• Thelma J Mielenz, Michael C Edwards, and Leigh F Callahan.
• Thurston Arthritis Research Center, University of North Carolina, Chapel Hill, NC 27599-7280, USA. thelma_mielenz@unc.edu
• J Clin Epidemiol. 2010 Mar 1;63(3):315-20.

ObjectivesTo conduct the initial modern measurement theory analyses because of its many advantages on the Tampa Scale for Kinesiophobia and emerging evidence suggesting that fear of movement influences functional disability in people with arthritis.Study Design And SettingSecondary analysis of 347 participants from a randomized controlled trial evaluating The People with Arthritis Can Exercise program. The original Tampa Scale for Kinesiophobia has 17 items and we collected 16 items (excluding item 6). An item response theory analysis was conducted using the graded response model in MULTILOG. Before this, a series of factor analyses assessed the unidimensionality assumption of this model.ResultsBased on the factor analyses, we removed the reverse-coded items (4, 8, 12, and 16). The item response theory analysis revealed that item 13 had an exceedingly low slope and was dropped.ConclusionItem response theory analyses looked at each item's performance and we can strongly suggest using our modified scale (11 items out of the 16 items), which provides relatively uniform precision of measurement across a wide range of fear of movement in people with arthritis. The item parameters from this study can build a computerized adaptive testing for this scale.

21 keywords...

hide…