Development of an item bank for a computerised adaptive test of upper-extremity function

Purpose.The purpose of this study was to determine the psychometric characteristics of an upper-extremity item bank as a precursor to developing a computer adaptive patient reported outcome instrument. The Activity dimension of the World Health Organization's International Classification of Functioning, Disability and Health (ICF) provided the conceptual framework for the items. Method.Factor and Rasch analyses were used to evaluate the psychometric properties of the item bank, including: monotonicity, local independence, dimensionality, item difficulty hierarchy and match between sample ability and item difficulty. Results.Monotonicity of the rating scale was supported. Nine item pairs were locally dependent, and thus one item from each pair was removed from subsequent analyses. There was evidence for two unidimensional constructs; gross upper-extremity and fine hand. Both constructs showed good internal consistency and person separation. In general, the order of item difficulty within each construct replicated the hypothesised item difficulty order. The fine hand construct had a ceiling effect. Conclusions.The above study of our newly developed upper-extremity item bank empirically verified the intended item difficulty order, identified separate constructs (i.e. gross upper-extremity and fine hand) and provided insights into eliminating the ceiling effect of one of the constructs. These findings are critical precursors to the development of upper-extremity components of the ICF Activity Measure, an ICF-based, CAT located on the web at: www.icfmeasure.phhp.ufl.edu.