SCREENING FOR DEPRESSION - CLINICAL VALIDATION OF GERIATRICIANS DIAGNOSIS, THE BRIEF ASSESSMENT SCHEDULE DEPRESSION CARDS AND THE 5-ITEM VERSION OF THE SYMPTOM CHECK LIST AMONG NONDEMENTED GERIATRIC INPATIENTS

In this study, two screening instruments for depression and geriatrici ans' diagnosis were compared against the Geriatric Mental State Schedu le (GMS), a standardized semi-structured psychiatric interview. The Br ief Assessment Schedule Depression Cards (BASDEC) achieved 91% sensiti vity and 85% specificity using a cutoff score of 7. Its receiver opera ting characteristics (ROC) had an area under the curve (AUC) of 0.88, with 95% confidence intervals of 0.78 and 0.98. The 5-item version of the Symptom Check List (SCL-5) achieved 77% sensitivity and 74% specif icity using a cutoff score of 10. Its AUC was 0.77, with 95% confidenc e intervals of 0.63 and 0.90. The p-value of the statistical differenc e between the two AUCs was 0.0554. The geriatricians' diagnosis had a kappa agreement coefficient of 0.39, sensitivity of 55% and specificit y of 96%. They missed 45% of depressed patients. Routine screening wit h BASDEC would considerably improve the detection of depression. Even among those patients who did not appear depressed to the geriatricians , BASDEC would detect one case in every 10 patients.