Computerized adaptive testing for PRWHE measurements using domains of pain and motor function.

Abstract

The Patient-Rated Wrist and Hand Evaluation is an outcome measure for patients with conditions affecting the wrist or hand. We evaluated the structural validity of the Patient-Rated Wrist and Hand Evaluation using psychometric techniques, then developed computerized adaptive testing algorithms. Factor analysis found two health constructs consistent with 'Pain' and 'Motor Function'. Two computerized adaptive tests were developed, which reduced the number of items from five to a median of one ('Pain'), and from 10 to two ('Motor Function'). Both computerized adaptive tests achieved high levels of precision (standard error of measurement <0.3) and similar scores to the full-length measure, demonstrated by Bland-Altman analysis. Dividing the Patient-Rated Wrist and Hand Evaluation into two subscales could reduce the response burden, improve standardization of outcome measurement for clinicians and provide precise insight into patient symptoms.Level of evidence: N/A.