The best assessment tests for long term trials on osteoarthritis (OA) of the hip and knee are the following, in order of relevance value (consensus of french experts): 1) the loss of joint space thickness on successive radiographies; 2, 3) the indices of the severity for OA of the hip and knee; the investigator's overall opinion; 4) the patient's overall opinion (visual analogue scale of handicap); 5, 6, 7) the pain level (visual analogue scale); the time for going up and down a standard flight of stairs and the time of pain in this distance; the limitation of two articular movements; 8) the increase of either analgesics or NSAIDs consumption; 9) concerning OA of fingers, the number of joints newly involved on successive radiographies. At present, only the radiological tests 1 and 9 are validated. The clinical tests 2 to 8, valuable for short term trials, have yet to be validated for long term follow-up. Recommended duration of trials is three years. A rigorous organisation is necessary to avoid erroneous inclusions: all data recorded in the pre-inclusion visit have to be checked by the principal investigator. Since we have not yet a validated chondroprotective agent as a reference drug, the trial must be randomised, double blind, placebo controlled, parallel group study.