UNLABELLED: Measuring depression among persons with stroke faces many challenges; diagnostic tools are lengthy and do not measure the extent of depression; screening tools are not stroke-specific; and metrics from the available indices do not provide a value that is mathematically or clinically meaningful. PURPOSE: To provide evidence for the development of a stroke specific Geriatric Depression Scale screening measure (SS-GDS) through Rasch methodology. METHODS: Secondary analyses of a randomized controlled trial post-stroke. Interviews from 91 subjects aged 71 (SD 10) over three time points or 240 interviews were analyzed. Rasch Analysis helped transform the 30-item GDS onto a logit scale. Unidimensionality, item fit, redundancy, and differential item functioning (DIF) were assessed. RESULTS: Seventeen items fit the model to form a hierarchical measure ranging in difficulty from +1.2 to -1.8 logits. Preliminary psychometric properties of reliability, validity, and responsiveness were adequate. Two items that demonstrated DIF, one for language and one for gender, were split. CONCLUSION: The 17-item SS-GDS Rasch measure was developed to screen for post-stroke depression (PSD) and provide an important step toward quantifying PSD. If revalidated in a larger sample, the SS-GDS could provide a mathematically valid index to screen for depression in stroke survivors.
RCT Entities:
UNLABELLED: Measuring depression among persons with stroke faces many challenges; diagnostic tools are lengthy and do not measure the extent of depression; screening tools are not stroke-specific; and metrics from the available indices do not provide a value that is mathematically or clinically meaningful. PURPOSE: To provide evidence for the development of a stroke specific Geriatric Depression Scale screening measure (SS-GDS) through Rasch methodology. METHODS: Secondary analyses of a randomized controlled trial post-stroke. Interviews from 91 subjects aged 71 (SD 10) over three time points or 240 interviews were analyzed. Rasch Analysis helped transform the 30-item GDS onto a logit scale. Unidimensionality, item fit, redundancy, and differential item functioning (DIF) were assessed. RESULTS: Seventeen items fit the model to form a hierarchical measure ranging in difficulty from +1.2 to -1.8 logits. Preliminary psychometric properties of reliability, validity, and responsiveness were adequate. Two items that demonstrated DIF, one for language and one for gender, were split. CONCLUSION: The 17-item SS-GDS Rasch measure was developed to screen for post-stroke depression (PSD) and provide an important step toward quantifying PSD. If revalidated in a larger sample, the SS-GDS could provide a mathematically valid index to screen for depression in stroke survivors.