The aim of the current study was to develop scales that assess symptoms of depression and anxiety and can adequately differentiate between depression and anxiety disorders, and also can distinguish within anxiety disorders. As point of departure, we used the tripartite model of Clark and Watson that discerns three dimensions: negative affect, positive affect and physiological hyperarousal.

Analyses were performed on the data of 1449 patients, who completed the Mood and Anxiety Symptoms Questionnaire (MASQ) and the Brief Symptom Inventory (BSI). From this, 1434 patients were assessed with a standardized diagnostic interview.

A model with five dimensions was found: depressed mood, lack of positive affect, somatic arousal, phobic fear and hostility. The scales appear capable to differentiate between patients with a mood and with an anxiety disorder. Within the anxiety disorders, somatic arousal was specific for patients with panic disorder. Phobic fear was associated with panic disorder, simple phobia and social anxiety disorder, but not with generalized anxiety disorder.

We present a five-factor model as an extension of the tripartite model. Through the addition of phobic fear, anxiety is better represented than in the tripartite model. The new scales are capable to accurately differentiate between depression and anxiety disorders, as well as between several anxiety disorders.

The Hospital Anxiety and Depression Scale (HADS) has established use with older adult populations in New Zealand but few studies have evaluated its psychometric properties. Research with the psychometric properties of the HADS in elderly populations has primarily used correlational methods that do not allow for the effects of measurement error to be observed. The hypothesized tripartite model of anxiety and depression within the HADS was evaluated using confirmatory factor analysis (CFA) methods.

Overall, 203 community-dwelling older adults who were recruited from older adult community groups completed the HADS. Competing two- and three-factor structures were trialled using CFA.

A three-factor model indicated a lack of differentiation between factors and poor clinical utility and was rejected in favor of a two-factor model. Significant correlations were observed between the anxiety and depression factors on the two-factor model, but it was considered to have validity for older adult samples. Good internal consistency was found for the HADS.

A two-factor model of the HADS was favored due to the lack of differentiation between factors on the three-factor model, and the higher clinical utility of a two-factor solution. The validity of the HADS may be limited by over-diagnosing anxiety in non-clinical populations. It is recommended that the HADS be used to measure change over time through treatment and not be used as a diagnostic tool until future research establishes appropriate norms and cut-offs.