Depression in Hospitalised Aphasic Stroke Patients-Identifying Valid Tools and Examining Outcomes

Objectives: This study aimed to assess the validity of aphasic depression and anxiety rating scales (Signs of Depression Scale [SODS], Hospital version of Stroke Aphasic Depression Questionnaire [SADQH-10], Behavioural Outcomes of Ansxiety scale [BOA]) in the local setting and describe clinical outcomes of a local aphasic stroke population. Methods: Records of 236 aphasic stroke patients from an existing database of a proactive post-stroke depression screening service in a tertiary stroke centre in Singapore with rehabilitation facilities met review criteria for having available data for the SODS, SADQH-10 and carer-rated version of BOA (BOA-C). The reference standard was a psychiatrist's diagnosis of post-stroke mood disorder. Clinical outcomes at 1-year post-stroke were analysed. Results: The areas under the Receiver Operator Characteristic curve (AUC) compared against the psychiatrist's diagnosis of post-stroke mood disorder were 0.805, 0.844 and 0.780 for the SODS, SADQH-10 and BOA-C respectively. Optimal cut-off scores were found for the SODS (four and above) and SADQH (seven and above), with a sensitivity of 60.00% and 55.56%, respectively, and a specificity of 94.62% and 96.17%, respectively. No appropriate cut-off score was found for the BOA-C. Of the patients, 58.6% were found to have poor outcomes at 1-year post-stroke, and recurrent and haemorrhagic strokes were significant predictors of poor outcomes. Conclusions: The SODS and SADQH-10 are appropriate tools for assessing post-stroke depression in hospitalised aphasic stroke patients in Singapore based on their predictive values and likelihood ratios. The BOA-C has poor validity. At 1-year post-stroke, more than 50% of aphasic stroke patients have poor clinical outcomes not associated with clinician-diagnosed depression status, but predicted by recurrent and haemorrhagic strokes.