Validity of the Patient Health Questionnaire-9 in neurologic populations

Background Because of symptom overlap, there is uncertainty about the validity of depression rating scales in neurologic populations. The objectives of this study were to evaluate the validity of the Patient Health Questionnaire-9 (PHQ-9) for detecting Diagnostic and Statistical Manual–defined major depressive episodes in people with neurologic conditions.

Methods Participants were recruited from outpatient clinics for multiple sclerosis, epilepsy, migraine, Parkinson disease, and stroke for this cross-sectional study. Participants were administered a questionnaire (this included the PHQ-9), chart review, and a follow-up telephone interview. The Structured Clinical Interview for Depression was used as the reference standard for psychiatric diagnoses. The performance of PHQ-9 was analyzed using sensitivity, specificity, diagnostic odds ratios (DORs), and receiver operator curve analysis.

Results All neurologic subpopulations had a specificity greater than 78% and sensitivity greater than 79% at a cut-point of 10. Using a random-effects model, the I-squared value was 13.7%, and Tau² was 0.05, showing homogeneity across the neurologic subpopulations. The pooled DOR was 25.3 (95% confidence interval [CI] 14.9–42.8). Meta-analytic analysis found that for sensitivity, the pooled estimate was 90% (95% CI 81–97), and for specificity, it was 85% (95% CI 79–90).

Conclusions Despite theoretical concerns about its validity, the PHQ-9 performed well at its standard cut-point of 10. Consistent with the literature, being able to use a validated, brief tool that is available publicly should improve case finding of depression in neurologic populations. When considering clinical practicality along with the findings of this analyzed, this study confirmed that the PHQ-9 is valid in a general outpatient neurologic population.