PURPOSE: The purpose of this study was to improve the effectiveness of adult hearing screens and demonstrate that interventions assessment methods are needed that address the individual's experienced hearing. Item response theory, which provides a methodology for assessing patient-reported outcomes, is examined here to demonstrate its usefulness in hearing screens and interventions. METHOD: The graded response model is applied to a scale of 11 items assessing perceived hearing functioning and 10 items assessing experienced social limitations completed by a sample of 212 persons age 55+ years. Fixed and variable slope models are compared. Discrimination and threshold parameters are estimated and information functions evaluated. RESULTS: Variable slope models for both scales provided the best fit. The estimated discrimination parameters for all items except for one in each scale were good if not excellent (1.5-3.4). Threshold values varied, demonstrating the complementary and supplementary value of items within a scale. The information provided by each item varies relative to trait values so that each scale of items provides information over a wider range of trait values. CONCLUSION: Item response theory methodology facilitates the comparison of items relative to their discriminative ability and information provided and thus provides a basis for the selection of items for application in a screening setting.
PURPOSE: The purpose of this study was to improve the effectiveness of adult hearing screens and demonstrate that interventions assessment methods are needed that address the individual's experienced hearing. Item response theory, which provides a methodology for assessing patient-reported outcomes, is examined here to demonstrate its usefulness in hearing screens and interventions. METHOD: The graded response model is applied to a scale of 11 items assessing perceived hearing functioning and 10 items assessing experienced social limitations completed by a sample of 212 persons age 55+ years. Fixed and variable slope models are compared. Discrimination and threshold parameters are estimated and information functions evaluated. RESULTS: Variable slope models for both scales provided the best fit. The estimated discrimination parameters for all items except for one in each scale were good if not excellent (1.5-3.4). Threshold values varied, demonstrating the complementary and supplementary value of items within a scale. The information provided by each item varies relative to trait values so that each scale of items provides information over a wider range of trait values. CONCLUSION: Item response theory methodology facilitates the comparison of items relative to their discriminative ability and information provided and thus provides a basis for the selection of items for application in a screening setting.