- Title
- Apples to apples? Comparison of the measurement properties of hospital anxiety and depression-anxiety subscale (HADS-A)...
- Creator
- Clover, Kerrie; Lambert, Sylvie D.; Oldmeadow, Christopher; Britton, Benjamin; King, Madeleine T.; Mitchell, Alex J.; Carter, Gregory L.
- Relation
- Current Psychology Vol. 41, Issue 7, p. 4592-4601
- Publisher Link
- http://dx.doi.org/10.1007/s12144-020-00906-x
- Publisher
- Springer
- Resource Type
- journal article
- Date
- 2022
- Description
- A range of anxiety measures is used in oncology but their comparability is unknown. We examined variations in measurement across three commonly used instruments: Hospital Anxiety and Depression-Anxiety subscale (HADS-A); Depression, Anxiety, Stress Scale - Anxiety subscale (DASS-A); and Generalised Anxiety Disorder scale (GAD-7). Participants (n = 164) completed the self-report measures and the Generalised Anxiety Disorder module of the Structured Clinical Interview for DSM-IV (SCID). We performed Rasch analysis and calculated diagnostic accuracy statistics. Instruments measured similar constructs of anxiety, but had different ranges of measurement, with the HADS-A including lower severity symptoms than the other two measures. Anxiety severity was similar for GAD-7 ‘mild’ and HADS-A ‘possible’ categories, but ‘mild’ anxiety on the DASS-A represented more severe symptoms. Conversely, DASS-A ‘severe’ anxiety represented less intense symptoms than GAD-7 ‘severe’ anxiety. Co-calibration indicated a score of eight on the HADS-A was equivalent in anxiety severity to scores of three on the DASS-A and six on the GAD-7. Area under the curve (AUC) was just acceptable for HADS-A and GAD-7 but not DASS-A. The HADS-A, DASS-A and GAD-7 displayed important differences in how they measured anxiety. In particular, categorical classifications of anxiety severity (mild/moderate/severe) were not equivalent across instruments. Thus, prevalence estimates of anxiety symptoms will vary as a consequence of the instrument used. The GAD-7 and HADS-A obtained more similar results and better AUC than the DASS-A. Our co-calibration could be used in future studies and meta-analyses of individual participant data to set cut-off points that provide more consistent classification of anxiety severity.
- Subject
- anxiety; cancer; Rasch-analysis; psychometrics; validation; psycho-oncology
- Identifier
- http://hdl.handle.net/1959.13/1452639
- Identifier
- uon:44472
- Identifier
- ISSN:1046-1310
- Language
- eng
- Reviewed
- Hits: 2177
- Visitors: 2173
- Downloads: 0