Abstract

Purpose. To examine the construct validity of the International Classification of Functioning Disability and Health (ICF) framework using multidimensional item response modelling and data collected in different regions from patients with five chronic health conditions. We assume that the ICF components should represent statistically called dimensions that are distinct although related. Method. Retrospective validation study using the ICF Core Sets from a convenience sample of patients in an international multicentre, cross-sectional database obtained in different rehabilitation centres. Health professionals working in 89 rehabilitation centres in 32 countries collected data from 3227 rehabilitation patients using the respective ICF Core Sets. Patients included had one of the following health conditions: low back pain (LBP), rheumatoid arthritis (RA), osteoarthritis (OA), obesity (OB) or post-stroke. Data from questions regarding a patient's functioning based on body structures and functions, activities, participation along the ICF Core Sets were analysed with multidimensional item response modelling. Results. The multidimensional models fit the data better than a model with few or no specifications regarding an underlying framework. For example, a model separating four dimensions 'body structures', 'body functions', 'activities' and 'participation' fits the data better than a model differentiating between 'body functions and structures' and 'activities and participation'. The ICF framework with its components represents underlying statistically called dimensions. Conclusion. The results of this study support the construct validity of the functioning part of the ICF. The distinct dimensions may facilitate the alignment of ICF components with other measures used clinically and in research. Based on our results it is justifiable to construct instruments integrating ICF categories within components.