Depression · 30 items · 0–30 · Yesavage JA, Brink TL, Rose TL, et al. Development and validation of a geriatric depression screening scale: a preliminary report. Journal of Psychiatric Research. 1983;17(1):37-49. https://doi.org/10.1016/0022-3956(82)90033-4
The Geriatric Depression Scale (GDS-30) is a 30-item self-report screening measure of depressive symptoms designed specifically for older adults. Each item is answered yes or no with reference to how the person has felt over the past week, and one point is given for each answer in the depressive direction, producing a total score from 0 to 30. The items concentrate on the cognitive and affective features of late-life depression - dissatisfaction with life, emptiness, hopelessness, withdrawal from activities - rather than symptom counts drawn from general-adult criteria.
What sets the GDS-30 apart is what it deliberately leaves out. Somatic complaints such as poor sleep, low energy and appetite change are common in older people for reasons unrelated to depression, so the developers excluded them to avoid inflating scores in medically ill or simply ageing respondents. The forced-choice yes/no format was likewise chosen to keep the cognitive demand low. The result is a screening instrument, not a diagnostic one: a high score signals that a clinical assessment for depression is warranted.
02 - Origin & purpose
Where it comes from.
The GDS was developed by Jerome Yesavage, T.L. Brink and colleagues at Stanford, published in the Journal of Psychiatric Research in 1982-83. The team began with a pool of 100 candidate questions generated by clinicians experienced in geriatric depression, administered them to older adults with and without depression, and retained the 30 items that correlated most strongly with the total score. Existing instruments of the day - the Hamilton, the Zung, the Beck inventory - had been built for younger adults and leaned heavily on somatic symptoms, which is precisely where they misfire in older populations.
The scale spread quickly through geriatric medicine, nursing homes and memory clinics, and spawned a family of short forms, most prominently the 15-item GDS-15 (Sheikh and Yesavage, 1986). The long form remains in use where a broader item sample is wanted, and the developers have placed the instrument in the public domain, with translations collected on the Stanford GDS website.
03 - Scoring & cutoffs
How scoring works.
Each of the 30 items scores one point when answered in the depressive direction - for twenty items that is 'yes', and for ten reverse-keyed items (such as being satisfied with life or feeling happy most of the time) it is 'no'. Totals are conventionally banded as 0-9 normal, 10-19 mild depression and 20-29 severe depression. In the original validation work a threshold of 11 or above correctly identified 84% of people with depression while correctly ruling it out in 95% of those without; a stricter threshold of 14 raised specificity to 100% at the cost of sensitivity (80%). The bands are a screening aid, not diagnostic categories - scores above threshold call for a clinical interview.
Score
Severity
Interpretation
0–9
Normal
No or minimal depressive symptoms. Monitor and reassure.
10–19
Mild depression
Mild depressive symptoms. Watchful waiting; consider repeating and further assessment.
20–30
Severe depression
Severe depressive symptoms. Clinical assessment and active treatment indicated.
A screening result in the depressive range warrants clinical assessment; the GDS-30 is a screen, not a diagnosis.
04 - Validation evidence
How well it performs.
The original Stanford validation compared the GDS against research diagnostic criteria in older adults with and without depression and reported strong internal consistency and stability alongside the classification figures above. Later meta-analytic work across community and clinical settings found the long form's performance more variable, with pooled specificity notably lower than in the derivation samples - a familiar pattern when screening instruments leave their home populations. Reliability has been replicated across translations and settings, including very old and cognitively impaired samples, though internal consistency falls as cognitive impairment deepens.
84% / 95%
Sensitivity / specificity (cutoff ≥ 11)
Yesavage et al., 1983
α = 0.94
Internal consistency
Yesavage et al., 1983
r = 0.85
Test–retest reliability (1 week)
Yesavage et al., 1983
77% / 65%
Pooled sensitivity / specificity
Wancata et al., 2006 meta-analysis
05 - How it compares
How it compares to the alternatives.
Instrument
Items
Time
When to reach for it
GDS-30
30
5–10 min
This page. The long form, used where a broader item sample is wanted.
-Moderate to severe dementia - self-report becomes unreliable; use an observer-rated instrument such as the Cornell Scale for Depression in Dementia
-Tracking week-to-week treatment response - no established minimal clinically important difference; the PHQ-9 or MADRS serve better
-Diagnosis - no screening scale diagnoses depression; a positive screen needs a clinical interview
07 - Confidence & precision
Reading the score with care.
No minimal clinically important difference has been established for the GDS-30, so score changes are hard to interpret as treatment response. For context, minimal detectable change estimates in the GDS family run around 4.3-4.5 points for the 15-item version (Snellman et al., 2024; Sugishita et al., 2016) and about 6.6 points for a 27-item version in cognitively impaired older adults - differences smaller than this fall within measurement noise. Treat the GDS-30 as a screening snapshot rather than an outcome measure.
08 - Limitations
What it cannot tell you.
Specificity outside derivation samples is modest: pooled estimates around 65% mean roughly one in three positive screens in community settings may be false positives (Wancata et al., 2006).
Validity degrades with cognitive impairment; internal consistency falls as MMSE scores drop, and self-report in moderate-to-severe dementia is unreliable (Conradsson et al., 2013).
The yes/no format sacrifices severity gradation; two respondents with the same score can differ considerably.
At 30 items it is long for routine screening; the GDS-15 captures most of the signal in half the time.
Excluding somatic items avoids false positives from physical illness but can miss depression presenting mainly somatically.
See how Aisel removes friction where it costs most. A 20-minute walkthrough tailored to your clinic.
We value your privacy
We use cookies to analyse site usage and improve your experience. Analytics and embedded media (e.g. YouTube) only load if you accept. Read our cookie policy.