The GAD-7 is the workhorse anxiety screen in primary care, talking-therapy services and psychiatry trials — seven items, a 0–21 total, and three numbers that change what you do next. Yet trainees routinely treat the total as a vague severity dial rather than reading what each band actually signals. This guide gives you the full picture of GAD-7 scoring and interpretation: how the seven items map onto DSM-5-TR criteria, where the 5/10/15 cut-points came from, and how to use serial scores to judge whether a treatment is working.
The instrument was developed by Spitzer, Kroenke, Williams and Löwe and published in 2006, with the diagnostic validation performed against the MINI structured interview in 2,740 patients across primary care and obstetrics–gynaecology clinics. Two decades later it remains one of the best-validated brief screens in psychiatry, and it is the anxiety counterpart to the PHQ-9 in almost every quality metric you will meet.
What the GAD-7 actually measures
Each of the seven items is scored 0 (not at all) to 3 (nearly every day) for how much the symptom has bothered the patient over the previous two weeks — the same two-week window used by DSM-5-TR and ICD-11 for a generalised anxiety disorder diagnosis. Add the items for a total of 0–21.
- Nervous, anxious or on edge
- Not being able to stop or control worrying
- Worrying too much about different things
- Trouble relaxing
- Being so restless that it is hard to sit still
- Becoming easily annoyed or irritable
- Feeling afraid as if something awful might happen
Read the item set against the DSM-5-TR criteria. GAD requires excessive anxiety and worry occurring more days than not for at least six months, plus three of six associated symptoms: restlessness, fatigability, concentration problems, irritability, muscle tension and sleep disturbance. The GAD-7 captures worry control, restlessness, irritability and a proxy for tension — but it contains no item on fatigue, concentration or sleep. That is why a low score never excludes GAD in an older adult presenting with fatigue and insomnia, and why the interview still has to cover the full criterion set.
GAD-7 scoring and interpretation: the 5/10/15 bands
The conventional interpretation bands, established in the original validation paper, are the ones examiners expect:
| GAD-7 total | Severity | What the band means and the suggested next step |
|---|---|---|
| 0–4 | Minimal | Anxiety unlikely to be clinically significant. No treatment indicated; rescreen at routine intervals or if the clinical picture changes. |
| 5–9 | Mild | Subthreshold symptoms. Watchful waiting with reassessment in 2–4 weeks, brief psychoeducation, sleep and caffeine review, and behavioural strategies. |
| 10–14 | Moderate | The working cut-point. Confirm the diagnosis with a structured assessment, quantify functional impairment, and initiate treatment — CBT or an SSRI/SNRI per NICE guidance. |
| 15–21 | Severe | Active treatment indicated; assess comorbid depression (they co-occur in over half of primary-care cases), panic, substance use and suicidality, and consider specialist referral. |
Bands describe severity, not diagnosis. A patient scoring 12 with clear functional impairment and six months of uncontrollable worry is a different proposition from a scoring 12 who is going through a difficult fortnight — the interview adjudicates, the score frames.
Exam trap
The clinically important cut-point is GAD-7 ≥10, not ≥5. A score of 5 is where symptoms become worth watching, but 10 is the threshold used for screening positivity and treatment decisions in the original validation and in downstream trials. Vignettes frequently offer "reassure and discharge" for a score of 11 because "10 is only moderate" — the correct reading is that ≥10 warrants diagnostic evaluation and a treatment decision. Conversely, a score of 6 should not be presented as a diagnosis of GAD.
Where the cut-points come from
In Spitzer and colleagues' 2006 study (Archives of Internal Medicine), the GAD-7 was compared with the MINI diagnosis across 2,740 patients. At a cut-off of ≥10, sensitivity was 89% and specificity 82% — excellent for a seven-item instrument, and the reason ≥10 became the standard positive-screen threshold. Internal consistency was high (Cronbach's α around 0.89), and the score correlated with functional impairment, self-reported difficulty, and health costs, so severity bands track real-world burden rather than statistical convenience alone.
The 5 and 15 bands were chosen as anchors: 5 as the point below which symptoms are unlikely to be clinically meaningful, 15 as the point identifying severe anxiety. Beyond GAD, the instrument performs reasonably as a screen for panic disorder, social anxiety disorder and PTSD — a useful fact when a vignette describes broad anxiety symptoms but it is the monitoring use that dominates practice.
Tracking response: what a change in score means
A 50% reduction in the GAD-7 total is the conventional definition of treatment response. A patient starting at 18 who reaches 9 after eight weeks of CBT has responded; a fall from 18 to 14 has not, despite looking like "some improvement". In trial work, a score of 4 or below is commonly used as remission.
In practice, re-administer the GAD-7 at every treatment contact or at fixed 4–12 week intervals, alongside the PHQ-9 when depression is comorbid. Services running routine outcome monitoring — including NHS Talking Therapies, where the PHQ-9 and GAD-7 are the standard paired measures — rely on this serial logic: one score is a photograph, a series is a trajectory.
Three checks before you sign the score off. First, verify the timeframe: is this genuinely the past two weeks? Second, ask what changed — new medication (salbutamol, steroids, thyroid replacement, stimulants), caffeine, alcohol withdrawal and thyroid disease can all inflate the total. Third, read item-level patterns: a patient scoring 3 on restlessness and 0 on worry may be describing akathisia or a physical illness rather than anxiety.
Limits worth remembering
The GAD-7 is a screening and severity measure, not a diagnosis. Positive screens require a diagnostic assessment that excludes bipolar disorder, substance-induced anxiety, hyperthyroidism, phaeochromocytoma (rare, but the exam loves it) and stimulant effects. It also says nothing about the six-month duration requirement — the score cannot distinguish six months from six days.
Its strengths are brevity, free availability and sensitivity to change. Its weaknesses are the missing fatigue/concentration/sleep items, performance anxiety in patients who under-report, and the fact that cut-offs are population-derived: an individual scoring 9 is not thereby free of GAD.
If you want the paired depression instrument covered at the same level, our guide to structured study notes shows how the PHQ-9 and GAD-7 are tested together in vignette format, and the concept library includes the DSM-5-TR criterion sets side by side.
Frequently asked questions
What do GAD-7 scores of 5, 10 and 15 mean in clinical practice?
5–9 is mild, subthreshold anxiety: watchful waiting and reassessment in 2–4 weeks. 10–14 is moderate: the positive-screen threshold that triggers diagnostic evaluation and a treatment decision. 15–21 is severe: active treatment indicated, with assessment of comorbid depression, substance use and suicidality. The bands describe severity; only a clinical interview establishes the diagnosis.
Is a GAD-7 score of 10 enough to diagnose generalized anxiety disorder?
No. At ≥10 the instrument screens positive with roughly 89% sensitivity and 82% specificity against a structured interview, which is excellent for case-finding but leaves real false positives. A diagnosis still requires the DSM-5-TR elements the questionnaire does not measure: excessive worry on most days for at least six months, difficulty controlling it, three of six associated symptoms, significant distress or impairment, and exclusion of substance, medication and medical causes.
How do you calculate treatment response using the GAD-7?
Compare follow-up scores with the baseline: a reduction of 50% or more is the conventional definition of response — 18 to 9 is a response, 18 to 14 is not. Many trials define remission as a final score of 4 or below. Percentage change requires an accurate baseline, so record the first score before treatment starts and repeat the questionnaire at fixed intervals rather than whenever the patient volunteers that things feel better.
How often should the GAD-7 be repeated during treatment?
At every treatment contact or every 4–12 weeks, whichever the service uses, and alongside the PHQ-9 when depression is present. Re-administration is what converts the score from a one-off severity label into an outcome measure, and it is the basis on which services justify continuing, changing or stepping up treatment.
Key takeaways
- Seven items, 0–3 each, past two weeks, total 0–21; bands at 5 / 10 / 15.
- ≥10 is the positive-screen and treatment-decision threshold — not 5.
- Spitzer 2006: at ≥10, sensitivity 89%, specificity 82% against the MINI.
- 50% score reduction = response; ≤4 commonly = remission; take a baseline first.
- The instrument misses fatigue, concentration and sleep items — interview for them.
- A score frames severity; the diagnostic interview, including the six-month duration, does the rest.
Turn the bands into recall you can retrieve under exam pressure with Preppsy's spaced-repetition flashcards, then test interpretation in context with a full mock exam.
Cover image: “GAD-7 Scoring and Interpretation (cover)” by Preppsy Editorial Team (Original artwork) via Preppsy.
