The factor structure of the Patient Health Questionnaire-9 in stroke: A comparison with a non-stroke population

Abstract

Background

It is unclear if certain post-stroke somatic symptoms load onto items of the Patient Health Questionnaire-9 (PHQ-9), a self-report depression questionnaire. We investigated these concerns in a stroke sample using factor analysis, benchmarked against a non-stroke comparison group.

Methods

The secondary dataset constituted 787 stroke and 12,016 non-stroke participants. A subsample of 1574 comparison participants was selected via propensity score matching. Dimensionality was assessed by comparing fit statistics of one-factor, two-factor, and bi-factor models. Between-group differences in factor structure were explored using measurement invariance.

Results

A two-factor model, consisting of somatic and cognitive-affective factors, showed better fit than the unidimensional model (CFI = 0.984 versus CFI = 0.974, p < .001), but the high correlation between the factors indicated unidimensionality (r = 0.866). Configural invariance between stroke and non-stroke was supported (CFI = 0.983, RMSEA = 0.080), as were invariant thresholds (p = .092) and loadings (p = .103). Strong invariance was violated (p < .001, ΔCFI = −0.003), stemming from differences in the tiredness and appetite intercepts. These differences resulted in a moderate overestimation of depression in stroke when using a summed score approach, relative to the comparison sample (Cohen's d = 0.434).

Conclusions

The findings suggest that the PHQ-9 measures a single factor in stroke. Because stroke patients may report higher tiredness on item 4, caution is advisable when classifying patients as depressed if they are near the cut-off and have significant post-stroke fatigue. Caution is also advised when comparing total scores between stroke and other populations.