Article Text

Download PDFPDF

The 15 item Geriatric Depression Scale had perfect sensitivity but low specificity for detecting depression in elderly people
  1. Rob Butler, MBBS, MRCPsych
  1. Edgeware Community Hospital Brain Injury Unit, Middlesex, UK

Statistics from

OpenUrlCrossRefPubMedWeb of Science

QUESTION: In elderly people (≥75 y of age), is the shorter, 15 item Geriatric Depression Scale (GDS15) a useful screening tool for detecting depression and anxiety?


Blinded comparison of GDS15 with diagnoses based on the ICD-10 criteria for mood disorders.


A large general practice in Melton Mowbray, Leicestershire, UK.


201 patients (median age 79 y, 57% women) who were ≥75 years of age, were not in residential care, and had completed the diagnostic interview after completing the GDS15 during a health check. All patients who scored ≥5 on the GDS15, 1 in 2 patients who scored 3 or 4, and 1 in 4 of remaining patients were asked to complete the diagnostic interview (n=257).

Description of test and diagnostic standard

The GDS15 was given by a practice nurse during regular annual home health checks. A geriatric psychiatrist and a research nurse, who were blind to the GDS15 results, used the Schedules for Clinical Assessment in Neuropsychiatry to make ICD-10 diagnoses.

Main outcome measures

Sensitivity, specificity, and likelihood ratios.

Main results

Results are shown in the table.

Test characteristics of GDS15 for detecting depression and anxiety*


In elderly people (≥75 y of age), the shorter, 15 item Geriatric Depression Scale (GDS15) had perfect sensitivity but low specificity for detecting depression when a cutoff score of 3 was used. The sensitivity and specificity of the GDS15 for detecting anxiety disorders were not high.


There is no doubt that most depression in older people goes undetected and untreated.1 Patients and doctors seem to collude to avoid the stigma and resignation that the diagnosis brings. This happens even though a simple intervention (a pill) has been shown to be helpful for over one half of this group, and a range of talking therapies seem to be equally effective.2 One way of improving the detection of depression is by using screening questionnaires. Checklists are generally disliked by healthcare professionals and may be tiresome for older clients. Hence the interest in looking at the efficacy of brief scales (even though the briefest scale possible [BSP] “Are you feeling depressed?” has yet to be implemented universally). In fact, the 15 questions in the GDS15 are similar to the 8 questions in the EBAS DEP, raising the possibility of one day unifying them (although the ensuing acronym would be unbearably long). The main difference between the studies is the populations they examine. The GDS15 was studied in people ≥75 years of age who lived in the community as part of their legislated (in the UK) annual health checkups. The EBAS DEP was used in a smaller group of people who were ≥65 years of age in Germany and lived in an old people's home or hospital or attended a clinic. Each study looked at how different cutoff scores for the scales compared with a gold standard of clinical criteria for depression. There was the usual tradeoff between sensitivity and specificity. If only 3 questions need to be answered “yes” to screen the patient as positive, then both scales are very sensitive (picking up all the small number of depressed patients in each study). The scales also identify a lot of “false-positives”, however, which is inevitable when the prevalence of a disorder is low. This may mean further unnecessary clinical interviews. No one knows whether those with false positive results may also benefit from further interventions. The take home message is that both scales are useful for screening elderly people. Every silver lining has a cloud, however. Once detected, depressed older people will need to be treated properly. Even with widely practicable treatments such as drugs, evidence exists that they are not always used well and that if quality is improved, so are outcomes.3

If the average geriatric psychiatrist covers the area of 50 general practitioners (GPs), Arthur et al's figures suggest that the GDS15 would screen 2100 people ≥75 years of age as positive, of whom 400 would be depressed. This would generate too much work for existing services. Perhaps this is one reason why GPs are not presently using depression scales in their annual screening of people who are ≥75 years of age.4



  • Source of funding: Nuffield Provincial Hospitals Trust.

  • For correspondence: Mr A Arthur, Department of Epidemiology and Public Health, University of Leicester, 22–28 Princess Road West, Leicester LE1 6TP, UK. Fax +44 (0)166 448 1427. Email aja{at}

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.