Persson C.U.1, Linder A.2, Hagell P.3
1Institution of Neuroscience and Physiology, Section for Clinical Neuroscience Rehabilitation Medicine Sahlgrenska Academy University of Gothenburg, Gothenburg, Sweden, 2NU-Hospital Group, Physiotherapy Department, Trollhattan, Sweden, 3PRO-CARE Group, School of Health and Society, Kristianstand University, Kristianstad, Sweden
Background: Postural control constitutes one of the prerequisites for everyday activities. After a stroke it is often impaired and also associated with falls. Postural control has demonstrated evidence to be influenced by physiotherapy, and after admission to a stroke unit a physiotherapeutic assessment of postural control is typically routine. For measuring postural control after a stroke, the Swedish modified version of the Postural Assessment Scale for Stroke Patients (SwePASS) is a new12 item rating scale. Since a scale´s outcome is an essential variable on which clinical decisions for further rehabilitation are made, it must be psychometrically sound. Previous research has shown the SwePASS to be responsive to change, reliable and valid in persons after an acute stroke. In persons with moderate deficits due to stroke, an adjusted SwePASS has in a Rasch analysis also shown to be unidimensional and to work the same regardless of age, gender and location of stroke lesion. However, this has not yet been evaluated for persons with severe neurological deficits following stroke.
Purpose: The aim was to evaluate the measurement properties of the SwePASS using Rasch analysis from a sample including persons with different severities of acute stroke.
Methods: Two hundred and fifty persons undergoing acute rehabilitation after stroke in two hospitals in Western Sweden were assessed with the SwePASS by their physiotherapist. A Rasch analysis, based on the assessments, was made using the software RUMM2030© to analyze targeting, reliability, overall fit, individual item and person fit, response category function and differential item functioning.
Results: The SwePASS covered a continuum of different levels of postural control but had suboptimal targeting with insufficient representation of lower and higher levels of postural control. The positive mean person ability value of 2.80 (SD 3.73) showed that the sample as a whole had a higher level of postural control than that represented by the scale. This indicated that the SwePASS was not optimally targeted to the sample. Initially, the SwePASS showed problems with disordered thresholds and local dependency. After creating subtests and rescoring, these problems were resolved. The SwePASS had a high reliability, an acceptable overall fit and provided the same response to an individual item in persons with stroke, regardless of gender, age and location of stroke lesion.
Conclusion(s): The SwePASS exhibited several promising measurement properties. However, poor targeting, illogical response categories and local dependency should be addressed to improve the scale.
Implications: The findings in this study may serve as a basis for further development of the scale which could involve a reduction of items and response categories. The items functioned similarly across all subgroups which is clinically important since there is a broad variation of persons suffering from a stroke.
Funding acknowledgements: The local Research and Development Board for Gothenburg and Södra Bohuslän, the Local Research and Development Board for Fyrbodal, Sweden.
Topic: Outcome measurement
Ethics approval: The study was approved by the Regional Ethical Review Board in Gothenburg (004-14, 25 February 2014).
All authors, affiliations and abstracts have been published as submitted.
