Abstract Introduction Nonsteroidal anti-inflammatory drugs are recommended for the relief of pain associated with hand osteoarthritis (OA) but do not alter the underlying structural changes that contribute to impaired physical function. The current analysis examined the relationship of pain relief with measures of function and global rating of disease in patients with hand OA. Methods This was a combined analysis of 2 prospective, randomized, double-blind, 8-week, multicenter, parallel-group studies comparing diclofenac sodium 1% gel with placebo gel (vehicle) in patients with radiographically confirmed mild to moderate hand OA. Patients (n = 783) aged ≥ 40 years applied diclofenac sodium 1% gel (2 g) or vehicle to each hand 4 times daily for 8 weeks. Outcome measures included pain intensity assessed on a 100-mm Visual Analog Scale (VAS); the Australian/Canadian Osteoarthritis Hand Index (AUSCAN) subscales for pain, stiffness, and physical function (100-mm VAS); and a global rating of disease (100-mm VAS). Change in VAS pain intensity from baseline to week 8 was categorized (<0%, 0%-<15%, 15%-<30%, 30%-<50%, 50%-<70%, and ≥ 70%) without regard to treatment and compared in each category with the mean change from baseline in each AUSCAN subindex and the global rating of disease. Pearson correlations between changes in outcome measures from baseline to week 8 were calculated. Results Changes in VAS pain intensity were accompanied by similar changes in AUSCAN scores and global rating of disease. Pearson correlations confirmed significant associations (P < 0.001) between change in VAS pain intensity and changes in AUSCAN pain (correlation coefficient [r] = 0.81), AUSCAN function (r = 0.75), AUSCAN stiffness (r = 0.66), and global rating of disease (r = 0.76). Conclusions Pain relief correlated with improvements in physical function, stiffness, and global rating of disease in patients with hand OA, irrespective of treatment. This suggests that pain or anticipation of pain inhibits physical function and influences patient perception of disease severity in hand OA. These results also suggest that any intervention to relieve the pain of hand OA may improve function and patient perception of disease severity, despite the absence of a disease-modifying mechanism of action. Trial registration Clinicaltrials.gov NCT00171652, NCT00171665.