Rasch validation of the German version of the Prosthesis Embodiment Scale for lower limb amputees and proposal of a revised version.

Abstract

To perform a detailed psychometric Rasch analysis of the Prosthesis Embodiment Scale (PEmbS) administered in adults with lower limb amputation (LLA). A convenience sample of German-speaking adults with LLA (<i>n</i> = 150), recruited from German state agencies' databases, was asked to complete the PEmbS, a 10-item patient-reported scale assessing prosthesis embodiment. The local dependency between two items was resolved by keeping for the global score only the lower score of these two items (#9 and #10). Collapsing the seven response categories to four (two expressing disagreement and two agreement) eliminated disordered thresholds. After that, the PEmbS demonstrated unidimensionality, acceptable item fit, and good reliability indices. A keyform plot was created to transform raw scores into linear measures of prosthesis embodiment, making it possible to compare the individual's item responses with those expected by the Rasch model, and to manage missing responses. The PEmbS is useful for assessing prosthesis embodiment in people with LLA, both for research and clinical purposes. We propose a revised version of the PEmbS for lower limb amputees; its appropriateness in other LLA contexts requires further investigation.