• Joshua Stott, Martin Orrell, and Georgina Charlesworth.
• Research Department of Clinical, Educational and Health Psychology, University College, London, United Kingdom. Electronic address: j.stott@ucl.ac.uk.
• Am J Geriatr Psychiatry. 2017 Nov 1; 25 (11): 1270-1280.

ObjectiveThe Hospital Anxiety and Depression Scale (HADS) is a self-report measure of anxiety and depression. It is recommended for clinical assessment and has been used as the primary outcome in large clinical trials with carers of people with dementia. Its validity and utility have never been examined in this population. The current study addresses this.MethodsIn a secondary analysis of baseline data from a recent intervention trial (N = 284) with cross-validation in baseline data from a second trial (N = 230), the authors used confirmatory factor analysis to test whether a one-, two-, or three-factor structure best fit the data and used indices of model misspecification to respecify. Internal consistency, concurrent validity of obtained factors, and measurement invariance across gender, age, kinship, and cohabitation status were assessed.ResultsA three-factor structure best fit the data. Removal of one item improved model fit. The factors showed good internal consistency and high levels of concurrent validity. Measurement invariance was adequate across gender and kinship but not age or cohabitation status. Results were replicated in the cross-validation sample, enhancing reliability.ConclusionIn this group the HADS measured three factors; depression, anxiety, and negative affectivity. The depression scale can be used as originally intended, supporting results of large clinical trials. The HADS does not validly measure distress or anxiety. Consequently, clinical practice recommendations could be revisited, and future research trials should not use HADS anxiety or distress as outcomes. Researchers should pay attention to measurement invariance when using HADS to compare carer subgroups.Copyright © 2017 American Association for Geriatric Psychiatry. Published by Elsevier Inc. All rights reserved.

29 keywords...

hide…