PHQ-9 vs GAD-7: how the depression and anxiety scores compare

What the PHQ-9 and GAD-7 each measure

The PHQ-9 and the GAD-7 are sister questionnaires from the Patient Health Questionnaire family, developed by Spitzer, Williams, Kroenke and colleagues with an educational grant from Pfizer, but they measure different things. The PHQ-9 grades depressive symptoms through nine items that follow the nine DSM-IV criteria for major depression, from loss of interest and low mood to sleep, energy, appetite, self-worth, concentration, psychomotor change and thoughts of self-harm. The GAD-7 was developed to identify probable generalized anxiety disorder and grade its severity, through seven items on nervousness, uncontrollable and excessive worry, trouble relaxing, restlessness, irritability and fear that something awful might happen. Both ask about the last 2 weeks and use the same four answers, scored 0 to 3: not at all, several days, more than half the days and nearly every day. The PHQ-9 total runs from 0 to 27 and the GAD-7 total from 0 to 21. Only the PHQ-9 asks about self-harm.

How to score the GAD-7 and read the result

To score the GAD-7, add its seven answers, each worth 0 for not at all, 1 for several days, 2 for more than half the days and 3 for nearly every day, for a total from 0 to 21. Spitzer and colleagues, in the Archives of Internal Medicine in 2006, proposed that scores of 5, 10 and 15 might be read as mild, moderate and severe anxiety, and the PHQ instruction manual repeats those cut points. A score of 10 or more was the point that best balanced sensitivity (89%) and specificity (82%) for generalized anxiety disorder, against a telephone interview by a mental health professional in 965 patients. A score of 15 or more had a sensitivity of 48% and a specificity of 95%. Rising GAD-7 scores were strongly linked with poorer function on all six scales of the SF-20 health survey and with more disability days. Unlike the PHQ-9, the GAD-7 has no moderately severe band, so 15 to 21 is one severe band.

The seven GAD-7 items, each asked about the last 2 weeks:

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

The GAD-7 is not limited to generalized anxiety disorder. In the same 965 primary care patients, Kroenke and colleagues (Annals of Internal Medicine, 2007) found that 19.5% had at least one anxiety disorder, and that the GAD-7 performed well as a screen for generalized anxiety disorder, panic disorder, social anxiety disorder and post-traumatic stress disorder, with areas under the ROC curve of 0.80 to 0.91 for the GAD-7 and its two-item version.

PHQ-9 vs GAD-7 accuracy at a cut-off of 10

Both questionnaires use 10 as their main screening cut-off, and at that point their accuracy against a mental health professional interview is similar. In Kroenke 2001, a PHQ-9 of 10 or more had a sensitivity of 88% and a specificity of 88% for major depression in 580 interviewed patients. In Spitzer 2006, a GAD-7 of 10 or more had a sensitivity of 89% and a specificity of 82% for generalized anxiety disorder in 965 interviewed patients. The PHQ-9 cut-off has since been tested far more widely: the Levis 2019 individual participant data meta-analysis in the BMJ pooled 58 studies and 17,357 participants and found that 10 or more again gave the best combined sensitivity and specificity, at 88% and 85% against semistructured interviews. Raising either cut-off to 15 trades sensitivity for specificity: 68% and 95% for the PHQ-9, and 48% and 95% for the GAD-7. Both remain screening and severity tools, and neither total makes a diagnosis.

Why use the PHQ-9 and GAD-7 together

Depression and anxiety often appear in the same patient, which is why the PHQ-9 and GAD-7 are often given together. In the 2006 GAD-7 study, the GAD-7 total correlated strongly with the PHQ-8 depression score (r = 0.75), yet factor analysis confirmed anxiety and depression symptoms as distinct dimensions, and each had its own independent effect on functional impairment and disability. A high score on one questionnaire therefore does not answer the question the other asks. The PHQ instruction manual uses the same thresholds across the family: 5, 10 and 15 mark mild, moderate and severe symptoms on the PHQ-9, GAD-7 and PHQ-15, a score of 10 or more is a yellow flag that draws attention to a possible clinically significant condition, and 15 or more is a red flag, where active treatment is probably warranted. For a quick first pass, the manual says a score of 3 or more on the two-item PHQ-2 or GAD-2 should prompt the full PHQ-9 or GAD-7 and a clinical interview.

The one item with no counterpart is PHQ-9 item 9, on thoughts of being better off dead or of self-harm. Any answer above 0 needs a prompt clinical assessment of self-harm risk, whatever the PHQ-9 or GAD-7 total, and the manual says a final decision about that risk requires a clinical interview. The PHQ-9 calculator flags it on its own line.

Worked example: one patient, both questionnaires

Take the worked example from the PHQ instruction manual, a 43-year-old woman who looks sad and has had fatigue for a month, and add a hypothetical GAD-7. On the PHQ-9 she answers 3, 1, 2, 3, 1, 2, 3, 0 and 1 to items 1 to 9: three items at 1, two at 2 and three at 3, for a total of 16, moderately severe. Her item 9 answer of 1, several days, needs a prompt clinical assessment of self-harm risk whatever her total; in the manual, the physician's questioning found no significant suicidal potential. Suppose her GAD-7 answers are 2, 1, 1, 2, 0, 1 and 1. They add to 8, in the mild band and below the cut-off of 10 for probable generalized anxiety disorder. Read together, the two scores point to depressive symptoms as the main problem, with milder anxiety alongside. A GAD-7 alone would not have shown her self-harm answer, because it has no such item.

Enter her PHQ-9 answers in the PHQ-9 calculator to see the total, the severity band and the item 9 flag; they are the answers it loads by default.