PHQ-9 Depression Test: Questions, Scoring and Accuracy

The PHQ-9 depression test (Patient Health Questionnaire-9) is a 9-question self-report form that asks how often each symptom of depression has bothered a person over the last 2 weeks^[6]. It is the depression module of the Patient Health Questionnaire, a self-administered version of the PRIME-MD diagnostic instrument for common mental disorders^[1].

Kurt Kroenke, Robert Spitzer and Janet Williams published its validation study in 2001, based on 6,000 patients in 8 primary care clinics and 7 obstetrics-gynecology clinics^[1]. Development was funded by an educational grant from Pfizer, which holds the copyright^[1].

Each item matches one of the DSM-IV symptom criteria for major depression, so the same form can grade severity and support a criteria-based diagnosis^[1]. Clinicians use it for screening, diagnosing, monitoring and measuring the severity of depression^[7].

How common depression is in the U.S. is tracked on the major depressive disorder and depression pages. This article covers how the test itself works, what scores mean and how accurate it is.

Each item has 4 answer choices: "not at all" (0), "several days" (1), "more than half the days" (2) and "nearly every day" (3)^[6]. The item scores are added for a total from 0 to 27^[7].

Totals of 5, 10, 15 and 20 are the lower edges of mild, moderate, moderately severe and severe depression^[1]. In the original study, the positive likelihood ratio for major depression rose from 0.04 for scores of 0 to 4 to 36.8 for scores of 20 to 27^[1].

A final question asks how difficult the checked problems have made it to work, take care of things at home or get along with other people, from "not difficult at all" to "extremely difficult"^[6]. In national survey data from August 2021 to August 2023, 87.9% of people age 12 and older with a PHQ-9 score of 10 or higher reported at least some difficulty with work, home or social activities^[8].

The diagnostic algorithm

The form can also be read as a checklist instead of a sum. Under the developers' algorithm, major depression is indicated when 5 or more of the 9 symptoms are present "more than half the days" and 1 of them is depressed mood or loss of interest^[1]. Question 9 counts if it is present at all, regardless of how often^[1].

A University of Washington clinical curriculum, citing the developers, pairs each band with a proposed action: none at 0 to 4, watchful waiting and a repeat PHQ-9 at 5 to 9^[7]. Scores of 10 to 14 are paired with a treatment plan, and 20 to 27 with immediate treatment and, when impairment is severe or response is poor, referral to a mental health specialist^[7].

In the 2001 validation study, a score of 10 or higher had a sensitivity of 88% and a specificity of 88% for major depression, checked against an independent interview by a mental health professional^[1]. The area under the ROC curve was 0.95, and internal reliability (Cronbach's alpha) was 0.89 in the primary care sample^[1].

Results held on repeat testing: scores from the clinic form and a telephone PHQ-9 within 48 hours correlated at 0.84^[1].

The largest accuracy study since then is a 2019 individual participant data meta-analysis in the BMJ, which pooled raw data from 58 studies and 17,357 participants, 2,312 of whom had major depression^[2]. Against semistructured interviews, the cutoff of 10 or higher had a sensitivity of 0.88 and a specificity of 0.85^[2].

The reference interview mattered. Across cutoffs of 5 to 15, sensitivity was 5% to 22% higher against semistructured interviews than against fully structured interviews designed for lay interviewers^[2]. The PHQ-9 also looked similarly sensitive but possibly less specific in younger patients, though the authors concluded that a cutoff of 10 can be used regardless of age^[2].

The USPSTF evidence review, drawing on an individual participant dataset of 47 studies and 11,234 participants, put both sensitivity and specificity at 85% for the same cutoff^[5].

No. The authors of a 2020 meta-analysis on this question state that depression symptom questionnaires are not for diagnostic classification^[3].

Their data show why. Across 44 studies and 9,242 participants, 24.6% scored 10 or higher, while 12.1% met criteria for major depression on the Structured Clinical Interview for DSM (SCID)^[3]. A cutoff of 14 or the diagnostic algorithm came closer, but study-level estimates still missed interview-based rates by an average of 4.8 to 5.6 percentage points^[3].

The developers built in exclusions as well. Before a final diagnosis, the clinician is expected to rule out physical causes of depression, normal bereavement and a history of a manic episode^[1]. Because the form asks only about depressive symptoms, it cannot separate unipolar depression from bipolar disorder on its own; tools such as the Mood Disorder Questionnaire screen for past mania.

Middle scores are the least decisive. In the original study, scores of 10 to 14 "represented a spectrum of patients," while scores below 5 almost always meant no depressive disorder and scores of 15 or higher usually indicated major depression^[1].

Question 9 asks how often a person has been bothered by "thoughts that you would be better off dead or of hurting yourself in some way"^[6]. Under the PHQ-9 algorithm it counts toward a diagnosis if present at all, regardless of duration^[1].

A response above zero is a flag, not a verdict. Clinical guidance attached to the scoring notes states that a patient who answers yes to question 9 needs further assessment of suicide risk by someone competent to assess it^[7]. The Columbia Suicide Severity Rating Scale is one structured follow-up assessment.

Item 9 answers carry real signal. Among more than 84,000 outpatients who completed the PHQ-9 at depression visits from 2007 to 2011, one-year risk of a suicide attempt rose from 0.4% for "not at all" to 4% for "nearly every day"^[4]. One-year risk of suicide death rose from 0.03% to 0.3% across the same answers^[4].

The same figures show that most people who report frequent thoughts of death do not go on to attempt suicide, and the researchers described the higher risk as an enduring vulnerability that built over several months rather than a short-term crisis^[4]. The USPSTF found the evidence insufficient to judge screening specifically for suicide risk in adults, while noting that the PHQ-9 includes a question on suicidal ideation^[5].

Help is available around the clock: the 988 Suicide & Crisis Lifeline answers calls and texts at 988, and Crisis Text Line answers texts sent to 741741 with the word HOME.

The PHQ-2 uses only the first two items, depressed mood and loss of interest, each scored 0 to 3^[9]. In its 2003 validation in the same 6,000-patient sample, a PHQ-2 score of 3 or higher had a sensitivity of 83% and a specificity of 92% for major depression^[9].

At a lower cutoff the PHQ-2 catches more cases. In the USPSTF review, a PHQ-2 score of 2 or higher identified 91% of people with major depression, more than the PHQ-9 at its standard cutoff^[5].

For teenagers, the PHQ-A is the full PHQ modified for adolescents^[10]. One study of the PHQ-A found a sensitivity of 0.73 and a specificity of 0.94, and the USPSTF notes that the PHQ-9 is the most commonly used screening instrument in clinical practice^[10]. Teen prevalence figures are on the adolescent depression page.

Other tools fit other groups. The USPSTF lists the Edinburgh Postnatal Depression Scale for pregnant and postpartum people, the Geriatric Depression Scale for older adults and the Center for Epidemiologic Studies Depression Scale (CES-D) alongside the PHQ^[5]. The Edinburgh Postnatal Depression Scale article covers the perinatal tool in detail.

The USPSTF recommends depression screening for all adults, including pregnant and postpartum people and adults 65 and older, a grade B recommendation issued June 20, 2023^[5]. It names the PHQ "in various forms" among the commonly used adult instruments^[5].

For adolescents 12 to 18, the USPSTF issued a grade B recommendation for major depressive disorder screening on October 11, 2022^[10]. It found the evidence insufficient for children 11 and younger^[10].

The PHQ-9 is also used to track change over time. In 434 older adults treated in the IMPACT late-life depression trial, the minimal clinically important difference for an individual was 5 points on the 0-to-27 scale, and the PHQ-9 was more responsive than the SCL-20 at 3 months^[11]. The authors concluded it is a responsive and reliable measure of treatment outcomes in both individual care and research^[11].

Trial results for the treatments themselves are covered in cognitive behavioral therapy for depression and anti-depression meds.