The GAD-7: Tracking Anxiety in Therapy
The GAD-7 is the most commonly used anxiety screening instrument in clinical practice. Seven items, two minutes, and a score that tells you more about your client's anxiety than an hour of session observation might.
But like any clinical tool, its value depends entirely on how you use it. A one-time administration gives you a snapshot. Repeated administrations over time give you something much more powerful: a trend.
This guide covers how the GAD-7 works, what the scores actually mean, how to track changes over time, and where the instrument fits alongside other assessments in your clinical workflow.
What the GAD-7 measures
The GAD-7 is a seven-item self-report questionnaire designed to screen for generalized anxiety disorder and measure anxiety severity. Each item asks how often the respondent has been bothered by a specific anxiety symptom over the past two weeks.
The seven items cover: feeling nervous or anxious, inability to stop or control worrying, worrying too much about different things, trouble relaxing, restlessness, irritability, and feeling afraid that something awful might happen.
Each item is scored from 0 ("not at all") to 3 ("nearly every day"). Total scores range from 0 to 21.
The instrument was developed by Spitzer, Kroenke, Williams, and Lowe and validated in primary care settings, but it's widely used in mental health practice. It's brief enough to administer frequently without burdening clients, and sensitive enough to detect clinically meaningful changes.
How to score it
Add the responses. Seven items, each scored 0 to 3. The total falls into one of four severity bands.
0–4: Minimal anxiety. The client isn't endorsing significant anxiety symptoms. This is your baseline for clients without anxiety as a presenting concern.
5–9: Mild anxiety. Some anxiety symptoms are present but not at a level that typically requires anxiety-specific treatment. Worth monitoring — particularly if the client has other presenting concerns that could be masking or exacerbating anxiety.
10–14: Moderate anxiety. This is the clinical threshold. A score of 10 or above has good sensitivity and specificity for generalized anxiety disorder, panic disorder, social anxiety disorder, and PTSD. Treatment planning should explicitly address anxiety at this level.
15–21: Severe anxiety. The client is endorsing most symptoms at high frequency. Active treatment is strongly indicated. Consider whether the current approach is adequate, whether referral for medication evaluation is appropriate, and whether the anxiety is impacting the client's ability to engage in therapy itself.
The GAD-7 and the PHQ-9: a natural pair
Anxiety and depression are the two most common presenting concerns in outpatient therapy, and they co-occur more often than not. The GAD-7 and PHQ-9 were designed as companion instruments — same response format, same two-week timeframe, same scoring structure.
Administering both gives you a two-dimensional picture of your client's mood. Depression improving but anxiety stable? That's clinically meaningful. Both worsening? Different story. Both improving? Now you have something to show the client.
The parallel structure also makes it easy for clients. Once they've completed one, the other feels familiar. Compliance stays high because the total time for both instruments is under five minutes.
Tracking changes over time
A single GAD-7 score tells you how anxious a client is right now. A series of scores tells you whether treatment is working.
What constitutes a meaningful change? A shift of 4 or more points is generally considered a reliable change — meaning it's unlikely to be measurement noise. A shift of 5 or more points is often cited as a clinically significant change. Crossing a severity band boundary (e.g., moving from moderate to mild) is always worth noting.
What patterns to watch for:
Steady decline is the ideal — scores dropping consistently over time as treatment takes effect. This is what you show clients when they say "I don't feel like I'm getting better." The data often tells a different story than their in-session self-report.
A plateau after initial improvement is common. The client got better quickly, then leveled off. This might mean the current approach has reached its ceiling for this client, or it might mean the remaining anxiety is situational and not amenable to the same intervention.
A spike after a period of stability usually points to a triggering event — a life stressor, a change in circumstances, sometimes a therapeutic deepening that temporarily increases anxiety. The trend before and after the spike matters more than the spike itself.
Gradual worsening is the pattern therapists most often miss without systematic tracking. A client who moves from a 7 to an 8 to a 9 to an 11 over two months doesn't set off alarms in any individual session. But the trend is clear when you see the data.
Beyond generalized anxiety
The GAD-7 was designed for generalized anxiety disorder, but research has shown it performs reasonably well as a screener for other anxiety-related conditions. Scores of 10 or above have acceptable sensitivity for panic disorder, social anxiety disorder, and PTSD.
That said, the GAD-7 is not a replacement for disorder-specific instruments. If a client screens positive on the GAD-7 and you suspect PTSD, the PCL-5 gives you much more detailed information about trauma-specific symptoms. If social anxiety is the primary concern, the Liebowitz Social Anxiety Scale or the SPIN provides better specificity.
Think of the GAD-7 as a first-line screen: it tells you whether anxiety is present and how severe it is. Disorder-specific instruments tell you what kind of anxiety it is and how to target treatment.
Common clinical considerations
Clients who underreport. Some clients minimize symptoms on self-report measures. If your clinical observation suggests more anxiety than the score reflects, trust your observation — but don't discard the score. Note the discrepancy. It may resolve over time as the therapeutic relationship deepens and the client becomes more comfortable reporting accurately.
Clients who overreport. Anxiety about completing an anxiety questionnaire is real. Some clients score higher on their first administration than on subsequent ones, not because their anxiety improved, but because the novelty of the measure inflated their responses. The second and third administrations are often more reliable than the first.
Somatic anxiety. The GAD-7 focuses primarily on cognitive and emotional symptoms of anxiety. Clients whose anxiety presents primarily as somatic symptoms (muscle tension, GI distress, chest tightness) may not score as high as their functional impairment would suggest. Consider supplementing with a measure that captures somatic anxiety if this is relevant to your client.
Medication effects. If a client starts or changes anxiety medication between administrations, factor that into your interpretation. A drop in GAD-7 scores after starting an SSRI might reflect medication response rather than therapy response — or both. The data doesn't tell you the cause, only the trajectory.
Integrating the GAD-7 into your workflow
The therapists who get the most value from the GAD-7 aren't the ones who administer it — they're the ones who review it before every session.
A two-minute review of your client's current GAD-7 score, trend over the past few months, and any changes since last session gives you a clinical lens to carry into the room. You're not starting from scratch every session. You're starting from data.
When a client says "I don't know, I think I'm the same," you can say "Actually, your anxiety scores have dropped from 15 to 9 over the past six weeks. That's a meaningful change." That's not just informative — it's therapeutic.
And when the data shows something the client hasn't noticed — a gradual worsening, a new spike, a plateau — you can address it proactively rather than reactively.
That's what outcome tracking is for: not to reduce therapy to a number, but to make sure the numbers don't get lost in the noise of session-to-session work.
Theracharts includes the GAD-7, PHQ-9, and 80+ other validated assessments — auto-scored with trend charts and clinical alerts. Get started free.
