Skip to content

Ruscio AM, et al. The case for eliminating excessive worry as a requirement for generalized anxiety disorder: a cross-national investigation. Psychol Med. 2024;54:3447-3458. PMID 39364896

One-paragraph summary

Using the WHO World Mental Health Survey Initiative — 133,614 adults from 12 surveys in low- and middle-income countries and 16 in high-income countries, assessed with the Composite International Diagnostic Interview — the authors identify people who meet every DSM-5 criterion for GAD except that their worry is not judged "excessive" relative to context, and compare them with full-criteria cases and with people without GAD. Removing the excessiveness requirement raises global lifetime GAD prevalence from 2.6% to 4.0%, with larger increases in LMICs than HICs. Non-excessive and excessive cases worry about many of the same things, though non-excessive cases worry more about the health and welfare of loved ones and less about personal or non-specific concerns. Non-excessive cases closely resemble excessive cases in sociodemographic characteristics, family history of GAD, and risk of temporally secondary comorbidity and suicidality. They are less severe on average but report comparable impairment and often seek treatment for GAD symptoms. The authors conclude that eliminating the excessiveness requirement would produce a more defensible diagnosis.

Key findings

  • Global lifetime prevalence 2.6% → 4.0% on removal of one criterion.
  • The increase is larger in LMICs — the requirement disproportionately excludes people living in objectively difficult circumstances, which is the paper's opening argument.
  • Non-excessive cases match excessive cases on family history of GAD, secondary comorbidity risk and suicidality: three validators that a criterion ought to separate if it is doing work.
  • Impairment is comparable despite lower average severity, and treatment-seeking occurs.

Limitations

  • Lay-administered structured interview; "excessiveness" is precisely the item with the worst measurement properties in that instrument (κ rises from 0.53 to 0.78 when it is removed; Wittchen 1995, PMID 7666382), so part of the finding may be that an unreliable criterion cannot separate groups.
  • Cross-sectional; course is inferred.
  • Excessive cases are more severe on average — the paper does not claim the criterion carries no information, only that it does not justify exclusion.
  • No treatment-response comparison between the two groups, which is the validator that would settle it.

Why it matters

This is the most consequential open question in GAD, published by the group that produced the field's reference epidemiology. If the criterion goes, prevalence rises by roughly half, the boundary with subthreshold GAD moves (subthreshold GAD is already about twice as prevalent as threshold GAD; Haller 2014, PMID 24886240), and the entire treatment evidence base — which recruits DSM-defined cases — becomes an evidence base for a subset of the disorder. It also converges with two independent lines: the reliability literature, where discordance concentrates on exactly this criterion; and the incremental-validity literature, where uncontrollability predicts clinical outcomes over and above excessiveness while the reverse does not hold (Hallion 2013, PMID 23713499). Nothing has changed in the manuals. That gap between published argument and diagnostic practice is the substance of the diagnostic boundary.

Cited by wiki pages

  • the-diagnostic-boundary.md
  • diagnosis-and-classification.md
  • epidemiology-and-burden.md
  • guidelines.md
  • red-flags-and-safety-concerns.md
  • patient-experience-and-advocacy.md
  • overview.md