Should You Retake the GAD-7? Using Your Anxiety Score to Tell If Treatment Is Working
Last reviewed: 09/08/2026
Reviewed by: Dr. Kiesa Kelly
A few weeks into anxiety treatment, most people hit the same wall: you cannot tell whether anything is actually changing. Some days feel easier. Some weeks feel exactly like the ones before therapy started. Memory is a poor instrument here: how you feel today quietly rewrites how you remember last month.
That is the real reason to retake the GAD-7 — not to get a better number, but to get a second point on a line. The GAD-7 anxiety screener was built to measure symptom severity, and repeated over time it turns a vague sense of "maybe better?" into a direction you can check. Anxiety is also one of the few areas where "is this working?" has a number attached: there is a published threshold for how far a score has to move before the change means anything. Below it, you cannot yet tell the change from the instrument's own margin of error.
In this article, you'll learn:
what a second GAD-7 score tells you that a first one cannot
how big a change has to be before it counts as real
when to retake the screener, and why spacing matters
what a flat score at week eight does and does not mean
what to bring to your clinician besides the number
This is not a guide to a single result — for what a given number means on its own, see our post on GAD-7 scoring. This one is about the second score, the third, and the shape they make together.
It is also not about session counts, which our post on how many CBT sessions anxiety treatment involves already covers. A session count is an average across many people; a score change is about you.
What a second GAD-7 score tells you that a first one cannot
One score is a location; two are a direction
A single result answers a narrow question: how often did seven anxiety symptoms bother you over the past two weeks. Useful the first time, and static — it cannot tell you whether you are moving or standing still.
Two scores at a sensible interval answer the question you actually care about: are things moving, in which direction, and how fast? That is not answerable from inside a hard week — which is exactly when people conclude nothing is working.
This is not an idea we invented. Clinical guidance for generalised anxiety disorder says practitioners delivering psychological treatment for GAD should use routine outcome measures and ensure the person is involved in reviewing whether it is working [1]. You are meant to be looking at your own trend, and repeating a screener is how you get one.
📈 Key takeaway: One score tells you where you are. A series of scores tells you which way you are facing.
The number was never built to carry the whole picture
Be clear what the GAD-7 is not. A 2025 Cochrane review pooled 35 studies that tested the GAD-7 against a full diagnostic interview and found that at the recommended cut-off of 10 or higher, it correctly identified about 64 percent of people who had generalised anxiety disorder; across the 22 studies that looked at any anxiety disorder, that figure was 48 percent [2]. It flagged few people who did not have the condition — specificity was about 91 percent in both analyses — and the reviewers are explicit that these summary figures are rough averages across very different populations. A useful flag, nowhere near a diagnosis.
The same guidance that asks clinicians to use outcome measures also says assessment should not rest on symptom number, severity, and duration alone, but on distress and impairment in daily life [1]. Our mental health screening tools work the same way: they organise the conversation rather than replace it [9].
Three things people get wrong about retaking the GAD-7
"If the score dropped, treatment is working." Sometimes. But a one- or two-point drop sits well inside ordinary week-to-week variation, and can be produced by nothing more than a quiet fortnight. Change has to clear a threshold before it means anything.
"If the score has not dropped, treatment is failing." This is the more damaging one, because it makes people quit at the wrong moment. A flat total can sit on top of real, uneven change, and it can also mean the plan needs a specific adjustment — information you can use, not a verdict on the whole effort.
"Any two scores show me a trend." Not if they overlap. The GAD-7 asks about the past two weeks, so scores taken days apart describe heavily overlapping stretches of your life, and the difference between them is mostly the last few days, dressed as a trend. Frequent measurement is genuinely useful when a clinician is reading it session by session; it is comparing two near-adjacent scores yourself that misleads.
🧭 Key takeaway: All three mistakes come from treating one number as a verdict instead of one point on a line.
None of this is unique to anxiety. If you have used a trauma screener, our post on retaking the PCL-5 will look familiar — the thresholds differ, reading trend over spike does not.
What counts as a real change, and what is just noise
Where the 4-point rule comes from
The threshold worth knowing is 4 points, and it comes from two different directions. The honest version includes both.
The first is a 2020 study of the GAD-7's sensitivity to change and its minimal clinically important difference. Working with 261 patients and comparing GAD-7 movement against independently rated improvement, the authors estimated that difference at 4 points on the total score [3].
The caveat is the population. Those participants came from a multisite trial of chronic depression, not a trial of anxiety treatment, and the authors said so plainly: "Confirmation of these findings and further investigation of the GAD-7 in populations and trials focusing on anxiety-specific treatment is highly recommended" [3]. The most-cited number in the field comes from a depression sample, and you deserve to know that before building a decision on it.
The second source lands closer to home. NHS Talking Therapies, the national service in England treating anxiety and depression, publishes a reliable change value for each of the symptom measures in its manual. For the GAD-7 that value is 4 or more points [4] — not a research estimate, but the working threshold applied daily across an anxiety service. Two different questions, and they land on the same number. One asks how much change patients notice as meaningful; the other asks how much change is larger than the instrument's own error. That they agree at 4 points is why the figure is worth carrying, though neither alone makes it precise.
Why a one- or two-point wobble is not progress
The idea underneath a reliable change threshold is older than the GAD-7. In 1991, Jacobson and Truax proposed a reliable change index precisely because clinicians needed a way to tell whether one person's change was larger than the measurement error built into the instrument [5]. Every questionnaire has some. A score is an estimate, not a reading off a scale.
So the practical rule: if your total moved 1, 2, or 3 points, treat it as flat — not bad news, not good news. If it moved 4 or more, in either direction, you have something worth putting in front of your clinician. That is also what makes the session-count question less urgent than it feels.
🔢 Key takeaway: Under 4 points, treat the change as noise. At 4 or more, treat it as a conversation starter — not as an instruction.
When to retake it, and why spacing matters
Let the two-week window do its job
Our GAD-7 asks how often each of seven symptoms has bothered you in the last two weeks, with four response options scored 0 to 3, for a total between 0 and 21 [6]. That frame is why spacing matters: two scores taken eight days apart share six days of the same life, so whatever moved between them is mostly a couple of days, amplified.
If your clinician asks you to complete the GAD-7 every session, do that — session-by-session measurement is standard practice in services built around these measures, and your clinician is reading it alongside everything else they see. What the two-week window does change is how you should read two scores side by side: unless they are at least two weeks apart, or anchored to a planned review point, they describe overlapping stretches of the same life rather than two different ones.
Anchor retakes to decisions, not to bad days
Re-score when something is genuinely being decided: a few weeks into a new approach, before a review appointment, after a medication change, or when circumstances have clearly shifted. Those are the points where a number can change what happens next. The least useful time is right after a hard evening, because you already know it was a hard evening.
Here is what that looks like. Imagine you start therapy with a GAD-7 of 16, and six weeks later you retake it and get 15. Your first instinct is that you have wasted six weeks. But look at what else changed: you went back to driving on the interstate, and you stopped rehearsing phone calls before making them. You are still sleeping badly and still tense most afternoons, which is most of what is holding the number up. That pattern — avoidance loosening while physical tension lags — is worth naming at your next appointment, and it is exactly the kind of change a single total is slow to show.
Now the opposite shape. You retake at week six and the score has dropped from 14 to 8, comfortably past the threshold, and it is tempting to read that as finished. But the drop landed in a stretch when work was quiet and you were sleeping well, and none of the situations you actually avoid have been tested yet. A real 6-point improvement is worth having. It is not the same as knowing it will hold when pressure returns, which is exactly what to raise next appointment. If you are weighing a structured approach like CBT for anxiety, give it enough weeks to show something before asking the questionnaire to rule on it.
⏱️ Key takeaway: Retake the screener when a decision is nearby, not when a feeling is loud.
What a flat score at week eight does and does not mean
Four things a flat total can hide
A total is a sum, and sums conceal. At least four things can hold one steady.
Item-level cancellation. Worry and restlessness may have eased while irritability and trouble relaxing got worse, landing the total in the same place. Four of the seven items moving in opposite directions is invisible in one number and obvious the moment you look at the rows.
Function moving while frequency does not. You may notice anxiety about as often and be doing far more with your week — keeping plans, going in to the office, saying yes to things you had been declining. The GAD-7 asks how often symptoms bothered you, not what you did anyway.
Timing. A rough fortnight inside an otherwise better three months will hold the score up by itself, which is why one flat retake means much less than two in a row.
The wrong ruler. This is the one almost nobody considers.
When the GAD-7 is the wrong ruler for your anxiety
The GAD-7 was developed to assess the severity of anxiety symptoms in generalised anxiety disorder [7]. It gets used far more broadly, because people with other anxiety disorders also score high on it — but NHS guidance is explicit about the cost. The GAD-7 does not cover the symptoms treatment for those conditions actually targets: avoided situations and whether you are alone or accompanied, panic attacks and the fear of having one, health anxiety, obsessions and compulsions, intrusive memories and avoidance of trauma reminders, and fear or avoidance of social situations [4].
The guidance goes further. A well-validated social anxiety measure can detect disabling social anxiety in people whose GAD-7 scores already sit near or below the caseness threshold that service uses [4]. The general number can look fine while the thing you came in for is still running your life, which is why the recommendation is to add a measure built for the specific disorder when that is the focus of treatment [4].
So if you have tracked a flat GAD-7 through eight weeks of treatment for panic, social anxiety, or health anxiety, the flat line may be telling you about the questionnaire rather than about you. Say so at your next appointment. If the plateau persists across measures and weeks, our post on what a plateau in anxiety therapy means covers what comes next.
It is also worth running a depression screener alongside. These symptoms co-occur constantly, and a GAD-7 that stalls while a PHQ-9 climbs points at a different treatment question.
🪞 Key takeaway: A flat score is a question, not an answer. Four things can hold it steady, and only one of them means treatment is not helping.
What to bring to your clinician
Four questions worth asking out loud
Numbers matter when they change a conversation. These four turn a score into a decision:
My total moved fewer than 4 points in six weeks. Is that what you would expect at this stage of this approach?
Which items moved, and does that pattern match what we have been working on?
Is the GAD-7 the right measure for my anxiety, or should we track something more specific alongside it?
If the trend is still flat at the next review, what would you change?
The last one matters most: it converts a plateau from a source of dread into a pre-agreed decision point.
A simple rule for reading your own trend
If your score moved 4 or more points down and daily life has widened alongside it, that is real movement — worth naming at your next appointment, along with what has been working. If it moved less than that but you are avoiding less, sleeping better, or doing more, treat it as early progress the total has not caught up with, and say so. If it is flat across two consecutive checkpoints and nothing in daily life has loosened, ask what would change — rather than concluding that treatment does not work for you.
Repeated measurement is not a formality. Across eleven meta-analyses, monitoring outcomes and feeding results back produced small but real average improvements, with larger effects when the data flagged cases that were not on track — though the same review is candid that it does not help everyone [8]. The benefit comes from what gets done with the number, which is why outcome tracking is built into the plan in our specialized therapy work rather than left as homework [10].
If you are unsure whether your trend means anything, you are welcome to get in touch.
🧩 Key takeaway: The score's job is to start a specific conversation. If it does not change what you and your clinician do next, it is not doing any work.
Where this leaves you
Retaking the GAD-7 is usually worth it, with two conditions: give it enough time that the two-week window measures something new, and decide in advance to read the result as a trend, not a grade.
A change of 4 or more points is worth a conversation. A smaller one is worth writing down and waiting on. A flat score is worth investigating, because four things can produce one and only one of them is bad news. And a number that contradicts your lived experience is not a reason to distrust yourself — it is a reason for a better question.
Anxiety running the show?
Evidence-based therapy can turn the volume down on anxiety — a clinician can help you find the approach that fits your life rather than a one-size-fits-all plan.
Frequently Asked Questions
Does the 4-point GAD-7 threshold come from research on anxiety treatment?
Not exactly, and it is worth knowing. The published GAD-7 minimal clinically important difference of 4 points comes from patients in a multisite chronic depression trial, and the authors specifically called for further work in anxiety-focused treatment samples. The same 4-point figure is used as the reliable change index in NHS anxiety and depression services, which is reassuring. Treat 4 points as a well-supported rule of thumb rather than a number tuned to your situation.
What should I do if my GAD-7 score goes up during anxiety treatment?
Bring it to your clinician rather than reading it as failure. A rise of 4 or more points is larger than measurement error, so it is worth explaining, but the explanation is often a specific stressor, a treatment step that temporarily raises distress, or simply a hard two weeks. Write down what was happening in that window and let your clinician weigh the number against everything else they are seeing.
Can retaking the GAD-7 too often make anxiety worse?
In our clinical experience, it can for some people. The questionnaire itself is not harmful, but frequent score-checking can slide into reassurance-seeking, which tends to feed anxiety rather than settle it. If you notice an urge to re-score every time you feel on edge, that is worth naming out loud to your clinician. Agreeing a schedule with your clinician keeps the screener working as a tracking tool instead of becoming part of the worry loop.
Should I track the PHQ-9 alongside the GAD-7?
Often, yes. Anxiety and depression symptoms overlap heavily, and a GAD-7 that stalls while low mood, hopelessness, or loss of interest keeps climbing points to a different treatment question than the one you have been working on. Running a depression screener alongside the anxiety one gives you two trends instead of one. Your clinician can help you decide which measures actually fit your plan.
Can I compare a GAD-7 I took online with one my clinician gave me?
Usually yes, as long as it is the same seven items, the same two-week instruction, and the same four response options. Comparability depends on holding the frame of reference constant. If one version was shortened, reworded, or asked about a different span of time, the two scores are answering slightly different questions, and the gap between them is not a clean trend you can read.
About ScienceWorks
ScienceWorks Behavioral Healthcare is a telehealth-forward practice serving Tennessee, with an in-person option at our Nashville office. Dr. Kiesa Kelly, our owner and licensed clinical psychologist, holds a PhD in Clinical Psychology with a concentration in Neuropsychology from Rosalind Franklin University of Medicine and Science, and completed practica, internship, and an NIH-funded postdoctoral fellowship across medical and academic settings, with more than 20 years of experience in psychological assessment and evidence-based treatment.
Our clinical team works with anxiety and panic, depression, OCD, trauma, insomnia, and ADHD and autism in adults and adolescents, using structured approaches including CBT, ACT, ERP, I-CBT, EMDR, and CBT-I. Outcome measures like the GAD-7 are part of how we and the people we work with check whether a plan is doing what it is supposed to do. Every article we publish is reviewed by a licensed clinician for accuracy before it goes live.
References
1. National Institute for Health and Care Excellence. Generalised anxiety disorder and panic disorder in adults: management. NICE clinical guideline CG113. 2011, last updated 2020. Available from: <https://www.nice.org.uk/guidance/cg113>
2. Aktürk Z, Hapfelmeier A, Fomenko A, Dulek D, Alnasser M, Linde K, et al. Generalized Anxiety Disorder 7-item (GAD-7) and 2-item (GAD-2) scales for detecting anxiety disorders in adults. Cochrane Database Syst Rev. 2025;3(3):CD015455. Available from: <https://pmc.ncbi.nlm.nih.gov/articles/PMC11934853/>
3. Toussaint A, Hüsing P, Gumz A, Wingenfeld K, Härter M, Schramm E, et al. Sensitivity to change and minimal clinically important difference of the 7-item Generalized Anxiety Disorder Questionnaire (GAD-7). J Affect Disord. 2020;265:395-401. Available from: <https://pubmed.ncbi.nlm.nih.gov/32090765/>
4. NHS England. NHS Talking Therapies for anxiety and depression Manual, version 7.1. Table 9: clinical cut-offs and reliable change index (p. 46-47); section 6.2.1, anxiety disorder specific measures (p. 47). Available from: <https://www.england.nhs.uk/wp-content/uploads/2018/06/nhs-talking-therapies-manual-v7.1-updated.pdf>
5. Jacobson NS, Truax P. Clinical significance: a statistical approach to defining meaningful change in psychotherapy research. J Consult Clin Psychol. 1991;59(1):12-19. Available from: <https://pubmed.ncbi.nlm.nih.gov/2002127/>
6. ScienceWorks Behavioral Healthcare. GAD-7 anxiety screener. Available from: <https://www.scienceworkshealth.com/gad-7>
7. Spitzer RL, Kroenke K, Williams JBW, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. 2006;166(10):1092-1097. Available from: <https://pubmed.ncbi.nlm.nih.gov/16717171/>
8. Barkham M, De Jong K, Delgadillo J, Lutz W. Routine outcome monitoring (ROM) and feedback: research review and recommendations. Psychother Res. 2023;33(7):841-855. Available from: <https://doi.org/10.1080/10503307.2023.2181114>
9. ScienceWorks Behavioral Healthcare. Mental health screening. Available from: <https://www.scienceworkshealth.com/mental-health-screening>
10. ScienceWorks Behavioral Healthcare. Specialized therapy. Available from: <https://www.scienceworkshealth.com/specialized-therapy>
Disclaimer
This article is for informational purposes only and is not a diagnosis, a treatment recommendation, or a substitute for individualized mental health care. A screener score is not a clinical decision — do not start, stop, or change treatment based on a number without talking to a qualified professional who can consider your full clinical picture. If you are in immediate danger or think you may harm yourself or someone else, call 911 or go to the nearest emergency room right away.