Objective: To study the validity of the American College of Rheumatology (ACR) criteria for hip osteoarthritis (OA) in patients receiving primary care. Three different sets of criteria are available, one with clinical symptoms only, and 2 with clinical symptoms and radiological signs combined. It is claimed that all 3 sets, separate from each other, can be used to diagnose hip OA for research purposes.
Methods: Consecutive patients (n = 227) aged 50 years or older, who had consulted a general practitioner for pain in the hip and had been referred for radiographic investigation of the hip, were recruited for a standardized history taking and physical investigation. The radiographs were evaluated according to a standardized protocol. The cross validity of the 3 different sets of ACR criteria was assessed by calculating the percentage agreement and the kappa between the separate sets.
Results: There was poor agreement between the set of clinical criteria and the 2 sets in which radiological signs were included (kappa 0.11 or lower). The 2 sets that included radiological signs showed high agreement (kappa 0.81-0.94, depending on cut off points for joint space narrowing). These sets also showed the highest agreement with the radiological OA defined as a Kellgren score 2 or more (kappa 0.13-0.48).
Conclusion: The clinical ACR criteria showed no cross validity with the 2 other ACR criteria sets.