Use the GAD-7 to structure a two-week anxiety review and compare complete results over time. Do not use its total to select psychotherapy, medication, dosage, treatment intensity, taper timing, or follow-up frequency. Those decisions require diagnosis, functioning, safety, history, context, and patient preference.
Score the symptoms first. Make treatment decisions later.
Keep severity bands descriptive
Survey Doctor adds seven item values, each scored 0 to 3. The total runs from 0 to 21. The original adult primary-care study established the following descriptive thresholds.
| Score | Severity | Defensible interpretation |
|---|---|---|
| 0-4 | Minimal anxiety | Few symptoms reported in the two-week window |
| 5-9 | Mild anxiety | Mild symptom frequency |
| 10-14 | Moderate anxiety | Moderate symptom frequency |
| 15-21 | Severe anxiety | Severe symptom frequency |
The bands do not encode treatment protocols. Minimal does not mean no care is needed, and severe does not identify a specific intervention. The plain-language GAD-7 score guide explains the same ranges for patients.
The originating study enrolled 2,740 adults, with 965 completing a mental-health interview. In that sample, 10 or higher had 89% sensitivity and 82% specificity for generalized anxiety disorder (Spitzer et al., 2006). These estimates support further assessment, not a universal rule.
Keep the sample visible.
A total cannot distinguish generalized anxiety disorder from another anxiety presentation, a medical condition, a medication effect, a substance effect, or situational distress.
Add the context the score omits
Before using a result in a decision, review:
- The seven item responses, not only the total
- Functional impact at work, home, and in relationships
- Symptom course beyond the past two weeks
- Safety and other mental health symptoms
- Medical conditions, medications, and substance use
- Previous care, response, adverse effects, and access
- Patient goals and preferences
Survey Doctor shows an optional functional-difficulty question when any GAD-7 symptom is endorsed. It stores the answer separately and does not add it to the 0-to-21 total. Published scoring guidance also handles difficulty separately. Ask about functioning directly rather than assuming the total represents impairment.
The GAD-7 does not contain a suicide or self-harm item. A low total cannot substitute for a separate safety assessment when the history or presentation raises concern.
If you need help now. In the US, Call 988 or text 988. Call 911 if you are in immediate danger. Outside the US, contact your local emergency number or find support in your country.
Use item patterns as prompts, not diagnoses
Items 1 through 3 cover nervousness and worry. Items 4 and 5 cover relaxing and restlessness. Items 6 and 7 cover irritability and fear.
These groupings can focus an interview. They are not validated subscales in Survey Doctor, and they should not be turned into separate scores or disorder labels. Ask the patient whether the pattern fits before assigning meaning to it.
An unchanged total can hide movement among symptoms. For example, worry may fall while restlessness rises. Item review makes that visible and can reveal changes that a single number obscures.
Interpret repeated scores cautiously
There is no universal minimum change or retest schedule for every population and care setting. Published estimates depend on baseline severity, sample, purpose, and method. Do not label a fixed point change as response, remission, relapse, or treatment failure without a source that fits the intended use.
For a comparable series:
- Use the same seven-item form.
- Require complete native responses.
- Keep the two-week timeframe consistent.
- Record major contextual changes.
- Review functioning and patient report with the score.
Survey Doctor's native results preserve all seven items and may include the optional functional-difficulty companion. A manual total does not preserve either. Incompatible imported summaries are withheld from current bands and trends rather than treated as native GAD-7 results.
Band crossings need the same caution. A move from 10 to 9 is a one-point difference, not proof that a disorder resolved. A stable score can still matter when daily function or a specific symptom changed.
Do not derive medication decisions from a total
A GAD-7 result does not support a medication name, starting dose, dose increase, augmentation strategy, benzodiazepine recommendation, or taper plan. Those choices require diagnosis, contraindications, interactions, prior response, adverse effects, patient preference, and current clinical guidance.
The same applies to psychotherapy. A total does not identify a modality or session frequency. The questionnaire can help frame goals, but the interview determines which problems need attention.
Avoid automated rules such as “increase medication after four weeks without a four-point drop.” The score does not establish that the trial was adequate, that adherence was consistent, or that the underlying diagnosis was correct.
A practical review workflow
Start with the current answers and ask whether they match the patient's experience. Review the previous complete native result if one exists. Then discuss functioning, context, safety, and goals.
Document the total and descriptive band without converting them into a diagnosis. If the score and clinical account conflict, investigate the difference. Do not choose whichever source looks more convenient.
Use the result to support questions such as:
- Which symptom changed most?
- What was happening during this two-week period?
- What effect did the symptoms have on daily life?
- Does the patient notice the same direction as the score?
- Did anything change in health, medication use, substances, sleep, or stress?
This preserves the value of standardized measurement without giving the questionnaire authority it does not have. The practice-level outcome measurement guide explains how to keep those source and denominator differences visible in aggregate reporting.
The clinical boundary
The GAD-7 describes seven anxiety symptoms over two weeks. It is not a diagnostic test and not a treatment algorithm. Its strongest role is to make part of the clinical picture explicit, repeatable, and discussable while the clinician and patient supply everything the total leaves out.