Objective: To estimate the predictive value of the 30-question Geriatric Depression Scale (GDS) in Spanish and calculate the most adequate cut-point for its use in Primary Health Care consultations.
Method: 218 patients over the age of 64 treated at three health centers of Area 10 in Madrid were selected. In the first phase, the subjects completed the GDS, the Mini-Mental State Examination (MMSE) and a questionnaire on health and socio-demographic variables. They were later interviewed using the Geriatric Mental Schedule (GMS), used as the gold standard by doctors who were unaware of the results of the GDS. Two categories were contemplated according to the results of the GMS: cases of depression (diagnosis of psychotic or neurotic depression) and non-psychiatric cases (no psychiatric diagnosis, although isolated symptoms could be present).
Results: 192 aged subjects were interviewed using the GDS and the GMS. Of these, 103 were considered "non-cases of depression" and 60 others made up the "cases of psychotic/neurotic depression" group. For the most effective cut-point (9/10), sensitivity was 86.7% and specificity 63.1%. Considering a prevalence of depression of 30%, the predictive value for positives was 50.2% and for negatives 91.7%. The Cronbach alpha coefficient was 0.82, and the area below the ROC curve obtained was 0.85. Those patients with cognitive deterioration had a mean GDS score similar to those that did not present deterioration (11.16 vs 10.52; p > 0.05).
Conclusions: The Geriatric Depression Scale is valid as a screening test in Primary Care consultations due to its high sensitivity and negative predictive value. The most effective Spanish GDS cut-point (9/10) is lower than that obtained in the original English version (10/11).
Copyright 2002 John Wiley & Sons, Ltd.