Cited by 306Psychometric Methodologies and TestingAnxiety, Depression, Psychometrics, Treatment, Cognitive ProcessesMental Health Treatment and Access

Measurement invariance of the patient health questionnaire‐9 (PHQ‐9) depression screener in U.S. adults across sex, race/ethnicity, and education level: NHANES 2005–2016

Open access
Patel, Jay S., Oh, Youngha, Rand, Kevin L., Wu, Wei, Cyders, Melissa A., Kroenke, Kurt, Stewart, Jesse C.·2019·Depression and Anxiety

Abstract

BACKGROUND: Despite its popularity, little is known about the measurement invariance of the Patient Health Questionnaire-9 (PHQ-9) across U.S. sociodemographic groups. Use of a screener shown not to possess measurement invariance could result in under/over-detection of depression, potentially exacerbating sociodemographic disparities in depression. Therefore, we assessed the factor structure and measurement invariance of the PHQ-9 across major U.S. sociodemographic groups. METHODS: U.S. population representative data came from the 2005-2016 National Health and Nutrition Examination Survey (NHANES) cohorts. We conducted a measurement invariance analysis of 31,366 respondents across sociodemographic factors of sex, race/ethnicity, and education level. RESULTS: Considering results of single-group confirmatory factor analyses (CFAs), depression theory, and research utility, we justify a two-factor structure for the PHQ-9 consisting of a cognitive/affective factor and a somatic factor (RMSEA = 0.034, TLI = 0.985, CFI = 0.989). On the basis of multiple-group CFAs testing configural, scalar, and strict factorial invariance, we determined that invariance held for sex, race/ethnicity, and education level groups, as all models demonstrated close model fit (RMSEA = 0.025-0.025, TLI = 0.985-0.992, CFI = 0.986-0.991). Finally, for all steps ΔCFI was <-0.004, and ΔRMSEA was <0.01. CONCLUSIONS: We demonstrate that the PHQ-9 is acceptable to use in major U.S. sociodemographic groups and allows for meaningful comparisons in total, cognitive/affective, and somatic depressive symptoms across these groups, extending its use to the community. This knowledge is timely as medicine moves towards alternative payment models emphasizing high-quality and cost-efficient care, which will likely incentivize behavioral and population health efforts. We also provide a consistent, evidence-based approach for calculating PHQ-9 subscale scores.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1002/da.22940

Cites

71 of 71 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • The PHQ-8 has good reliability (α=.86-.89) and has shown validity across different racial and ethnic groups.

    “The PHQ-8 has good reliability (α=.86-.89) and has shown validity across different racial and ethnic groups [69].”

    From App-Based Mindfulness Meditation for People of Color Who Experience Race-Related Stress: Protocol for a Randomized Controlled Trial (2022) · cites this paper as [69]

    The Results (page unavailable) find measurement invariance across racial and ethnic groups, and Methods > Measures describes high internal consistency—but both findings concern the PHQ-9. The cited paper does not establish these properties for the PHQ-8 or report the claimed α=.86–.89 range. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Authors
Patel, J. S., Oh, Y., Rand, K. L., Wu, W., Cyders, M. A., Kroenke, K., & Stewart, J. C.
Affiliations
  • Indiana University – Purdue University Indianapolis (US)
  • Texas Tech University (US)
  • Regenstrief Institute (US)
Corresponding author
Stewart, Jesse C.
Venue
Depression and Anxiety
Publisher
Wiley
Published
2019-07-29
Volume
36
Issue
9
Pages
813-823
ISSN
  • 1091-4269 (ISSN-L)
  • 1520-6394
OpenAlex
W2786060782
Cited by
306
References
71
Citations by year
2026: 47 · 2025: 49 · 2024: 83 · 2023: 49 · 2022: 36 · 2021: 25 · 2020: 11 · 2019: 5 · 2017: 1
FWCI
13.73
Type
journal article
Language
en
Open access
Locations
  • Depression and Anxiety (publishedVersion)
  • PubMed (publishedVersion)
  • PubMed Central (acceptedVersion)
  • IUScholarWorks (Indiana University) (submittedVersion)
  • PMC
Funders
National Heart, Lung, and Blood Institute (R01HL122245)
Topics
  • Social Sciences › Decision Sciences › Management Science and Operations Research › Psychometric Methodologies and Testing
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Anxiety, Depression, Psychometrics, Treatment, Cognitive Processes
  • Social Sciences › Psychology › Social Psychology › Mental Health Treatment and Access
Keywords
PHQ-9, measurement invariance, depression screening, factor structure, somatic depressive symptoms, sociodemographic disparities, sex differences, race and ethnicity, educational attainment, NHANES
MeSH terms
Patient Health Questionnaire, Depression, Depressive Disorder, Educational Status, Ethnicity, Factor Analysis, Statistical, Female, Humans, Male, Mass Screening, Middle Aged, Nutrition Surveys, Psychometrics, Racial Groups

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review