Depression in older adults is both common and commonly missed. Major depression affects 6-10% of older adults in primary care, 12-20% of nursing home residents, and up to 45% of hospitalized elderly patients. Clinical judgment alone misses most cases—systematic screening is essential.
The challenge is that late-life depression looks different. Research shows older adults present with more somatic symptoms, more agitation, more hypochondriasis, and more gastrointestinal complaints than younger adults—but less guilt and less emphasis on sad mood. Many experience "depression without sadness," where apathy and physical complaints dominate while mood symptoms stay muted. This atypical presentation makes detection harder.
The instruments
The GDS-15 was designed specifically for older adults in the 1980s. Its key feature: it deliberately excludes somatic symptoms. All 15 items focus on psychological and functional aspects—life satisfaction, interests, hope, energy, social withdrawal. By avoiding sleep, appetite, and fatigue questions, the GDS prevents medical illness from inflating depression scores. The yes/no format is easier for patients with cognitive or sensory impairments.
The PHQ-9 is the standard primary care screener, validated across age groups including older adults. It maps directly to DSM criteria for major depression. Four of its nine items assess somatic symptoms (sleep, fatigue, appetite, psychomotor changes)—which can be problematic when medical illness causes those same symptoms.
What the research shows
Head-to-head studies find comparable performance. A Swedish study of community-dwelling adults 65+ found AUC of 0.97 for GDS-15 and 0.95 for PHQ-9—no significant difference. A Chinese study of 1,546 older adults showed 96% concordance between instruments.
At standard cutoffs, the GDS-15 (cutoff ≥5) has sensitivity around 81-94% and specificity 78-88%. The PHQ-9 (cutoff ≥10) shows sensitivity 71-84% and specificity 81-95%. Notably, a 2024 meta-analysis found GDS-15 ≥8 provided the closest match to structured clinical interview diagnoses.
One important finding: PHQ-9's standard cutoff of 10 may be too high for older adults. Lowering to 5 yielded 100% sensitivity (but only 81% specificity) in one study. Consider adjusting based on whether you prioritize catching cases versus avoiding false positives.
When they diverge
The instruments measure overlapping but distinct constructs. GDS-15 captures psychological distress without somatic confounding. PHQ-9 captures the full DSM symptom picture, including physical manifestations.
GDS-15 advantages: In medically complex patients—multiple chronic conditions, recent hospitalizations, prominent physical symptoms—the PHQ-9 may overestimate depression because patients endorse somatic items reflecting illness rather than mood. The GDS-15 sidesteps this problem. Its yes/no format also works better for patients with mild cognitive impairment.
PHQ-9 advantages: Some older adults experience depression primarily through physical symptoms—the somatic presentation that characterizes late-life depression. For these patients, PHQ-9 may capture the clinical picture better than GDS-15's purely psychological focus. PHQ-9 is also already integrated into most EHRs, familiar to clinicians across specialties, and required by some quality programs.
Choosing an instrument
Use GDS-15 when: Medical comorbidity is significant and could confound somatic symptoms. Patients have mild cognitive impairment. Your practice is geriatric-focused (nursing homes, geriatric clinics, home care). Research protocols require it.
Use PHQ-9 when: You want one instrument across all ages for workflow simplicity. Patients are relatively healthy without major somatic confounding. DSM-aligned assessment matters. Quality measures require it. You're tracking patients longitudinally from middle age.
Use both when: You want comprehensive baseline assessment. Diagnostic uncertainty exists about whether symptoms are depression versus medical illness. Treatment targets both psychological and somatic symptoms.
Discrepant results are clinically informative. High PHQ-9 with low GDS-15 suggests somatic symptoms may be driving the score—explore medical causes. High GDS-15 with low PHQ-9 suggests psychological distress without prominent physical manifestations.
Practical considerations
Both instruments can be self-administered, interviewer-administered, or given by phone. For older adults with sensory or cognitive limitations, reading questions aloud often improves completion and accuracy.
Screen at annual wellness visits at minimum. Repeat at 2-4 week intervals when tracking treatment response. Care transitions (hospital discharge, nursing home admission) warrant fresh screening given elevated depression risk.
For patients with moderate-to-severe dementia, self-report instruments become unreliable. Consider observational measures like the Cornell Scale for Depression in Dementia.
After a positive screen
Review which specific items were endorsed—this guides clinical interview. Explore symptom onset, precipitants, severity, and functional impact. Assess for medical explanations and medication effects. Ask directly about suicidal thoughts; older adults have the highest suicide rates, particularly older white men.
Both psychotherapy (especially problem-solving therapy and behavioral activation) and antidepressants work in older adults. Start medications low and go slow. Monitor for falls, cognitive effects, and hyponatremia.
---
For most primary care settings, the PHQ-9 handles geriatric depression screening adequately—it's familiar, integrated, and performs comparably to instruments designed for older adults. Switch to the GDS-15 when medical complexity makes somatic symptom interpretation unreliable, or when your practice focuses primarily on geriatric populations. The PHQ-2 works as an ultra-brief first screen if positive results trigger full assessment.