Lena Spangenberg1, Heide Glaesmer2, Maren Boecker3, Thomas Forkmann3. 1. Department of Medical Psychology and Medical Sociology, University of Leipzig, Philipp-Rosenthal-Str. 55, 04103, Leipzig, Germany. Lena.Spangenberg@medizin.uni-leipzig.de. 2. Department of Medical Psychology and Medical Sociology, University of Leipzig, Philipp-Rosenthal-Str. 55, 04103, Leipzig, Germany. 3. Institute of Medical Psychology and Medical Sociology, University Hospital of RWTH Aachen, Aachen, Germany.
Abstract
PURPOSE: To extend knowledge about measurement equivalence of depression measures assessed by tablet and paper-pencil administration, the present study evaluated the effect of mode of administration (MoA) on scale and item level for the Patient Health Questionnaire (PHQ-9) and the Aachen Depression Item Bank (ADIB) in elderly patients. METHODS: Primary care patients (N = 193, ≥60 years) were assessed following a crossover design in Leipzig, Germany. All participants filled out the PHQ-9 and the ADIB in both MoAs under study. Effects of MoA were analyzed by intra-class correlation, mixed-effects regression, and differential item functioning (DIF). Additionally, detection rates between both MoAs were compared using receiver operating characteristics analysis compared to a diagnostic interview (SCID-I, N = 163). RESULTS: No effect of MoA was found in the PHQ-9 on scale score or item level. Two ADIB items showed DIF according to MoA. In terms of discriminatory power, MoA did not influence detection rates of both instruments. CONCLUSIONS: In summary, our findings suggest that no severe effect of mode of administration on self-report assessments of depression should be expected. It can be concluded that tablets provide a valid way to electronically assess depressive symptoms in elderly patients. Yet changes in item presentation can influence the psychometric properties and require equivalence testing using sophisticated analyses on item level such as DIF.
PURPOSE: To extend knowledge about measurement equivalence of depression measures assessed by tablet and paper-pencil administration, the present study evaluated the effect of mode of administration (MoA) on scale and item level for the Patient Health Questionnaire (PHQ-9) and the Aachen Depression Item Bank (ADIB) in elderly patients. METHODS: Primary care patients (N = 193, ≥60 years) were assessed following a crossover design in Leipzig, Germany. All participants filled out the PHQ-9 and the ADIB in both MoAs under study. Effects of MoA were analyzed by intra-class correlation, mixed-effects regression, and differential item functioning (DIF). Additionally, detection rates between both MoAs were compared using receiver operating characteristics analysis compared to a diagnostic interview (SCID-I, N = 163). RESULTS: No effect of MoA was found in the PHQ-9 on scale score or item level. Two ADIB items showed DIF according to MoA. In terms of discriminatory power, MoA did not influence detection rates of both instruments. CONCLUSIONS: In summary, our findings suggest that no severe effect of mode of administration on self-report assessments of depression should be expected. It can be concluded that tablets provide a valid way to electronically assess depressive symptoms in elderly patients. Yet changes in item presentation can influence the psychometric properties and require equivalence testing using sophisticated analyses on item level such as DIF.
Authors: Lizabeth A Goldstein; Mary Beth Connolly Gibbons; Sarah M Thompson; Kelli Scott; Laura Heintz; Patricia Green; Donald Thompson; Paul Crits-Christoph Journal: J Behav Health Serv Res Date: 2011-07 Impact factor: 1.505
Authors: Femke Lamers; Catharina C M Jonkers; Hans Bosma; Brenda W J H Penninx; J André Knottnerus; Jacques Th M van Eijk Journal: J Clin Epidemiol Date: 2008-02-14 Impact factor: 6.437
Authors: Derek D Satre; Alexandra N Anderson; Amy S Leibowitz; Tory Levine-Hall; Sally Slome; Jason Flamm; C Bradley Hare; Jennifer McNeely; Constance M Weisner; Michael A Horberg; Paul Volberding; Michael J Silverberg Journal: Contemp Clin Trials Date: 2019-08-22 Impact factor: 2.226
Authors: Maud Wieczorek; Christine Rotonda; Jonathan Epstein; Francis Guillemin; Anne-Christine Rat Journal: Qual Life Res Date: 2018-06-12 Impact factor: 4.147
Authors: Lily Chen; Julia L Chapman; Brendon J Yee; Keith K H Wong; Ronald R Grunstein; Nathaniel S Marshall; Christopher B Miller Journal: BMJ Open Date: 2018-03-08 Impact factor: 2.692