Measurement properties of a modified measure of participation for persons with spinal cord injury

Objective: The primary aim of this study was to examine and refine a modified measure of participation for adults with spinal cord injury (SCI) based on a conceptual model of participation. Method: This study involved secondary analysis of data from a larger study designed to identify a standard measure of participation for use in SCI research. The larger study recruited 634 community-dwelling adults with SCI from seven collaborating SCI Model Systems Centers, of whom 520 subjects (average age 45.1 ± 13.6 years, 76% were male) completed the survey that is the focus of the present analysis. Content review, confirmatory factor analysis (CFA), Rasch analysis, and precision analysis were employed to select the items for the modified participation measure. Results: Three participation domains were supported: Productivity, Social, and Community, that displayed good model-fit (CFI=0.984, TLI=0.982, RMSEA=0.043) in CFA and good item-fit (infit= 0.6 to 1.4) in Rasch analysis. Differential Item Functioning (DIF) was found in one item, however its magnitude was small. The precision of each scale was better for participants in the middle range of participation and was lower for participants with extremely low or high participation. Conclusion: The study results support the proposed three-dimensional construct of participation by demonstrating good model-fit and item-fit. Ongoing efforts are needed to expand the domain coverage and increase the precision of the instrument.