Abbreviating the brief. Approach to ultra-short versions of the Yesavage questionnaire for the diagnosis of depression

Abstract

Objectives. To assess the validity of small questionnaires of 4.5 and 1 questions, based on the 15-question Spanish version of the Yesavage geriatric depression scale (GDS-VE), for diagnosing depression in the over-65s. Design. Study to validate questionnaires. Setting. Health centre consultations. Participants. 258 patients were selected to adapt and validate the original questionnaire (GDS-15). Then validity data were assessed by using a group with one question (GDS-1), a group with 4 questions (GDS-4) and a group with 5 (GDS-5). The questions were chosen for the greater sensitivity and specificity shown in the validation study of the GDS-VE. Main measurements. Inter and intra-observer reliability, internal consistency, convergent validity (Spearman's correlation coefficient: CC), criterion validity (sensitivity, S, and specificity, E), and the diagnostic capacity of each questionnaire (area under the ROC curve: AUC), at their respective 95% confidence intervals (CI), were all analysed. Results. 249 patients concluded the study, with the following results (the cut-off point was taken as 2 or more points on the 4 and 5-question versions): GDS-1: S: 71.1% (61.7-80.4), E: 74.1% (67.3-80,9), AUC: 0.726 (0.659-0.793), CC: 0.679 (0.673-0.684). GDS-4: S: 73.3% (64.2-82.4), E: 77.7% (71.2-84.2), AUC: 0.821 (0.765-0.876), CC: 0.803 (0.754-0.852). GDS-5: S: 81.1% (73.1-89.1), E: 73.2% (66.3-80.1), AUC: 0.833 (0.789-0.887), CC: 0.866 (0.824-0.908). Conclusions. The sensitivity and specificity results are similar to results of other ultra-short GDS studies. The high reliability of these ultra-short versions leads us to recommend that new studies be run to enable these versions to be consolidated as a valid and very rapid option for diagnosing depression among the elderly.