Poster 162: Depression assessment in rehabilitation patients with communication difficulties: the Cornell depression scale.1

Abstract Objective: To use the Cornell Depression Scale (CDS) to diagnose major depression in rehabilitation inpatients with communication difficulties, using the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV), as standard assessment tool. Design: Prospective clinical trial. Setting: Rehabilitation medicine unit in a tertiary teaching hospital. Participants: Consecutive general rehabilitation patients (61.8% strokes, others being head injuries, spinal cord injuries, deconditioning) admitted between May and August 2002. Patients who did not complete their rehabilitation program for various reasons (including medical instability) were excluded from analysis. Interventions: All 78 patients admitted were screened by the same investigator for depressive signs and symptoms using DSM-IV criteria on admission, and then on a weekly basis until discharge. Of the 68 patients who completed rehabilitation, 14 (20.6%) had communication difficulties (defined as a score of Main Outcome Measure: Presence or absence of major depression, as determined by the CDS and/or DSM-IV criteria. Results: 13 (19.1%) of the 68 patients were diagnosed with major depression using the DSM IV. Of these, 4 patients had communication difficulty; they also fulfilled the criteria for major depression on the CDS. Importantly, the remaining 10 patients with communication difficulty did not have depression as assessed by the DSM-IV and the CDS. There was hence a 100% correlation between the 2 tools. Conclusions: Diagnosis of major depression in the presence of communication difficulties can be challenging. The CDS, a nonverbal assessment tool focusing on observers’ ratings and originally constructed for patients with dementia, showed accurate correlation with the well-established DSM-IV and may hence be useful in rehabilitation patients with communication difficulties. A larger study, to ensure accuracy and statistical significance of this correlation, is necessary.