The GDS-5: a five-item depression screen for older adults who tire of long forms
- In 200 community-dwelling Iranian older adults, the five-item Geriatric Depression Scale (GDS-5) reproduced a single coherent factor on exploratory factor analysis — the shortened form did not fragment into noise.
- Diagnostic accuracy against a criterion was high: AUC = 0.924, with sensitivity of 0.98 and specificity of 0.77 at the optimal cut-off. As a first-pass screen, it almost never misses a depressed older adult.
- Concurrent validity against the CES-D was strong (r = 0.71, p < .001), and test–retest reliability was excellent (ICC = 0.809, p < .001) — the score is stable across administrations.
- The instrument is five yes/no items; administration takes under two minutes and requires no clinician training, making it usable by nurses, GPs, and social workers, not only mental-health specialists.
Late-life depression is the screening problem nobody enjoys. The patient is fatigued, often cognitively slowed, frequently somatically preoccupied, and rarely volunteers low mood as a complaint. The standard GDS-15 is already a concession to brevity, yet in a busy geriatric clinic or a home visit, fifteen items is still fifteen items. This study asks whether you can cut to five and keep the signal. The answer, in this sample, is essentially yes.
What the data shows
The authors validated the Persian GDS-5 in 200 older adults using exploratory factor analysis, concurrent validity against the CES-D, ROC analysis against a depression criterion, and test–retest reliability. The single-factor solution is the headline: a five-item self-report measuring one latent construct, not a grab-bag of weakly related symptoms. The discrimination is what makes it clinically interesting — an AUC of 0.924 places it firmly in the "good-to-excellent" band, and the sensitivity of 0.98 is the property you actually want from a screen. A screening instrument earns its place by rarely letting a true case through; specificity (0.77 here) is the cost you pay in false positives, which a second-stage assessment then filters. The CES-D correlation of 0.71 and the retest ICC of 0.809 confirm the obvious worry — that brutal item reduction would buy speed at the price of stability — does not materialise here.
For your practice
Use the GDS-5 as a triage gate, not a diagnosis. Its design point is the setting where a fuller assessment is not yet feasible: the first contact in primary care, the routine geriatric review, the community nurse's home visit. A positive screen means "do the full interview", not "start an antidepressant". Because the items are dichotomous and the language is plain, you can delegate administration to non-specialist staff and reserve your own time for the positives. Two practical cautions for real clinics: the high sensitivity means you will generate false positives, so build in a confirmatory step rather than acting on the five-item score alone; and in patients with established cognitive impairment, self-report grief and mood scales degrade — pair the screen with a collateral or observer measure when dementia is on the table.
A five-item screen with 0.98 sensitivity is not a diagnostic instrument — it is a door you open quickly so that no depressed older patient walks past it unseen.
Single-country (Iranian) cross-sectional sample of 200, so cut-offs and specificity will need local re-checking before transfer to other health systems; the criterion was screening-based rather than a structured diagnostic interview, and generalisation to clinical and long-term-care populations awaits further validation.