Statement of Purpose: To determine a short measure of physical function using existing KOOS, HOOS and WOMAC data from people with knee or hip osteoarthritis (OA).
Methods: Rasch analysis was conducted on existing data from individuals with OA treated medically or surgically (post meniscectomy, osteotomy) as well as those immediately pre total hip or knee replacement. Fit of the data to the Rasch model was evaluated by model chi-square and item fit statistics (chi-square, size of residual, F-test). Differential item functioning was evaluated by gender, age and country. Unidimensionality was evaluated by factor analysis and residual analysis. Individual data sets were analyzed and data pooled and re-analyzed for fit to the Rasch model. Once the final set of items was determined by the above procedures, correlations between physical function scores based on the Rasch scoring and a simple sum score were calculated to determine how well the summed score represented the data.
Results: Nine sets of knee data (n=1662) were included with ages 26-101 years. The male/female ratio was 964:698. Of the 17 physical and 5 sports items of the KOOS, 11 items fit the Rasch model, are free from differential item functioning and represent a unidimensional, interval level scale. Table 1 shows the 11 items. The Rasch scores and the sum of the 11 items were correlated (R=0.83). Nine sets of hip data (n=5266) were included with ages 28-102 years. The male/female ratio was 2002:3152. Based on the above criteria, 6 HOOS items fit the model (Table 2). The Rasch scores and the raw sum were correlated (R=0.99).
Conclusion: The original activity and sports items of the KOOS and HOOS (which subsume the WOMAC items) were reduced to 11 and 6 items, respectively, achieving a feasible, short scale with interval level properties that can be used as a function component of an OA severity scoring system.
Table1: Physical scale for knee OA
Table 2: Physical scale for hip OA
Methods: Rasch analysis was conducted on existing data from individuals with OA treated medically or surgically (post meniscectomy, osteotomy) as well as those immediately pre total hip or knee replacement. Fit of the data to the Rasch model was evaluated by model chi-square and item fit statistics (chi-square, size of residual, F-test). Differential item functioning was evaluated by gender, age and country. Unidimensionality was evaluated by factor analysis and residual analysis. Individual data sets were analyzed and data pooled and re-analyzed for fit to the Rasch model. Once the final set of items was determined by the above procedures, correlations between physical function scores based on the Rasch scoring and a simple sum score were calculated to determine how well the summed score represented the data.
Results: Nine sets of knee data (n=1662) were included with ages 26-101 years. The male/female ratio was 964:698. Of the 17 physical and 5 sports items of the KOOS, 11 items fit the Rasch model, are free from differential item functioning and represent a unidimensional, interval level scale. Table 1 shows the 11 items. The Rasch scores and the sum of the 11 items were correlated (R=0.83). Nine sets of hip data (n=5266) were included with ages 28-102 years. The male/female ratio was 2002:3152. Based on the above criteria, 6 HOOS items fit the model (Table 2). The Rasch scores and the raw sum were correlated (R=0.99).
Conclusion: The original activity and sports items of the KOOS and HOOS (which subsume the WOMAC items) were reduced to 11 and 6 items, respectively, achieving a feasible, short scale with interval level properties that can be used as a function component of an OA severity scoring system.
Table1: Physical scale for knee OA
| Item | Logit (difficulty) |
| taking off socks (easiest) | 1.422 |
| getting on/off toilet | 1.380 |
| walking | 1.325 |
| rises from bed | 1.311 |
| standing | 1.086 |
| rises from sitting | 0.448 |
| twisting | -1.189 |
| jumping | -1.398 |
| squatting | -1.430 |
| running | -1.454 |
| kneeling (most difficult) | -1.504 |
Table 2: Physical scale for hip OA
| Item | Logit (difficulty) |
| sitting (easiest) | 1.887 |
| descending stairs | 0.835 |
| getting in and out of bath | 0.546 |
| twisting/pivoting on loaded leg | -0.015 |
| walking on uneven surface | -0.678 |
| running (most difficulty) | -2.575 |
A.M. Davis, None; A.V. Perruccio, None; M. Canizares, None; G.A. Hawker, None; P.G. Conaghan, None; A. Tennant, None; M. Dougados, None; J. Jordan, None; E. Roos, None; S. Lohmander, None.
See more of: Epidemiology and Health Services Research III
See more of: ACR Abstract Submission Poster Sessions
See more of: ACR Abstract Submission Poster Sessions
![[ Visit Client Website ]](images/banner.gif)