GAD-7 Administration Guide for Therapists: Scoring and Follow-Up

Authored by Dr. Hannah Lin, counseling psychologist trained in CBT, ACT, and IFS, with over a decade of clinical practice across anxiety and complex trauma.

A client describes weeks of free-floating worry, restless sleep, and a short fuse at home. You already hear generalized anxiety in the room. What you still need is a repeatable number you can place next to last month’s number, without turning the hour into a testing battery.

That is the clinical job of the GAD-7. It is a seven-item self-report screen for anxiety severity that takes about one minute, scores from 0 to 21, and supports measurement-based care when you administer it on a stable cadence and review the trend with the client. It does not replace formulation, differential diagnosis, or the judgment call about which anxiety cycle is maintaining the problem this week.

This guide is for licensed therapists using the measure in outpatient therapy. It covers when to use it, how to administer and score it, how to document and review change, and how to turn a flat or rising trajectory into a concrete follow-up move. For the broader MBC frame, start with the measurement-based care practical guide for therapists. For plan writing after the score is in, use the anxiety treatment plan template.

Educational content for licensed therapists, not clinical or legal advice. Screening is not diagnosis. Adapt cadence, framing, and follow-up to presentation, risk, culture, setting, and supervision. Obtain the free clinical instrument from a legitimate source and follow its terms of use.

When therapists reach for this anxiety screen

Use this brief screen when anxiety is a primary concern, a comorbid track you are actively treating, or a severity signal you want on a stable schedule. Common fits:

• Intake and diagnostic evaluation when generalized worry, tension, or mixed anxiety is on the table
• Early treatment to set a baseline before exposure, worry work, or medication coordination
• Every 2 to 4 sessions during active anxiety-focused care
• Planned reviews, step-up or step-down decisions, and pre-discharge checks
• Cases where the client and you disagree about whether things are moving

It is a weaker first pick when the presentation is clearly panic-only, phobia-only, or trauma-primary and you need a measure that maps those symptom clusters more tightly. Even then, many clinicians still keep a brief overall anxiety load score while using a second tool for the focal cycle.

Do not use a single elevated total as proof of generalized anxiety disorder. Duration, excessiveness, difficulty controlling worry, impairment, substance and medical differentials, and the client’s story still decide the diagnosis. The score is evidence inside that process.

How to administer without turning the session into a test

Framing language that protects the alliance

Introduce the measure as a shared view of severity over time, not as a grade:

“These seven questions take about a minute. They help us both see whether the anxiety load is moving, stuck, or rising. There is no pass or fail. We will look at the number together and decide what it means for this week’s plan.”

If a client bristles at forms, name the purpose once, keep the measure short, and always close the loop in session. A completed scale that never gets discussed teaches clients the form is paperwork, not care. That same loop supports client engagement between sessions when scores arrive before the visit and open the hour.

Administration practicalities

• Mode: paper, secure portal, or in-room tablet. Match whatever your chart can store reliably.
• Window: the standard stem asks about the last two weeks. Say that aloud so clients do not answer only about today.
• Timing: many practices collect it in the waiting room or the first three minutes. Others send it 24 hours before telehealth. Pick one workflow and keep it.
• Assistance: read items aloud when literacy, vision, language, or cognitive load requires it. Note assisted administration in the chart.
• Safety: item content can raise distress. Have your usual risk pathway ready. A high score plus hopelessness or suicidal ideation is a clinical event, not a scoring footnote.

Pairing with other measures

Pairing this anxiety screen with the PHQ-9 is common in outpatient care because anxiety and depression travel together. Keep each measure’s job clear: PHQ-9 for depressive severity and item-9 risk flag; the anxiety total for severity and trend. Do not average them into one vague “symptoms” number.

Scoring and interpretation boundaries

Each of the seven items is rated 0 (not at all), 1 (several days), 2 (more than half the days), or 3 (nearly every day). Sum the seven items for a total from 0 to 21. Higher totals indicate greater anxiety severity on the scale.

Severity bands clinicians use in conversation

Total score	Common severity band	How to use it clinically
0–4	Minimal	Monitor; confirm the client’s language still matches low load
5–9	Mild	Track trend; keep interventions specific to the active cycle
10–14	Moderate	Prioritize active treatment targets; recheck cadence
15–21	Severe	Intensify formulation review, risk check, and session focus

Original validation work on the GAD-7 supported a screening cut point around 10 for identifying likely generalized anxiety cases in primary care, with solid sensitivity and specificity in that setting (Spitzer et al., 2006). Later work has also discussed nearby cut points in other samples. For therapy practice, treat ≥10 as a prompt for fuller clinical assessment, not as an automatic GAD label.

Many versions also include a non-scored functional-impact question (how difficult these problems have made work, home, or getting along with others). Capture that rating when available. A moderate total with severe functional impact is a different clinical picture from a moderate total with minimal day-to-day interference.

Interpretation guardrails

• Screen ≠ diagnosis. A total above threshold supports interview; it does not finish differential.
• Band ≠ protocol. Moderate does not automatically mean “start exposure this session.” Match technique to the maintaining cycle.
• Single score ≠ story. Session context, avoidance, sleep, substances, medical issues, and alliance still matter.
• Cross-disorder sensitivity. Elevated scores also appear in panic, social anxiety, PTSD, and other presentations. Read the total as anxiety load, then formulate.
• Change over time is the clinical gold. A drop from 16 to 11 may matter more than whether 11 sits in the moderate band.

Documenting scores and reviewing trends

What belongs in the note

A defensible entry is short and specific:

• Measure name and date
• Total score and severity band
• Functional-impact rating if collected
• Any clinically relevant item pattern (for example, persistent restlessness and irritability with lower worry items)
• Whether you reviewed the score with the client
• Clinical implication for this visit and the next review point

Weak documentation: “Anxiety measure completed.” Stronger: “Anxiety screen total = 14 (moderate) on 2026-07-14; impact = very difficult at work. Reviewed graph with client. Worry postponement continued; next readministration session 8 or sooner if panic spikes return.”

Cadence that builds a usable trend

Phase	Suggested cadence	Review move
Intake	Baseline once	Set initial band and targets
Active anxiety treatment	Every 2–4 sessions	Compare to baseline and last score
Plateau or deterioration signal	Next session	Formulation and plan check
Maintenance / step-down	Every 4–6 sessions	Confirm stability before spacing
Pre-discharge	Final administration	Document end-state and relapse signs

Plot totals across sessions on a simple line. Clients often respond to the shape of the line more than to a single band label. When the line is flat or rising, say so directly and invite the client into the problem-solving, rather than burying the number in the chart.

What a trend is allowed to change

A meaningful shift in trajectory can justify:

• Tightening or loosening session focus
• Switching from pure cognitive work to exposure or uncertainty experiments
• Addressing therapy-interfering behavior or homework barriers
• Coordinating with prescribers using shared severity language
• Extending care with a clearer medical-necessity narrative

A trend should not, by itself, force a new diagnosis code or override risk assessment.

Worked follow-up example

Client: “M,” adult outpatient, generalized worry with sleep disruption and work avoidance.
Baseline total (session 1): 16 (severe). Functional impact: very difficult.
Sessions 2–5: worry postponement, values-based scheduling, and reduced evening reassurance texts.
Session 6 total: 15 (still severe). Client says nights feel “a little quieter,” but the number barely moved.

Clinical read: subjective relief without score movement is a signal to test whether the active cycle is still untreated. In session, M reveals the main daytime behavior is repeated mental review of emails “so nothing bad happens.” Worry postponement at night never touched the daytime control strategy.

Follow-up moves tied to the score:

1. Name the flat trajectory out loud and show the two points on paper.
2. Reformulate the maintaining cycle as intolerance of uncertainty plus daytime cognitive avoidance.
3. Add one uncertainty experiment: send one short work email without pre-reading it four times; log predicted vs actual outcome.
4. Keep the same anxiety screen on a two-session cadence for the next month.
5. Document: “Anxiety total 16→15 over five sessions; plan adjusted toward daytime uncertainty experiment; recheck session 8.”

If session 8 is 11 with clearer work functioning, you have convergent evidence. If it is still 15 with new panic symptoms, widen the assessment rather than repeating the same homework louder.

Downloadable administration pack

The shared PHQ-9 and GAD-7 administration pack is the artifact for this page. It is built for the chart desk, not as a substitute for the copyrighted item stems:

• Pre-session administration checklist and client-framing lines
• Score log with date, total, band, functional impact, and review checkbox
• Multi-session trend grid for side-by-side anxiety (and optional depression) totals
• Follow-up decision prompts for improvement, flat line, and deterioration
• Note language stems that keep screening language out of diagnosis fields

Use the official free clinical instrument for the seven items themselves. Use the pack to administer on cadence, score consistently, and turn the number into a next step.

Download the PHQ-9 and anxiety-screen administration pack (PDF)

Where Emosapien fits

Anxiety measurement only helps when the score, the session focus, and the next experiment stay connected in the record. Emosapien helps therapists keep that thread visible across visits: scores and plan language stay available when you draft the note, and you remain responsible for diagnosis, formulation, and sign-off.

Start your journey with Emosapien and keep anxiety-screen trends tied to the clinical moves you actually make between sessions.

References

1. Spitzer, R. L., Kroenke, K., Williams, J. B. W., & Löwe, B. (2006). A brief measure for assessing generalized anxiety disorder: The GAD-7. Archives of Internal Medicine, 166(10), 1092–1097. Spitzer et al. DOI record
2. Kroenke, K., Spitzer, R. L., Williams, J. B. W., Monahan, P. O., & Löwe, B. (2007). Anxiety disorders in primary care: Prevalence, impairment, comorbidity, and detection. Annals of Internal Medicine, 146(5), 317–325. Kroenke et al. DOI record
3. Plummer, F., Manea, L., Trepel, D., & McMillan, D. (2016). Screening for anxiety disorders with the GAD-7 and GAD-2: A systematic review and diagnostic metaanalysis. General Hospital Psychiatry, 39, 24–31. Plummer et al. DOI record
4. National Institute for Health and Care Excellence. CG113: Generalized anxiety disorder and panic disorder in adults: management.