Validade da “escala de depressão geriátrica” em unidades primárias de saúde na cidade de Fortaleza, Ceará

Abstract

Introduction – Depression is one of the major potentially treatable psychiatric disorders affecting geriatric patients. At primary care facilities up to 30-50% of depressed patients are misdiagnosed and therefore not treated adequately. Recognition of depression is best improved by routine screening using efficient and easily applicable instruments. One of the most widely used instruments, the Geriatric Depression Scale (GDS), has been sufficiently validated in samples from specialized or out-sourced services. However, few studies, none of them from Brazil, have evaluated the performance of the GDS at the level of primary care in which most geriatric patients with depression are treated, making it hard to generalize existing results to this context. Objectives – 1) to determine the sensitivity, specificity, positive predictive value, negative predictive value and likelihood ratio in relation to the GDS comparing findings to structured clinical interviews for disorders of Axis I of the DSM-IV (SCID-I); 2) to build a Receiver Operator Characteristic Curve (ROC) determining the ideal cut-off point for primary care patients; 3) to determine the prevalence of episodes of major depression in geriatric primary care patients; 4) to identify the socio-demographic profile of geriatric primary care patients; 5) to examine geriatric patients presenting with episodes of depression and/or dysthymic disorders while taking anti-depressants. Method – cross-sectional study involving 220 geriatric patients from four health facilities. A questionnaire providing a global profile of the study population was applied, as well as the GDS and structured clinical interviews for disorders in Axis I of the DSM-IV (SCID-I). The first two instruments were applied by a trained interviewer and the SCID (gold standard) was applied by a trained psychiatrist. Results – The prevalence of episodes of major depression and/or dysthymic disorders was 17.27%. The best cut-off point with GDS-30 was 10/11 (case/non-case) with a sensitivity of 92.1% (CI 77.5–97.9), specificity 79.7% (CI = 72.9–85.1), accuracy 81.8% and likelihood ratio 4.5. The best cut-off point with GDS-15 was 4/5 with a sensitivity of 86.8% (CI = 71.1–95.1), specificity 82.4% (CI = 75.9–87.5), accuracy 83.1% and likelihood ratio 4.9. The best cut-off point with GDS-4 was 0/1 with a sensitivity of 84.2% (CI = 68.1–93.4), specificity 74.7% (CI = 67.7–80.7), accuracy 76.30% and likelihood ratio 3.9. With GDS-1 the sensitivity was 52.6% (CI = 36.0–68.7%). Conclusions – GDS-30 is an excellent screening instrument for episodes of major depression and/or dysthymic disorders at primary care facilities due to its high sensitivity and high negative predictive value. Reduced versions with 15 and 4 items proved good screening tests requiring less time to be applied. The systematic use of GDS at primary care facilities is recommended in order to identify a greater number of cases of depression and thereby offer a more adequate treatment.