Psychometric analyses to improve the Dutch ICF Activity Inventory

Janna E Bruijning,Ger van Rens,Dirk Knol,Ruth van Nispen
DOI: https://doi.org/10.1097/OPX.0b013e318282cdce
Abstract:Purpose: In the past, rehabilitation centers for the visually impaired used unstructured or semistructured methods to assess rehabilitation needs of their patients. Recently, an extensive instrument, the Dutch ICF Activity Inventory (D-AI), was developed to systematically investigate rehabilitation needs of visually impaired adults and to evaluate rehabilitation outcomes. The purpose of this study was to investigate the underlying factor structure and other psychometric properties to shorten and improve the D-AI. Methods: The D-AI was administered to 241 visually impaired persons who recently enrolled in a multidisciplinary rehabilitation center. The D-AI uses graded scores to assess the importance and difficulty of 65 rehabilitation goals. For high-priority goals (e.g., daily meal preparation), the difficulty of underlying tasks (e.g., read recipes, cut vegetables) was assessed. To reduce underlying task items (>950), descriptive statistics were investigated and factor analyses were performed for several goals. The internal consistency reliability and test-retest reliability of the D-AI were investigated by calculating Cronbach α and Cohen (weighted) κ. Finally, consensus-based discussions were used to shorten and improve the D-AI. Results: Except for one goal, factor analysis model parameters were at least reasonable. Internal consistency reliability was satisfactory (range, 0.74 to 0.93). In total, 60% of the 65 goal importance items and 84.4% of the goal difficulty items showed moderate to almost perfect κ values (≥0.40). After consensus-based discussions, a new D-AI was produced, containing 48 goals and less than 500 tasks. Conclusions: The analyses were an important step in the validation process of the D-AI and to develop a more feasible assessment tool to investigate rehabilitation needs of visually impaired persons in a systematic way. The D-AI is currently implemented in all Dutch rehabilitation centers serving all visually impaired adults with various rehabilitation needs.
What problem does this paper attempt to address?