How doctors screen for depression and anxiety: the PHQ-9 and GAD-7, explained

What these two questionnaires are

If you are not safe right now, use the crisis numbers above before you read any further. Nothing below is urgent by comparison.

In the early 1990s Robert L. Spitzer, Janet B.W. Williams and Kurt Kroenke built a diagnostic instrument for primary care called PRIME-MD, with an educational grant from Pfizer. It was accurate and it was slow. The self-administered short version of it became the Patient Health Questionnaire. The PHQ-9 is its depression module: nine questions, one for each of the nine criteria a clinician would otherwise ask about. The GAD-7 came later, seven questions built on the same response set, originally to identify generalized anxiety disorder and now used more widely than that.

Both ask the same opening question — how often, over the last two weeks, you have been bothered by each problem — and both offer the same four answers: Not at all, Several days, More than half the days, and Nearly every day. That is the entire design. There is no trick question and no hidden item.

They are also, unusually for clinical instruments, free. The authors' own distribution site states that all PHQ and GAD-7 screeners and translations are downloadable from that website and no permission is required to reproduce, translate, display or distribute them, and each form repeats the same grant in its footer. That is why they are printed below in full, and why almost every clinic, app and website you have seen them on could print them too.

What follows is a reference copy. Nothing on this page is scored, and nothing on this page is about you.

The PHQ-9, printed in full

The instruction at the top of the form reads: Over the last 2 weeks, how often have you been bothered by any of the following problems? The four answer columns are Not at all, Several days, More than half the days, and Nearly every day.

1. Little interest or pleasure in doing things

2. Feeling down, depressed, or hopeless

3. Trouble falling or staying asleep, or sleeping too much

4. Feeling tired or having little energy

5. Poor appetite or overeating

6. Feeling bad about yourself — or that you are a failure or have let yourself or your family down

7. Trouble concentrating on things, such as reading the newspaper or watching television

8. Moving or speaking so slowly that other people could have noticed? Or the opposite — being so fidgety or restless that you have been moving around a lot more than usual

9. Thoughts that you would be better off dead or of hurting yourself in some way

Below the nine, the form asks one more question that is deliberately not part of the score: if you checked off any problems, how difficult have these problems made it for you to do your work, take care of things at home, or get along with other people? The four answers are Not difficult at all, Somewhat difficult, Very difficult and Extremely difficult. That single item is not part of any PHQ score; its authors note it is strongly associated with symptom severity and with impairment, which is why clinicians ask about it.

Reproduced from the PHQ-9 exactly as its authors distribute it. The form's own footer reads: Developed by Drs. Robert L. Spitzer, Janet B.W. Williams, Kurt Kroenke and colleagues, with an educational grant from Pfizer Inc. No permission required to reproduce, translate, display or distribute. No changes were made to the wording of the nine items.

The GAD-7, printed in full

The instruction at the top of this form reads: Over the last 2 weeks, how often have you been bothered by the following problems? The same four answer columns apply.

1. Feeling nervous, anxious or on edge

2. Not being able to stop or control worrying

3. Worrying too much about different things

4. Trouble relaxing

5. Being so restless that it is hard to sit still

6. Becoming easily annoyed or irritable

7. Feeling afraid as if something awful might happen

Reproduced from the GAD-7 exactly as its authors distribute it. The form's own footer reads: Developed by Drs. Robert L. Spitzer, Janet B.W. Williams, Kurt Kroenke and colleagues, with an educational grant from Pfizer Inc. No permission required to reproduce, translate, display or distribute. No changes were made to the wording of the seven items.

Print this page, or copy the two lists onto a sheet of paper, and answer them in your own time before an appointment. A form filled in calmly at your kitchen table is a better account of your fortnight than one filled in on a clipboard in a waiting room.

How the authors read the numbers, and why this page shows none

The people who built these scales published how they read them, and what follows is their reading, not ours.

For the PHQ-9, each answer is worth 0 to 3 and the nine items give a range of 0 to 27. In the validation paper, PHQ-9 scores of 5, 10, 15, and 20 represented mild, moderate, moderately severe, and severe depression, respectively — the authors read 5, 10, 15 and 20 as the thresholds of mild, moderate, moderately severe and severe. For the GAD-7 the seven items give a range of 0 to 21, and the authors' instruction manual reads 5, 10 and 15 as the cutpoints for mild, moderate and severe anxiety, and describes a score of 10 or greater as a recommended point at which further evaluation is warranted.

Those bands belong to the instruments and are printed here because withholding them while printing the items would be the worse choice: you would look them up anyway, somewhere with no crisis note on the page. They are a description of how a clinician reads a form. They are not a verdict, they are not a diagnosis, and they are not a statement about you.

MyTestAtlas does not add these answers up for you, and the reason is the ninth question. Item 9 of the PHQ-9 asks about thoughts that you would be better off dead or of hurting yourself in some way. A clinician asks that question because they are in the room, can ask the next question, and can act on the answer within the hour. The authors' own manual says the same thing in its own terms: a final decision about the actual risk of self-harm requires a clinical interview. A web page can do none of that. It cannot see your face, cannot ask what you meant, and cannot keep you safe. Scoring that item behind a Calculate button would be pretending otherwise, so this site does not score the PHQ-9 at all.

There is a version without it. The PHQ-8 is all the items of the PHQ-9 except the ninth, and its authors describe it as mainly used in non-depression research studies. That is a study design decision, not a safer consumer product, and it does not change what this site does: no scored questionnaire on either count.

If you answered anything other than Not at all to item 9 while reading it, treat that as the reason to contact someone today rather than as a number to work out. The crisis numbers at the top of this page are the shortest route.

What happens after a screen comes back high

A high score does not start a diagnosis; it starts a conversation, and the conversation is longer than the form. In England the NHS describes the first step plainly: if you see a GP about symptoms of generalised anxiety disorder (GAD), they'll ask you some questions about your worries and physical symptoms. Routes differ by country, but the content of the appointment does not vary much.

• Time and duration. How long this has run, what was happening when it started, and whether anything like it has happened before.
• The physical possibilities. Blood tests are common — thyroid function above all, and often iron and vitamin levels — because an underactive thyroid can produce many of the same answers in a person who is not depressed.
• Sleep, alcohol, other drugs, and every prescribed medicine, including anything started in the weeks before the change.
• Whether there have ever been stretches of unusually high energy and low sleep need. This is asked because it changes which treatments are safe, and it is the question most often skipped when someone arrives during a low period.
• Safety, asked directly and out loud. This is the part the form cannot do, and it is the reason the form exists in a room with a person in it.
• What you want to change. Talking therapy, medication, or both; who refers; how long the wait is; and a date to be seen again.

Then the same questionnaire often comes back, weeks later, and is compared with the first one. That is the use it was designed for and the use it is best at: not labelling a person, but showing whether a chosen treatment is moving the number. A single score is a baseline. Two scores are information.

You can also start this yourself. In England anyone over 18 can refer themselves to an NHS talking therapies service without going through a GP first.

What to bring, and where this site's self-checks fit

Appointments are short and adjectives are hard to interpret. Bring specifics.

• Two weeks of one-line notes: mood, energy, and hours slept.
• Two or three concrete things that have changed, with dates — work missed, a plan called off, a reply not sent.
• The two lists above, filled in at home, so the ten minutes are spent on what they mean rather than on filling them in.
• Every medicine, supplement and substance you take, and any recent change to them.
• The one thing you are most worried about, written down before you go, so it survives a short appointment.

MyTestAtlas has two self-checks in this category. The Low Mood and Worry Self-Check reports five patterns from the last two weeks — interest and pleasure, energy and sleep, worry that will not switch off, body tension and restlessness, and how it costs the day. The Social Situations Self-Check reports four patterns about social life: the build-up before, what happens during, avoidance and escape, and the replay afterwards. Neither has a cut-off, a total or a verdict, and neither contains a question about self-harm, because the reasoning above applies to our own items as much as to anyone else's.

What they are for is language. People often arrive at an appointment able to say only that they have not been themselves. Leaving instead with named patterns and dated examples turns that into something a clinician can work with. That is the whole of the claim, and this page has printed the actual instruments so you can see the difference between the two things.

And if you are not safe, the crisis numbers at the top of this page matter more than anything else on it.