Objective: The Group for the Respect of Ethics and Excellence in Science (GREES) organized a working group to assess the value of time to joint surgery as a potential therapeutic failure outcome criterion for osteoarthritis (OA) of the hip or knee in the assessment of potential structure modifying agents. Methods: PubMed was searched for manuscripts from 1976 to 2004. Relevant studies were discussed at a 1-day meeting. Results: There are no accepted guidelines for 'time to' and 'indications for' joint replacement surgery. A limited number of trials have examined joint replacement surgery within the study population. Several parameters, particularly joint space narrowing (interbone distance), correlate with surgical intervention. However, at the level of the knee, none of the parameters have positive predictive value for joint replacement surgery better than 30%. In contrast, lack of significant joint space narrowing has a strong negative predictive value for joint replacement surgery (>90%), that remains after controlling for OA pain severity. Conclusion: At this time, GREES cannot recommend time to joint surgery as a primary endpoint of failure for structure modifying trials of hip or knee OA-as the parameter has sensitivity but lacks specificity. In contrast, in existing trials, a lack of progression of joint space narrowing has predictive value of >90% for not having surgery. GREES suggests utilizing joint space narrowing (e.g. >0.3-0.7 mm) combined with a lack of clinically relevant improvement in symptoms (e.g. ≥ 20-25%) for 'failure' of a secondary outcome in structure modifying trials of the hip and knee. © 2004 Osteo Arthritis Reserch Society International. Published by Elsevier Ltd. All rights reserved. ; SCOPUS: ar.j ; info:eu-repo/semantics/published