Skip to content

Spitzer RL, et al. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. 2006;166:1092-7. PMID 16717171

One-paragraph summary

A criterion-standard study across 15 US primary-care clinics between November 2004 and June 2005. Of 2,740 adult patients completing a study questionnaire, 965 had a telephone interview with a mental health professional within one week. A seven-item self-report anxiety scale was developed and evaluated against those independent diagnoses, against functional status measures, disability days and health-care use. The GAD-7 showed good reliability and criterion, construct, factorial and procedural validity; a cut-point was identified that optimised sensitivity (89%) and specificity (82%). Increasing scores were strongly associated with impairment on all six Medical Outcomes Study Short-Form General Health Survey scales and with disability days. Although GAD and depression symptoms frequently co-occurred, factor analysis confirmed them as distinct dimensions with differing but independent effects on functional impairment and disability. Self-report and interviewer-administered versions agreed well.

Key findings

  • Sensitivity 89% / specificity 82% at the identified cut-point (conventionally ≥10).
  • Dose–response association between score and every domain of functional impairment measured, plus disability days and health-care use.
  • GAD and depression symptoms loaded as separate factors with independent effects on impairment — a psychometric argument for the diagnosis's separability that predates and partly counterweights the genetic evidence for near-complete overlap.
  • Agreement between self-report and interviewer administration, which is what made large-scale deployment possible.

Limitations

  • Development sample: the 89%/82% figures are internal to the study that produced the instrument. The 2025 Cochrane synthesis of 48 studies and 19,228 participants puts pooled sensitivity at 0.64 (95% CI 0.56–0.72) with specificity 0.91 (Aktürk 2025, PMID 40130828). The difference may reflect development-sample optimism, case mix, settings and thresholds; the pooled estimate should not be reduced to a single cause.
  • US primary care only, one payer environment, one language at development.
  • The scale operationalises DSM-IV GAD. It is now widely deployed in systems moving toward ICD-11, which has neither an excessiveness requirement nor a fixed symptom count.
  • Later work identified cultural differential item functioning: Black/African American participants with high symptoms scored lower than others at equivalent symptom levels (Parkerson 2015, PMID 25725310).

Why it matters

The GAD-7 made brief measurement of generalized-anxiety symptoms practical across surveys, primary care, digital trials and service datasets. Its reach creates a recurring interpretation risk: a screening threshold is not a diagnosis. Screening-scale symptom prevalence (18.2% of US adults reporting past-two-week symptoms; Terlizzi 2024, PMID 39591466) and interview-based 12-month GAD prevalence (1.8%; Ruscio 2017, PMID 28297020) measure different constructs, populations and time frames and should not be presented as interchangeable.

Cited by wiki pages

  • screening-and-measurement.md
  • diagnosis-and-classification.md
  • the-diagnostic-boundary.md
  • epidemiology-and-burden.md