels (8.3 & 2.7), and three or more levels (10.4 & 4.2), respectively. The clinically-meaningful improvement based on PtGA among OA patients were approximately 3.8 and 1.4 for PCS and MCS, respectively. The clinically-meaningful changes for eight SF-36 domains were 7.2 for physical function, 13.1 for Role physical, 11.6 for bodily pain, 2.1 for general health, 5.6 for vitality, 6.2 for social function, 7.7 for role emotion, 2.7 for mental health. CONCLUSION: Clinically-meaningful improvement varies between SF-36 domains and summary scores among OA patients. These results provide guidance in interpreting HRQoL results and planning clinical trials.