Objective: Accurate knowledge is central to effective self-care of osteoarthritis (OA). This study aimed to assess the measurement properties of the Osteoarthritis Knowledge Scale (OAKS) with versions for the hip and knee.

Methods: Participants with hip OA (n = 144), knee OA (n = 327), and no OA (n = 735) were recruited. Rasch analysis was conducted to assess psychometric properties using data from all participants with hip OA and 144 randomly selected participants with either knee OA or no OA. Test-retest reliability and measurement error were estimated among those with hip (n = 51) and knee (n = 142) OA.

Results: Four items from the draft scales were deleted following Rasch analysis. The final 11-item OAKS was unidimensional. Item functioning was not affected by gender, age, educational level, or scale version (hip or knee). Person separation index was 0.75. Test-retest intraclass correlation coefficient was 0.81 (95% CI 0.74, 0.86; hip version 0.66 [0.47, 0.79]; knee version 0.85 (0.79, 0.90)). Smallest detectable change was 9 points (scale range 11–55; hip OA version 11 points; knee OA version 8 points).

Conclusion: The OAKS is a psychometrically adequate, unidimensional measure of important OA knowledge that can be used in populations with and without hip and knee OA. Caution is needed when using with populations with only hip OA as test-retest reliability of the hip version did not surpass the acceptable range.