OA patients' response to analgesic treatment was highly individual and the response to one drug was not predictive of that to another drug. A significant proportion of patients were not treated satisfactorily with diclofenac or tramadol alone. The results obtained from a descriptive analysis of group effects (means, medians) were inappropriate for drawing conclusions on individual treatment benefits. Improvement of functional capability apparently was a consequence of pain relief. Effective pain relief should therefore be the main therapeutic goal in patients with OA where inflammation is less prominent.