Reliability and equivalence of the 8-item version of the Patient Health Questionnaire (PHQ-8) to screen for depressive symptoms in adults with chronic pain: A representative U.S. population study.

BACKGROUND

The 8-item Patient Health Questionnaire (PHQ-8) is the most widely implemented clinical depression screener in the world and is increasingly influential in population surveillance of unmet mental health needs. However, in the context of chronic pain (CP), depression screening could be complicated by overlapping symptoms. Inaccurate attribution of CP symptoms to depression could lead to overestimation of depressive symptom severity or even artifactual inflation of depression prevalence estimates. Results of small, non-representative studies are heterogenous; suitability of the PHQ-8 to screen for depression in those with CP remains unclear. We aimed to determine the reliability and equivalence of PHQ-8 in adults with and without CP.

METHODS

This cross-sectional observational population study used representative data from the 2019 National Health Interview Survey. Descriptive statistics and visualizations were generated; reliability, cross-group equivalence and measurement invariance were assessed.

FINDINGS

The final sample contained 30,983 U.S. adults. Prevalence of clinically significant depressive symptoms was 3.5% in those without CP, 20.1% in those with CP, and 34.8% in those with high-impact chronic pain (HICP). Reliability, measurement invariance and cross-group equivalence were observed at the configural, metric, and scalar levels. No evidence consistent with overestimation of depression prevalence or severity in the contexts of CP or HICP was observed.

INTERPRETATION

The PHQ-8 is reliable and measures depressive symptoms equivalently in the context of CP, supporting the continued use of PHQ-8 to screen for depression in clinical and research settings and in population surveillance of adults with and without CP. CP was associated with five times higher prevalence of clinically significant depressive symptoms; findings suggest this reflects true cross-group differences in prevalence and severity of depressive symptoms and is unlikely to result from measurement issues.

FUNDING

This work was funded in part by the Comprehensive Center for Pain & Addiction, University of Arizona, and by NIH K23HD104934.