Skip to content

Ruscio AM, et al. Cross-sectional Comparison of the Epidemiology of DSM-5 Generalized Anxiety Disorder Across the Globe. JAMA Psychiatry. 2017;74:465-475. PMID 28297020

One-paragraph summary

The first population-based data on GAD as defined in DSM-5, from the WHO World Mental Health Survey Initiative: 147,261 adults aged 18–99, interviewed face to face in representative household samples across 26 countries, surveyed between 2001 and 2012 with a single consistent protocol and instrument (the Composite International Diagnostic Interview), weighted mean response rate 69.5%. Combined lifetime prevalence was 3.7% (SE 0.1%), 12-month 1.8% (SE 0.1%), 30-day 0.8%. Prevalence varied widely by country income: lifetime 5.0% in high-income, 2.8% in middle-income, 1.6% in low-income countries. GAD typically began in adulthood and persisted, with later onset and a more persistent course in lower-income countries. Lifetime comorbidity was 81.9% (SE 0.7%), driven by mood (63.0%) and other anxiety disorders (51.7%). Severe role impairment affected 50.6% of cases. Roughly half sought treatment (49.2%), more often those with severe impairment (59.4%), comorbidity (55.8%) or residence in a high-income country (59.0%).

Key findings

  • DSM-5 GAD is more prevalent than DSM-IV GAD — the criteria change moved the number.
  • A threefold high-income to low-income gradient in lifetime prevalence under an identical protocol.
  • Onset in adulthood, distinguishing GAD from the childhood-onset phobias.
  • Half of all lifetime cases experience severe role impairment; half seek treatment.
  • The comorbidity figure (81.9% lifetime) that is routinely quoted — and routinely mis-transferred to point prevalence, where primary-care data show pure GAD outnumbering comorbid GAD/MDE 3.8% to 1.6% (Wittchen 2002, PMID 12044105).

Limitations

  • Lay-administered fully structured interview (CIDI). Concordance between CIDI and clinician SCID diagnosis of lifetime GAD has been estimated at κ=0.35, rising to 0.66 when the excessiveness criterion is removed (Wittchen 1995, PMID 7666382).
  • Surveys were fielded 2001–2012 and DSM-5 criteria applied retrospectively to that data.
  • Cross-sectional: course and persistence are reconstructed, not observed.
  • The income gradient cannot distinguish true prevalence differences from measurement, expression and detection differences — the paper says so.

Why it matters

This is the reference epidemiology for GAD, and every subsequent prevalence claim is either this study, a national survey compared against it, or a screening-scale figure being mistaken for it. It also set up the field's central nosological argument: the same group later used the same data to show that removing the excessiveness requirement raises global lifetime prevalence from 2.6% to 4.0% while identifying cases that resemble diagnosed cases on family history, comorbidity, suicidality and impairment (Ruscio 2024, PMID 39364896). The 2017 paper establishes what GAD is at population scale; the 2024 paper argues that what it is depends on a criterion the same investigators would delete.

Cited by wiki pages

  • epidemiology-and-burden.md
  • the-diagnostic-boundary.md
  • diagnosis-and-classification.md
  • comorbidity-and-primary-care.md
  • course-relapse-and-long-term-outcome.md
  • overview.md