OBJECTIVE: The Hospital Anxiety and Depression Scale (HADS) is commonly used to assess distress among individuals with cancer. However, previous studies cast doubt on the most appropriate dimensional structure for the HADS, suggesting that 1- or 3-dimensional structures might offer superior fit to the original 2-dimensional one. This article is the first to use Rasch analysis to examine the psychometric properties of the subscales corresponding to each of these alternative structures. METHOD: The HADS was completed by 1,360 cancer survivors. Rasch analyses were conducted to examine summary and individual model fit statistics, person separation index, response format, item bias, redundancy, and dimensionality. RESULTS: The HADS-Total scale was found to be multidimensional, and it was necessary to remove almost half of the items to achieve fit. Analyses only partially supported the original structure, as both HADS-Anxiety and HADS-Depression showed initial model misfit and item deletion was necessary to achieve fit. Within the 3-dimensional structures, the Rasch statistics for the anxiety subscales were within acceptable range and no adjustment was needed. Analyses did not support adding Item 7 to HADS-Depression. CONCLUSIONS: Results supported modified versions of the HADS-Anxiety and HADS-Depression; however, combining all items to form HADS-Total is not recommended. Numerous studies using classical test theory and Rasch analyses have corroborated the exclusion of some items (e.g., Item 7) and appropriateness of the subscales defined by a 3-dimensional structure. Further research is required to identify the incremental validity of potential revised subscales. PsycINFO Database Record (c) 2013 APA, all rights reserved.
OBJECTIVE: The Hospital Anxiety and Depression Scale (HADS) is commonly used to assess distress among individuals with cancer. However, previous studies cast doubt on the most appropriate dimensional structure for the HADS, suggesting that 1- or 3-dimensional structures might offer superior fit to the original 2-dimensional one. This article is the first to use Rasch analysis to examine the psychometric properties of the subscales corresponding to each of these alternative structures. METHOD: The HADS was completed by 1,360 cancer survivors. Rasch analyses were conducted to examine summary and individual model fit statistics, person separation index, response format, item bias, redundancy, and dimensionality. RESULTS: The HADS-Total scale was found to be multidimensional, and it was necessary to remove almost half of the items to achieve fit. Analyses only partially supported the original structure, as both HADS-Anxiety and HADS-Depression showed initial model misfit and item deletion was necessary to achieve fit. Within the 3-dimensional structures, the Rasch statistics for the anxiety subscales were within acceptable range and no adjustment was needed. Analyses did not support adding Item 7 to HADS-Depression. CONCLUSIONS: Results supported modified versions of the HADS-Anxiety and HADS-Depression; however, combining all items to form HADS-Total is not recommended. Numerous studies using classical test theory and Rasch analyses have corroborated the exclusion of some items (e.g., Item 7) and appropriateness of the subscales defined by a 3-dimensional structure. Further research is required to identify the incremental validity of potential revised subscales. PsycINFO Database Record (c) 2013 APA, all rights reserved.
Authors: Sylvie D Lambert; Julie F Pallant; Kerrie Clover; Benjamin Britton; Madeleine T King; Gregory Carter Journal: Qual Life Res Date: 2014-04-01 Impact factor: 4.147
Authors: Sylvie D Lambert; Afaf Girgis; Patrick McElduff; Jane Turner; Janelle V Levesque; Karen Kayser; Cathrine Mihalopoulos; Sophy T F Shih; Daniel Barker Journal: BMJ Open Date: 2013-07-24 Impact factor: 2.692
Authors: Sylvie D Lambert; Lindsay R Duncan; Janet Ellis; John Wellesley Robinson; Carly Sears; Nicole Culos-Reed; Andrew Matthew; Manon De Raad; Jamie Lynn Schaffler; Daniel Santa Mina; Paramita Saha-Chaudhuri; Helen McTaggart-Cowan; Stuart Peacock Journal: Pilot Feasibility Stud Date: 2021-03-20