Screening Instrument Comparison
Anxiety Screening & Assessment Instrument Comparison
Quick Reference Tool | PsychHQ Source: Module 2, Evaluation & Diagnosis (Lessons 2, 6) | Last Updated: March 2026
Why Instrument Selection Matters
Primary care physicians detect GAD in 29% of cases, panic disorder in 14%, and social anxiety in just 2.2% (Vermani et al., 2011). Structured screening instruments roughly double GP sensitivity for anxiety detection with only a marginal drop in specificity. In pediatric settings, the gap is even wider: community clinicians identified anxiety in 26.2% of youth, while the ADIS-C/P (Anxiety Disorders Interview Schedule — the semi-structured diagnostic gold standard) identified it in 93.2% of the same sample.
The tool you pick depends on three things: who you're evaluating, what clinical question you're answering, and what you can afford.
Pediatric Instruments (Ages 6–18)
| Feature | SCARED | SCAS | RCMAS-2 | MASC-2 | ADIS-C/P | PARS |
|---|---|---|---|---|---|---|
| Ages | 8–18 | 6–18 | 6–19 | 8–19 | 6–18 | 6–18 |
| Items | 41 | 44 (child), 38 (parent) | 49 | 50 | Semi-structured | Clinician-rated |
| Format | 3-point scale (0–2) | 4-point scale (0–3) | Yes/No | 4-point scale | Interview (0–8 severity) | Clinician synthesis |
| Time | 10 min | 10 min | 10–15 min | 15 min | 90+ min per family | 20–30 min |
| Cost | Free | Free | Proprietary | Proprietary | Licensed | Free |
| Informants | Child + Parent | Child + Parent | Child | Child | Child + Parent (parallel) | Child + Parent + Clinician |
| DSM mapping | 5 factors: GAD, panic/somatic, separation, social, school avoidance | 6 factors: adds OCD + specific phobias | 3 subscales: physiological, worry, social | Anxiety Probability Score + GAD Index | Full diagnostic interview with dimensional severity | Single severity metric integrating all sources |
| Key cutoffs | Total ≥25 (sens 71%, spec 61–71%); ≥30 improves specificity | T-scores: <60 normal, 60–64 elevated, 65–69 high, ≥70 very high | T >60 clinically elevated | T 65–69 subclinical, ≥70 clinical | Severity ≥4 = criteria met with impairment | 11.5 (5-item) or 17.5 (7-item) |
| Reliability | Change sensitivity r = 0.51–0.57; cross-ethnic invariance established | α 0.87–0.94; cross-cultural stability (52 studies) | Validity scales (defensiveness, inconsistency) | Advanced predictive metrics | ICC 0.87 (child), 0.96 (parent) | ICC 0.97 |
| Best for | Primary care/specialty screening | Broadest diagnostic coverage (includes OCD, specific phobias) | Younger children (simple yes/no format) | Tertiary care; diagnostic precision | Diagnostic gold standard (research, specialty) | Treatment outcome tracking |
Instrument-by-Instrument Notes
SCARED — The workhorse. Free, fast, validated across African American, Hispanic, White, and Asian youth (Skriner & Chu, 2014). Change sensitivity is strong: percentage reductions in SCARED scores correlate r = 0.51–0.57 with clinician-rated global improvement (Caporino et al., 2017). Parent-child agreement is modest (r = 0.20–0.60) — this reflects the inherent difficulty parents have in observing internal cognitive distress, not a flaw in the instrument. Subscale cutoffs allow disorder-level triage: panic ≥7, GAD ≥9, separation ≥5, social ≥8, school avoidance ≥3.
SCAS — Covers OCD and specific phobias, which the SCARED misses. Cross-cultural stability validated across 52 studies. Preschool adaptations extend it to early childhood. Age- and gender-stratified T-scores (standardized scores where 50 = average and each 10 points = one standard deviation) provide normative comparison.
RCMAS-2 — The yes/no format reduces cognitive load for younger children (ages 6–8 particularly). Built-in validity scales catch defensiveness and inconsistent responding — useful when you suspect a child is minimizing.
MASC-2 — The premium option for specialty settings. The Anxiety Probability Score and dedicated GAD Index give precision that broader tools lack. Per-use licensing limits high-volume use.
ADIS-C/P — Gold standard, but 90+ minutes per family makes it impractical outside specialty and research settings. Parallel child and parent interviews with excellent interrater reliability. If you need a definitive diagnosis with dimensional severity ratings, this is it.
PARS — The treatment-tracking instrument. Clinician-rated, synthesizes child, parent, and clinical observation into a single severity score. ICC (intraclass correlation coefficient — a measure of how consistently different raters assign the same score) of 0.97 — as reliable as it gets. Used as the primary outcome measure in the major pediatric anxiety pharmacological trials (RUPP, CAMS). Free.
Adult Instruments
| Feature | GAD-7 | GAD-2 | PSWQ | BAI | SPIN | LSAS | PDSS | ADIS-5 |
|---|---|---|---|---|---|---|---|---|
| Items | 7 | 2 | 16 | 21 | 17 | 24 | 7 | Semi-structured |
| Score range | 0–21 | 0–6 | 16–80 | 0–63 | 0–68 | 0–144 | 0–28 | 0–8 severity |
| Cost | Free | Free | Free (non-commercial) | Proprietary | Proprietary | Free | Free | Licensed |
| Time | 1–2 min | 30 sec | 2–5 min | 5–10 min | 5 min | 10–15 min | 5 min | 60+ min |
| Key cutoffs | 5 mild, 10 moderate, 15 severe | ≥3 triggers full GAD-7 | 61–65 distinguishes GAD from SAD | No standard clinical cutoffs | ≥19 clinical social phobia | 30–49 mild, 50–64 moderate, 65–79 marked, ≥80 severe | ≥9 needs assessment; ≤5 remission | ≥4 = diagnosis met |
| Psychometrics | At cutoff 10: sens 89%, spec 82% | Sens 86%, spec 83% for GAD | Excellent internal consistency; GAD-specific | Somatic-heavy; poor in elderly | Responsive to pharmacological + CBT treatment | Rates fear and avoidance separately | α 0.91; ICC 0.81; 40% reduction = response | Definitive adult gold standard |
| Best for | Universal screening + monitoring | Ultra-brief pre-screen | Worry quantification; GAD vs other anxiety | Panic-spectrum (somatic emphasis) | Social anxiety screening | Social anxiety severity staging | Panic disorder tracking | Comprehensive diagnostic evaluation |
Key Clinical Notes
GAD-7 — The universal starting point. Beyond GAD, it detects panic disorder (74% sensitivity), social anxiety (72%), and PTSD (66%) — making it a reasonable cross-diagnostic screen. The minimal clinically important difference (MCID) is 4 points, so a change of 3 or fewer from visit to visit is measurement noise, not real clinical change. Measurement invariance (the scale measures the same construct consistently across different demographic groups) holds across sex and racial/ethnic groups, though the anxiety-depression boundary may be experienced differently across populations (the latent anxiety and depression factors correlate more strongly in Black patients).
GAD-2 — Two questions, 30 seconds. Score ≥3 triggers the full GAD-7. Use this as universal triage in any setting where time is scarce.
PSWQ — Measures pathological worry specifically, which is the defining cognitive feature of GAD. Where broader scales often fail to distinguish GAD from other anxiety disorders, the PSWQ does this reliably. A cutoff between 61 and 65 best separates GAD from social anxiety.
BAI — Heavily somatic: numbness, tingling, palpitations. Correlates strongly with panic but systematically underestimates the cognitive worry of GAD. In elderly or medically ill patients, somatic overlap with medical conditions inflates scores artificially (sensitivity 0.70, specificity 0.60 in older adults). The Geriatric Anxiety Inventory (GAI-20) is the better choice for patients 65+ (sensitivity 0.89, specificity 0.80).
PDSS — The panic disorder tracker. Covers attack frequency, distress, anticipatory anxiety, agoraphobic avoidance, interoceptive avoidance (avoiding body sensations that mimic panic, like exercise or caffeine), and functional impairment — all in 7 items. Remission = PDSS ≤5. Clinical response = 40% or greater reduction from baseline.
Clinical Decision Framework
| Clinical Situation | Recommended Instrument(s) |
|---|---|
| Primary care screening (pediatric) | SCARED (free, validated, subscale cutoffs for disorder-level triage) |
| Primary care screening (adult) | GAD-2 → if ≥3, full GAD-7. Pair with PHQ-9 (Patient Health Questionnaire-9, a depression screener) to catch the common anxiety-depression overlap |
| Broadest pediatric diagnostic coverage | SCAS (covers OCD and specific phobias that SCARED misses) |
| Specialty diagnostic evaluation (pediatric) | ADIS-C/P (gold standard) or MASC-2 (if time-limited) |
| Specialty diagnostic evaluation (adult) | ADIS-5 (gold standard) + disorder-specific scale |
| Treatment monitoring (pediatric) | PARS (clinician-rated, ICC 0.97, free) |
| Treatment monitoring (adult) | GAD-7 at every visit (2-min burden, MCID of 4 points); add PDSS or LSAS during targeted therapy |
| Worry quantification / GAD vs. other anxiety | PSWQ (cutoff 61–65 for GAD vs SAD) |
| Social anxiety staging | LSAS (24-item, fear + avoidance rated separately) |
| Panic disorder tracking | PDSS (7-item, response and remission thresholds defined) |
| Older adults (65+) | GAI-20 (sens 0.89, spec 0.80) — avoid BAI (somatic weighting inflates scores) |
| Budget-constrained practice | SCARED + PARS (pediatric) or GAD-7 + PSWQ (adult) — all free |
When Informant Reports Disagree
Cross-informant agreement for pediatric internalizing symptoms is low: correlations between parent, teacher, and child self-report typically range from r = 0.14 to 0.28 (Miller et al., 2014; Dirks et al., 2012). This is not a measurement failure. Different informants observe different contexts — a child may have severe separation anxiety at home that the teacher never sees, or social anxiety at school that parents don't witness.
When reports diverge, the clinical question is not "who is right?" but "what is different about these environments?" A child who scores low on teacher report but high on parent report may have strong classroom structure compensating for deficits that emerge in less structured home settings — or the reverse.
Weighting shifts with development. For young children where self-report is unreliable, parent report plus clinical observation carries the weight. As adolescents develop introspective capacity, self-report becomes the primary source for internalizing symptoms. If the chief concern involves school refusal or peer difficulties, teacher reports move from optional to mandatory.
Functional Assessment: A Separate Measurement
Symptom severity only modestly predicts functional impairment (mean correlation r = 0.34 across six primary anxiety disorders; McKnight et al., 2016). Two patients with identical GAD-7 scores can have entirely different functional trajectories. Validated tools for measuring impairment independently include the Sheehan Disability Scale (SDS — 3 items, score >5 on any domain = significant), WHODAS 2.0, CGAS (Children's Global Assessment Scale, 1–100), and WFIRS (domain mean >1.5 = significant impairment). Anxiety symptom scales do not replace functional assessment.
Key References: Birmaher et al. (1997/1999, SCARED); Spence (1998, SCAS); Spitzer et al. (2006, GAD-7); Kroenke et al. (2007, GAD-2); Shear et al. (1997, PDSS); Liebowitz (1987, LSAS); Meyer et al. (1990, PSWQ); Skriner & Chu (2014, SCARED invariance); Caporino et al. (2017, SCARED change sensitivity); Vermani et al. (2011, GP detection rates); McKnight et al. (2016, symptom-function correlation); Atchison et al. (2024, GAI-20); Miller et al. (2014) & Dirks et al. (2012, cross-informant agreement)