Validation of screening measures for assessing mood in stroke patients

There are few validated measures to assess mood in stroke patients and even fewer suitable for all stroke patients, including those with communication problems. The aim of this study was to compare the Stroke Aphasic Depression Questionnaire Hospital version (SADQ-H), Signs of Depression Scale (SODS), Visual Analogue Mood Scale (VAMS) and Visual Analogue Self-esteem Scale (VASES) in screening for mood problems after stroke.

Fifty healthy older adults and 100 stroke patients in hospital completed the VAMS and VASES. A nurse completed the SADQ-H and SODS in relation to the stroke patients. A relative/carer completed the SADQ-H and SODS in relation to the healthy older adults. Those without communication problems also completed the Hospital Anxiety and Depression Scale (HADS).

The internal consistency of the scales was low in healthy older adults. In stroke patients the internal consistency of the SADQ-H, VAMS, and VASES was high (alpha = .71-.84) but that of the SODS was low (alpha = .53). In healthy older adults, correlations between the HADS and the VAMS and VASES were high but low between the HADS and SADQ-H and SODS. In stroke patients, the HADS depression scale correlated significantly with all the scales (0.35-0.55) but only the SADQ-H 10, VAMS, and VASES were significantly correlated with the HADS anxiety scale (0.40-0.52). Appropriate cut-offs were found for the SADQ-H (17/18), SADQ-H 10 (5/6), SODS (1/2), and VAMS 'sad' item (22/23) in comparison to depression on the HADS. No appropriate cut-offs were identified in comparison to anxiety on the HADS.

The SADQ-H, SADQ-H10 and SODS were all appropriate for screening for possible depression after stroke but not for screening for possible anxiety. The SADQ-H 10 had greater internal consistency and higher sensitivity and specificity than the SODS and is shorter than the SADQ-H. It was also significantly correlated with both the anxiety and depression scales of the HADS. The SADQ-H 10 was therefore recommended as the most appropriate for screening purposes. The VAMS and VASES provided no clear cut-offs for use in screening but scores were highly correlated with the HADS. They are therefore more suitable for assessing severity of low mood rather than for screening purposes. The cut-offs identified need further validation in an independent sample of stroke patients, including a higher proportion with low mood.