Development and initial psychometric evaluation of an item bank created to measure upper extremity function in persons with stroke.

Abstract

OBJECTIVE: To create and illustrate the development of a method to parsimoniously and hierarchically assess upper extremity function in persons after stroke. DESIGN: Data were analyzed using Rasch analysis. SETTING: Re-analysis of data from 8 studies involving persons after stroke. SUBJECTS: Over 4000 patients with stroke who participated in various studies in Montreal and elsewhere in Canada. METHODS: Data comprised 17 tests or indices of upper extremity function and health-related quality of life, for a total of 99 items related to upper extremity function. Tests and indices included, among others, the Box and Block Test, the Nine-Hole Peg Test and the Stroke Impact Scale. Data were collected at various times post-stroke from 3 days to 1 year. RESULTS: Once the data fit the model, a bank of items measuring upper extremity function with persons and items organized hierarchically by difficulty and ability in log units was produced. CONCLUSION: This bank forms the basis for eventual computer adaptive testing. The calibration of the items should be tested further psychometrically, as should the interpretation of the metric arising from using the item calibration to measure the upper extremity of individuals.